Timestamps are in GMT/BST.
[0:00] <Hamled|Erp|> downloaded the one from freenetproject.org/snapshots
[0:00] <Hamled|Erp|> just seednodes.ref.bz
[0:00] <Hamled|Erp|> not one of the unstable ones
[0:01] <jabawok_h2> still no connections?
[0:02] <Hamled|Erp|> nope
[0:02] <salahx> Try thesee instead: http://home.earthlink.net/~salahx/freenet/unstable.ref.gz
[0:02] <jabawok_h2> hmm double check your node can make connections out (ie not firewalled)
[0:02] <salahx> a lot of nodes have been choking on the seednodes.ref from the site
[0:03] <Hamled|Erp|> well it was working fine just yesterday or so, jabawok_h2
[0:03] <Hamled|Erp|> salahx, still no luck
[0:04] <salahx> it takes a few minutes...
[0:04] <Hamled|Erp|> k
[0:07] * Ash-Foxeus (Hal-9000@aak47.neoplus.adsl.tpnet.pl) Quit (Connection timed out)
[0:10] <Hamled|Erp|> well so far no connections
[0:10] <Hamled|Erp|> I'm gonna try reinstalling freenet
[0:10] <salahx> ok
[0:11] <Hamled|Erp|> I most likely messed something up without realizing
[0:13] <KenMan> hi salahx
[0:13] <salahx> Hi KenMan
[0:14] <KenMan> thelema probably off to sleep, huh ?
[0:15] <KenMan> i'll take his silence as a 'yes'
[0:16] <Hamled|Erp|> I think I figured out what the problem probably was
[0:16] <Hamled|Erp|> I only replaced freenet.jar, not the rest of the rest of the files like freenet-ext.jar
[0:16] <KenMan> well, i don't know what toad is doing with this new unstable network, but it looks 'okay' so far... nothing *obviously* broken yet
[0:17] <KenMan> Hamled|Erp|: but you're pushing packets now, eh ?
[0:17] <Hamled|Erp|> dunno, it's still starting
[0:17] <KenMan> oh, well my fingers are crossed for you...
[0:17] <Hamled|Erp|> :P
[0:18] <Hamled|Erp|> thanks
[0:18] <Hamled|Erp|> wondering why it's taking so long to startup...
[0:18] <KenMan> somehow unstable is showing success rates over 10% for 3 hours straight. I'm sure this is just coincidence, or something to do with a new network reset... not 'the solution' yet. It's too soon for that.
[0:20] <KenMan> I wonder who's doing what... it doesn't appear to be from someone running my silly fetcher script...
[0:21] <Hamled|Erp|> hmmm
[0:21] <KenMan> no luk choi ??
[0:21] <Hamled|Erp|> oh, looks like it was getting OOM errors so it wasn't starting up
[0:21] <KenMan> grrr
[0:22] <KenMan> maybe you should reduce the size of seednodes.ref, if it is over 20 megs... just go to about 50% into the file, find the word end, and delete everything after that.
[0:22] <Hamled|Erp|> honestly windows has very odd memory management... shutdown one process using 50MB of ram, and suddenly you have 150MB freed :P
[0:23] <salahx> GDI usage probably
[0:23] <Hamled|Erp|> got over 200MB free, and java is reporting OOM exceptions...
[0:23] * qumqats (qumqats@outel.org) has joined #freenet
[0:24] <KenMan> did you get a freenet.log file ?
[0:24] <Hamled|Erp|> yeah, taht's where I'm seeing the exceptions reported
[0:25] <Hamled|Erp|> last line before the errors is "
[0:25] <Hamled|Erp|> Aug 10, 2004 12:40:48 AM (freenet.node.Main, main, NORMAL): Loaded bucket factory"
[0:25] <KenMan> how far did it get before the exceptions showed up ? did you get to "Starting interfaces.."
[0:25] <KenMan> ah, the next step is to read seednodes. Do what i suggest just above...
[0:26] <Hamled|Erp|> *shrug* ok
[0:26] <KenMan> just delete the last half of that seednodes.ref file.
[0:27] <KenMan> in a perverse way, i suppose it is a good sign that we have enough nodes to overflow this file...
[0:28] <Hamled|Erp|> of course, it's perhaps not a good sign that loading a 20MB file should require over 300MB of ram
[0:28] <KenMan> no, that simply indicates that windows is working as designed...
[0:28] <Hamled|Erp|> :P
[0:29] <KenMan> sorry, couldn't resist
[0:29] <Hamled|Erp|> I had no problem with that large a file before, but I haven't rebooted in a while :|
[0:29] <salahx> THe Windows GDI is leakier than Fred
[0:29] <KenMan> what editor ??
[0:29] <KenMan> ouch, salahx. That is pretty harsh. Even MS doesn't deserve that !!
[0:29] <Hamled|Erp|> eh, no I mean freenet was running fine before
[0:29] <Hamled|Erp|> with the same seednodes.ref file
[0:29] <KenMan> well, try it anyway.
[0:30] <Hamled|Erp|> I don't think this is related to the ref file...
[0:30] <Hamled|Erp|> tried salahx's 5MB file
[0:30] <Hamled|Erp|> still getting OOM errors
[0:30] <KenMan> i accept your logic, and raise you one confusion
[0:30] <Hamled|Erp|> didn't want to just cut the other file in half
[0:31] <Hamled|Erp|> looked like there were probably more logical points at which to remove the rest, but I didn't want to spend the time to determine that
[0:31] <Hamled|Erp|> I think I'll just reboot and see if that helps
[0:31] <KenMan> the next line after "Loaded bucket factory" should be "read seed nodes" , FWIW
[0:31] <Hamled|Erp|> k
[0:32] * Hamled|Erp| (~hamled@pool-68-163-62-23.phil.east.verizon.net) Quit (Read error: 104 (Connection reset by peer))
[0:32] <KenMan> yes, the primary recommendation of Windows. Windows NT should be rebooted at least one a week. This was in an actual product guide.
[0:32] <KenMan> From MS !!
[0:33] <KenMan> I think it is somewhere in their knowledge base...
[0:33] <KenMan> not that this should raise any eyebrows or anything. Business as usual.
[0:35] * dERMOTH (~dERMOTH@dsl-131-244.aei.ca) has joined #freenet
[0:37] * Hamled (~hamled@pool-68-163-62-23.phil.east.verizon.net) has joined #freenet
[0:38] * dERMOTH is now known as dERMzzz
[0:43] <Hamled> well, I have restarted
[0:43] <Hamled> and in doing so, I have determined that my installation of freenet must be the problem
[0:51] <Hamled> hmm... installed the unstable version and it's working, have only one connection, but that's a start
[0:52] <KenMan> cool... 60187 ?
[0:53] <KenMan> I'm seeing some unbelievable success rates, but surely this will not last as more people join in...
[0:53] <KenMan> it must just be a fluke. I've never served out 275 items in an hour's time. Surely it is just a fluke.
[0:54] <Hamled> actually 60185
[0:54] <KenMan> odd, sentData is reading 100% , and it shows only 101 items in that time. Oh no, I wonder what the no. transfers is doing...
[0:55] <Hamled> updating
[0:55] <Hamled> maybe this'll help it out
[0:55] <Hamled> that or it'll break
[0:55] <Hamled> :P
[0:55] <KenMan> heh, it shows zero active xfers right now. Well, 101 items out in an hour is pretty good in any case.
[0:55] <Hamled> it's like a real internet! wow!
[0:56] <KenMan> what do you mean ?
[0:56] <Hamled> I'm just making a joke about the speed at which things crawl in freenet
[0:56] <Hamled> atleast, from what I've seen
[0:56] <KenMan> oh. :(
[0:56] <KenMan> well, anonymity has its costs
[0:57] <Hamled> Built on a distributed server network of laptops from 1992... Freenet is home to FreeSpeech and actual internet turtle races
[0:58] <KenMan> heh... it has a lot of room for improvement.
[0:59] <Hamled> well lots of things do
[1:00] <KenMan> toadnode is backed off for 86235 seconds. So much for removing the MRI cap of 600 seconds...
[1:00] <Hamled> new version seems to be stuck on Aug 10, 2004 1:12:38 AM (freenet.node.states.announcing.Announcing, YThread-1, NORMAL): Found 0 announcement targets for this node.
[1:01] <KenMan> something wrong with your seednodes. You need the new one, let me post it for you...
[1:02] <KenMan> Hamled - fetch : http://mywebpages.comcast.net/jkcorson/FreeNet/seednodes.ref
[1:02] <Ash-Fox> Right... On tv, there was a spammer who claimed spam didn't hurt anyone.... I've lost big buisness deals because I couldn't find the e-mails sent to me in the amount of spam I get a week: http://shell.quickfox.org/~Ash-Fox/theespam.jpg now he tells me this didn't hurt me?
[1:03] <KenMan> delete rt* ngrt* ls* and 'node'
[1:03] <KenMan> Ash-Fox: into every life a little spam must fall
[1:03] <Ash-Fox> KenMan, look at the url, that's how much spam I get in a week
[1:04] <KenMan> you must be running freenet, access is slow...
[1:04] <salahx> the mRI are insane on this experematal net
[1:04] <KenMan> salahx: all of them ?
[1:04] <salahx> 2 of them
[1:04] <KenMan> toadnode surely is insane, no doubt.
[1:05] <KenMan> I hope the other one is not me - 68.34...
[1:05] <salahx> 14695049ms !
[1:05] <Hamled> now it's sitting on
[1:05] <Hamled> Aug 10, 2004 1:21:50 AM (freenet.node.Main, main, NORMAL): seeded routing table
[1:07] <KenMan> Ash-Fox: it's going nowhere... i can't get a connection. Browser sits on 'Connecting to shell.quickfox.org...' for a long time
[1:08] <KenMan> somehow, I had 1000 incoming requests for the last complete hour...
[1:08] <Ash-Fox> KenMan inbox says: 7066 new messages
[1:08] <KenMan> so my MRI to /someone/ must be okay.
[1:08] <KenMan> Ash-Fox: that's a lot of messages to have to click delete on !! I do feel your pain.
[1:09] <Ash-Fox> KenMan I filter an average of 750 messages a day at the mailserver
[1:09] <KenMan> I have decent luck with mozilla's email spam detection code in the client...
[1:09] <Ash-Fox> KenMan It's not like I don't have a spam filter... this spam filter is rather effective to the amount i used to get
[1:09] <Ash-Fox> KenMan, the system in thunderbird and mozilla doesn't seem so great, constantly flags legitimate messages
[1:10] <KenMan> operation timed out attempting to connect to AshFox. Must be all the spam your server is busy slinging... eating all your bw.
[1:10] <Ash-Fox> KenMan, probably
[1:10] <KenMan> :(
[1:10] <Ash-Fox> try http://shells.quickfox.org/~Ash-Fox/theespam.jpg sometimes that works
[1:12] <KenMan> not looking good... anyway, how many users do you have on your mailserver ? Do you have a handle on the ratio of good to spam ?
[1:12] <Ash-Fox> I receive most of the spam, and as for users, there are about 12
[1:12] <Ash-Fox> When I mean most of the spam, I mean 99.9% of the spam
[1:12] <KenMan> many businesses and providers report some ungodly ratios of spam to good stuff...
[1:13] <KenMan> perhaps spam captured your address as administrator of the server or something... ??
[1:13] * Ash-Fox has a e-mail address that is 9 years old
[1:14] <Ash-Fox> That's the one that collects the most spam
[1:14] <KenMan> i think spamming deserves the death penalty. Public canings, followed by a hanging.
[1:14] <KenMan> then followed by a beheading.
[1:15] <salahx> in the electric chair
[1:15] <Hamled> bleh, how can I get this thing to use a different reference file?
[1:15] <KenMan> and finally, releasing wild hyenas on the remains...
[1:15] <salahx> in a gas chamber..while beign shot
[1:15] <salahx> PETA will never let the hyena do that!
[1:16] <KenMan> okay then, gasoline and flame to the corpse.
[1:16] <KenMan> followed by a jig on the ashes by the greatest of victims of spam .
[1:17] <Ash-Fox> That's not fair
[1:17] <KenMan> "well, my shoes got a little dirty in the process, but I am consoled by the opportunity to dance on the remains of the person who sent me 479 offers to enlarge my penis. For god's sake, I am a woman."
[1:18] <Ash-Fox> I'd prefer they live out their lives... under constant problems, such as, a spammer may not approach any kind of computer or use one, so no ATMs, no cable/satelite decoders etc..
[1:18] <Hamled> blah, I'm sick of trying to get this thing to work
[1:18] <KenMan> that's not half bad... and they have to wake up every time one of their victims feels the need to call them on the telephone.
[1:19] <salahx> i dunno..still woudln't to bad to pput them in a locked room with a printer, adn an e-mail address
[1:19] <salahx> every time someoe esends a spam e-mail the printer prints it...
[1:19] <salahx> ...until the whole room is filled with papaer and the spammer sufficates!
[1:20] <KenMan> plus, victims are allowed to enter the home of the convicted spammer , and take whatever they like. Such as food. This causes the spammer to have to go to the store everytime he opens the refridgerator. Seems like minor payback for all the time of others they have wasted
[1:20] <Ash-Fox> That's destroying our rain forests
[1:20] <salahx> good point
[1:21] <KenMan> the spammer must wear a vibrating beeper (much like people under home-arrest) and they are not allowed to turn it off.
[1:21] <KenMan> victims are allowed to call the spammer any time they wish, and watch their reaction on a public webcam installed on the forehead of the spammer.
[1:22] <Ash-Fox> spammers must pay for a 0800 number
[1:23] <KenMan> community highway service while wearing a jumpsuit that identifies them as a convicted spammed. Do you know how many of the paid highway workers get spit on and have things thrown at them ??
[1:23] <KenMan> At least here in the good ol' USA...
[1:24] <KenMan> but then , I imagine these guys would just get run over in their first day of service
[1:24] <KenMan> kinda takes the fun out of it. Not that I've ever tried to hurt a regular highway worker. I can't imagine why people do it.
[1:25] <KenMan> But there are a lot of stupid people in this world.
[1:25] <Ash-Fox> they can hurt someone and get away with it i suppose
[1:25] <Ash-Fox> we don't have that here
[1:26] <KenMan> A guy was telling me, he gets a glass bottle thrown at him at least once a week. Moving at 100kph !!
[1:26] <KenMan> That is just plain crazy.
[1:27] <Ash-Fox> Americans...
[1:27] <KenMan> well, some of them, anyway.
[1:29] <KenMan> Believe it or not, we have our share of moral and decent folk too, but they are becoming a minority :(
[1:29] <salahx> yeah :(
[1:30] <KenMan> and they are not the ones who make the headlines for disemboweling their mother because she brought them a diet soda, instead of a regular one.
[1:31] <Ash-Fox> never knew that happened untill you told me
[1:31] * qumqats (qumqats@outel.org) has left #freenet
[1:32] <KenMan> here in baltimore, a man got mad at his girlfriend (about a month ago), so he cut the heads off her 3 young children .
[1:32] <KenMan> about 2 years ago, a man cut the head off his mother (with whom he lived) with a japanese harikari sword
[1:33] <salahx> well there are nuts everythese, that's why its news :)
[1:33] <Ash-Fox> eww
[1:33] <KenMan> i don't remember the reason for that one. He /might/ have had a mental problem.
[1:33] <KenMan> that was over in my sister's neighborhood.
[1:34] <KenMan> but then, there are 1 million people in my greater metropolitan area. Simple stats say some of they are gonna be kooks.
[1:34] <KenMan> You have to do something fairly creative to make national news in this country.
[1:36] <KenMan> Those guys down in florida who killed 6 people with baseball bats (over a stolen video game console), that isn't worthy. Gang violence like that been going down for decades.
[1:36] <Ash-Fox> mhmm
[1:43] * Iakin (Iakin@as25-4-6.kp.g.bonet.se) has joined #freenet
[1:45] <KenMan> now those anthrax attacks, likely committed by an american, that was news worthy. At least, when news channels tend toward the more sensational news.
[1:46] <KenMan> hi iakin, how are things ?
[1:46] <KenMan> still busy as ever at work ?
[1:47] <KenMan> 4 of 5 routes backed off for over 200 seconds each...
[1:49] <KenMan> there we go, success dropped to 20% of what it was... i knew it was too good to be real. 18% success on 1521 queries, i mean, really. Like I'm gonna believe that is normal...
[1:51] <KenMan> wow, i'm giving out MRIs to all my peers of under 7 seconds. Now, salahx , is it fair in a network like this for individual nodes to be so divergent in their calculated MRI values ??
[1:51] <salahx> THey never did catch the Antrax guy, did they ?
[1:51] <KenMan> not yet...
[1:51] <KenMan> they just started looking into some guy in new york state this past week. Finally decided the other guy had been harassed enough.
[1:52] <salahx> maybe the FBI will be lucky and the brother of theanthrax killer will turn him in :)
[1:52] <KenMan> Three years of the FBI calling Hatfield "an interesting person" but "not a suspect." He can't get a job anywhere.
[1:52] <KenMan> good reference :)
[1:54] <KenMan> i wonder if Anthrax guy will stay anonymous for 20 years too...
[1:54] <Iakin> Hello kenman
[1:54] <KenMan> god morgan :)
[1:54] <Iakin> Nope, work is getting better nodw
[1:55] <Iakin> But I have been on vacation
[1:55] <salahx> well as long as he doesn't get published any 45kword rants^W manifestos maybe :)
[1:55] <KenMan> let's hope Anthrax guy doesn't use start using Freenet. We don't need that kind of attention.
[1:57] <KenMan> The last use of real poison stuffs (castor bean crap, what's it called again? oh ricin) was accompanied by some guy ranting about US interstate trucking regulations.
[1:57] <salahx> ricin
[1:57] * Iakin notices that nowadays it seems like routingtime, not bw, limits his node
[1:57] <KenMan> See, we don't need to fear the Arab terrorists, we have sufficient homegrown ones now.
[1:57] <salahx> yep
[1:58] <KenMan> routing time while requests tick down 20 RNFs ...
[1:58] <KenMan> did you know that somewhere around 40% of all requests that reach your node wind up without a single route available to move further ? This requires your node to RNF on 40% of the incoming queries, without any useful processing whatsoever.
[1:59] <salahx> that like back when we were QRing
[1:59] <Iakin> Kenman: Right
[1:59] <KenMan> And the ones that DO make it though your node face the same challenge on every step of their circuitous journey...
[1:59] <KenMan> :(
[1:59] <Iakin> So.. I should then try to speed up routing even further then
[2:00] <KenMan> wink !
[2:00] <Iakin> Hmmm...
[2:00] <Iakin> Wonder if people tolerates yet some jbigi updates
[2:00] <KenMan> that is exactly like when we were QRing...
[2:00] <Iakin> (with methods that doesn't really belong in it..
[2:01] <KenMan> whatcha got in the pipeline, Iakin ??
[2:01] <KenMan> more nuclear software ?
[2:01] * salahx (salahx@sc1-24.217.174.147.charter-stl.com) Quit ()
[2:01] <Iakin> Number of uncontactable nodes 5
[2:01] <Iakin> Contacted and attempted to contact node references 568
[2:01] <Iakin> Contacted node references 68
[2:02] <Iakin> How come we dont consider all those nodes that have 100% failed connctions as uncontactable?
[2:02] <Iakin> ..only those that has a void/void address
[2:02] <Iakin> KenMan: I'll see what I can do..
[2:02] <KenMan> good question, put it on the stack of good questions, please ! ;)
[2:03] <Iakin> As usual the problem is tha bigint math
[2:03] <Iakin> but I have to finish the dynamic-linked lib and that readme forst
[2:03] <Iakin> ok, bbl
[2:03] <KenMan> see ya.
[2:05] <KenMan> strange. We have some pretty consistent traffic on unstable. I think it is (mostly) all running through my node, but still, it is suspicious.
[2:05] <KenMan> Because there are a fair number of successes.
[2:06] <Iakin> Current estimated requests per hour: 5668.866799268669
[2:06] <Iakin> Current global quota (requests per hour): 5766.865714072736
[2:06] <Iakin> Is any of these what the node currently experiences?
[2:06] <KenMan> on stable ?
[2:06] <KenMan> I think the first one is an instant calculation / projection...
[2:07] <KenMan> watch it and see how fast it fluctuates, that's all i can suggest.
[2:08] <KenMan> okay, now everything makes more sense. I was resolving almost everything with the local datastore. 18% success - no wonder. RoutingsOut haven't reached above 2% so far.
[2:10] <Iakin> Hmm is requestDelayCutoff still in use?
[2:11] <KenMan> i seriously doubt it. I just tripped over that option for the first time about 2 days ago.
[2:11] <iip_i2p> <Newsbyte> blasfemy! Everyone knows there is no obsolete code left in there!
[2:11] <iip_i2p> <Newsbyte> ;-)
[2:12] <KenMan> Iakin, it appears to be obsolete.
[2:12] <KenMan> heh, i think 75% of it is obsolete, either on purpose or by practice...
[2:13] <KenMan> oh wait, it may still be used in Node.java ...
[2:15] <KenMan> yeah, looks like it is. But it may just be a disguise.
[2:16] <KenMan> night all !! zzz
[2:39] <ShaunMacPherson> see u KenMan
[2:40] <iip_i2p> <Newsbyte> pff...who is sleeping this early?
[2:41] <ShaunMacPherson> not me :P
[2:47] * Zorix (Brandon@fl-64-45-226-120.dyn.sprint-hsd.net) Quit (Read error: 60 (Operation timed out))
[3:11] * hobx (~hobx@h196n1fls21o1077.bredband.comhem.se) Quit (Remote closed the connection)
[3:21] <theLostFloppy> 'morning
[3:51] <iip_i2p> <Newsbyte> ok, managed to create the new FreenetSCIP, based on 5090
[3:52] <iip_i2p> <Newsbyte> evening
[3:58] <iip_i2p> <Newsbyte> 'evening, I said
[3:58] <iip_i2p> <Newsbyte> what's wrong with eyekon this time?
[3:59] <iip_i2p> <Newsbyte2> test
[4:00] <iip_i2p> <mazzanet> *slaps Newsbyte around a bit with a large trout*
[4:01] <mazzanet> .
[4:14] * leex-zzzz is now known as leex-work
[4:16] <leex-work> all sleeping probly
[4:16] * Zorix (Brandon@fl-64-45-226-120.dyn.sprint-hsd.net) has joined #freenet
[4:43] <iip_i2p> <Newsbyte2> mazza ain't (wasn't)
[4:52] * Hory (~Miranda@81.196.25.110) has joined #FreeNet
[5:00] * thelema (~thelema@adsl-65-65-200-192.dsl.stlsmo.swbell.net) Quit (Read error: 110 (Connection timed out))
[5:01] * Iakin3 (~nb@212.105.104.162) Quit (Read error: 110 (Connection timed out))
[5:01] * Zorix (Brandon@fl-64-45-226-120.dyn.sprint-hsd.net) Quit (Read error: 110 (Connection timed out))
[5:23] * guido^pe (~unknown@dsl-082-082-156-230.arcor-ip.net) has joined #freenet
[5:47] <theLostFloppy> I AM AWARE
[5:47] <theLostFloppy> (sorry, couldn't resist)
[6:11] <iip_i2p> <Newsbyte2> looking at old SF-movies, are you? ;-)
[6:13] <iip_i2p> <Newsbyte2> so, read my latest post? Anyone knows of a 'hobbex' that would be an ex-coder of freenet? I'm more then 4 years with the freenetproject, but the name doesn't ring a bell (then again, I'll forget mazza too, in a copple of days, if he's not around ;-)
[6:13] * mazzanet points to hobx_
[6:16] * thelema (~thelema@adsl-65-65-203-187.dsl.stlsmo.swbell.net) has joined #freenet
[6:16] * Zorix (Brandon@fl-64-45-226-120.dyn.sprint-hsd.net) has joined #freenet
[6:17] <hobx_> mazzanet: Points to me about what?
[6:18] <mazzanet> [20:31:47] <iip_i2p> <Newsbyte2> so, read my latest post? Anyone knows of a 'hobbex' that would be an ex-coder of freenet? I'm more then 4 years with the freenetproject, but the name doesn't ring a bell (then again, I'll forget mazza too, in a copple of days, if he's not around ;-)
[6:18] <hobx_> so he is either lying about how long he has been around, or a somewhat thick.
[6:18] <mazzanet> heh
[6:20] <iip_i2p> <Newsbyte2> thick? thick?!
[6:20] <iip_i2p> <Newsbyte2> I'm fat, even!
[6:21] <hobx_> hobbex is my sourceforge user name, so it appears in every CVS log.
[6:21] * sanity (~ian@81-178-89-164.dsl.pipex.com) has joined #freenet
[6:21] * ChanServ sets mode +o sanity
[6:22] <iip_i2p> <Newsbyte2> so, how about the anonimty of freenet?
[6:23] <iip_i2p> <Newsbyte2> hey...didn't I see you yesterday here too?
[6:23] <hobx_> me?
[6:23] <iip_i2p> <Newsbyte2> /me makes a point of rapidly forgetting nicks :-)
[6:23] <hobx_> I've been in here more or less nonstop for five years...
[6:24] <hobx_> And freenet offers some obscurity of identity, but not much for anonymity.
[6:24] <theLostFloppy> AAH Distcc pisses me off
[6:24] <sanity> hobx: does anything?
[6:24] <iip_i2p> <Newsbyte2> do other coders agree with that?
[6:25] <hobx_> I mean, the basic idea is that you can see what people request but they have deniability because they could just be forwarding it.
[6:25] <sanity> hobx: even dining cryptographers is vulnerable to attack
[6:25] <iip_i2p> <Newsbyte2> well, you basically only see a hash, no?
[6:25] <hobx_> But factor in high HTL values and do some statistical analysis and you can get the false positive rate down pretty low (and as oppressive regimes usually aren't that worried about false positives)
[6:26] <hobx_> sanity: Of course.
[6:26] <hobx_> My point is just that Freenet is considerably weaker then just basic onion routing as far as anonymity is concerned.
[6:27] <iip_i2p> <Newsbyte2> anyway, that will become fixed when it uses a pre-route onion-style thingy, no?
[6:28] <sanity> hobx: freenet decouples transmitter and receiver through caching in a way that onion routing does not
[6:28] <hobx_> Newsbyte: Starting with onion-routing steps would help. It also needs better traffic analysis protection though.
[6:29] <hobx_> sanity: I'm talking about trolling for identities by running nodes in the network.
[6:29] <iip_i2p> <Newsbyte2> and why do you doubt there are many thousands of nodes being used? On stable, at least, there seem to be as much
[6:29] <sanity> hobx: identities of whom?
[6:29] <sanity> hobx: publisher or consumer?
[6:29] <hobx_> sanity: The easiest would be people requesting from lists of banned keys.
[6:30] <hobx_> But publisher works too. Run many nodes, save all the logs. Wait for a bad website to come up, crosscheck the keys of all the site elements against your log, and look for a node that sent several of them with high HTL values.
[6:30] <sanity> hobx: the likelihood that a given request for a key you are looking for going through your node is pretty low in a large Freenet. the likelihood of it happening repeatedly is even lower.
[6:31] <iip_i2p> <Newsbyte2> so, so...you are hobbex! What a clever way to hide your nick into hobx! I never would have guessed! ;-)
[6:31] <hobx_> Do it right and I think you can get a high chance of guilt for a single person a lot of the time.
[6:31] <sanity> hobx: i await your paper
[6:31] <iip_i2p> <Newsbyte2> speaking of which...weren't you going to make a paper, Ian?
[6:32] <sanity> newsbyte: no, theo was
[6:32] <hobx_> sanity: Ok. "Why I am disilliusioned about freenet" is something I can talk about at conferences...
[6:33] <iip_i2p> <Newsbyte2> and about the traffic analysis: making a small chunks thingy of 64kb (or 32 as toad wants it) would help with that
[6:33] <hobx_> Newsbyte: I started respelling my nick when I was ahead of the company I stole the name from on searches.
[6:34] <iip_i2p> <Newsbyte2> ah well, we're all a bit dessillusioned, now and then, about freenet. even Ian, I suspect. But it still has it's potential.
[6:34] <sanity> hobx: i think Brandon has cornered that market
[6:35] <hobx_> Yeah, that's what I mean.
[6:35] <hobx_> Brandon and Steven showed me that one can do that!
[6:35] <iip_i2p> <Newsbyte2> well...I doubt you NEVER felt dissapointed with Freenet, Ian
[6:36] <hobx_> sanity: Except I wouldn't say "Freenet sucks because they didn't do what I said", I would say "They did what I said mostly, but I was wrong."
[6:39] <iip_i2p> <Newsbyte2> that said, one can't say it's totally non working neither, at least with the latest build
[6:39] <iip_i2p> <Newsbyte2> and..ermm...well, being desillusionned and all that...do you still code much on freenet? (and if so: why?)
[6:40] <hobx_> Kenman told me yesterday that his node was rejecting 30-70% of all queries because the node has no open routes.
[6:40] <hobx_> No. The last substantial coding I did was in the winter of 2002/2003
[6:41] <hobx_> Now I just hand around here and annoy Ian.
[6:41] <iip_i2p> <Newsbyte2> yes, ok, then my next question: why do you keep hanging around, then? ;-)
[6:42] <iip_i2p> <Newsbyte2> oh, annoying Ian...ok, that's a good reason! ;-)
[6:42] <iip_i2p> <Newsbyte2> lol
[6:42] <hobx_> I'm a bitter old fart with nowhere else to go.
[6:42] <iip_i2p> <Newsbyte2> true, true
[6:42] <iip_i2p> <Newsbyte2> ;-)
[6:42] <sanity> hobx: i got a mac
[6:42] <sanity> hobx: its pretty darn cool
[6:43] <sanity> i have set it up so that when i move my mouse off the left edge of the screen, it appears on the right edge of my old linux laptop
[6:43] <hobx_> sanity: Sudden urge to mod down comments commenting that DRM is DRM setting in?
[6:43] <iip_i2p> <Newsbyte2> so, what? You're running freenet on OSX, these days
[6:43] <iip_i2p> <Newsbyte2> ?
[6:43] <sanity> hobx: indeed, i wouldn't have got a mac had i not been protected from such badthink
[6:43] <sanity> hobx: are you still banned?
[6:43] <sanity> hobx: that is fucking crazy
[6:43] <hobx_> yes
[6:43] <hobx_> tried it this morning
[6:44] <sanity> hobx: given that i suspect between us we account for about half the highly moderated comments on that bloody website
[6:44] <theLostFloppy> Banned, why?
[6:44] <iip_i2p> <Newsbyte2> banned? from where? slashdot?
[6:44] <hobx_> Newsbyte: This comment: http://slashdot.org/comments.pl?sid=114185&cid=9673617
[6:44] <sanity> newsbyte: no, i run freenet on my TV box, which runs linux
[6:45] <iip_i2p> <Newsbyte2> I read you whining, I mean, complaint about ppl banning you and that other dude of /. , Ian. I took pitty, and since I had moderationpoints to spare at that time, I benevolently gave them away to you guys :-)
[6:46] <iip_i2p> <Newsbyte2> I believe it had something to do with OSX as well
[6:46] <hobx_> Did you get mythtv working?
[6:47] <sanity> hobx: of course not
[6:47] <hobx_> :-)
[6:47] <iip_i2p> <Newsbyte2> but my memory is not good, as hob can confirm
[6:47] <sanity> hobx: only San Francisco-based lawyers can get mythtv working, not a lowly computer scientist like me
[6:48] <hobx_> I would do it, but I don't have a TV
[6:48] <iip_i2p> <Newsbyte2> lemme think...it was about how dreadfull it was that slashdot banned your IP(s) and were unwilling to lift it, because you complained about OSX, as did that other guy
[6:48] <iip_i2p> <Newsbyte2> and that, because the points went up and down like a yoyo
[6:49] <iip_i2p> <Newsbyte2> since I like yoyo's I added my points to it too ;-)
[6:49] <hobx_> In fact I got the entire math department here banned from slashdot - though I'm not sure anybody else has noticed.
[6:50] <iip_i2p> <Newsbyte2> does that ring any bell, or did I waste my modpoints on some other sanity&co? ;-)
[6:52] <iip_i2p> <Newsbyte2> logic would dictate, seen that you just said you had a mac, that it was, indeed, you. Feel free to throw away modpoints in my direction, when you have them. ;-)
[6:55] <iip_i2p> <Newsbyte2> "Freenet doesn't use onion routing (last time I checked), but it does use the concept of sending messages through mutiple hops. But the main difference between Freenet and Tor is that Freenet is an anonymous publishing system and Tor is an anonymizing layer that can work with almost any application."
[6:55] <iip_i2p> <Newsbyte2> now is this fair or unfair, I have to decide...
[6:55] <iip_i2p> <Newsbyte2> now is this fair or unfair, I have to decide...
[6:55] <iip_i2p> <Newsbyte2> I mean, it seems to be correct, but is it fair?
[6:55] <hobx_> You know slashdot threads are dead after 12 hours.
[6:56] <hobx_> Why do you keep revisiting that one?
[6:56] <iip_i2p> <Newsbyte2> as long as it's in my list of metamodpoints, I'm visiting it
[6:58] <iip_i2p> <Newsbyte2> /me is a karmawhore :-)
[6:58] <hobx_> Why?
[6:59] <hobx_> I could go from zero to excellent karma in 12 hours today.
[6:59] <iip_i2p> <Newsbyte2> goodie for you!
[6:59] <hobx_> Get a sport that is a little difficult. Back when I was trying to catch Signal 11 and Enoch Root, that was real karma whoring.
[6:59] <iip_i2p> <Newsbyte2> the only good answer: because I like it :-)
[7:00] <hobx_> I mean, the highest you can get is like 30. Mine peaked at 211 before they started capping it.
[7:01] <hobx_> And that was when their were one or two +5 per story, now there are ten or twenty.
[7:03] <iip_i2p> <Newsbyte2> hmm...I don't see your point. Take, for instance, a regular slut, that likes to have sex. Would it be less fun for her to get more laid then what she got in the times she only got laid two times a week?
[7:04] <jabawok_h2> rofl
[7:04] <iip_i2p> <Newsbyte2> :-)
[7:06] <hobx_> Sex is still sex even if you have a lot of it.
[7:06] <hobx_> When everybody is being moderated up the value of it diminishes.
[7:06] * theLostFloppy goes down to reboot
[7:06] * theLostFloppy (~michaelku@fia41-111.dsl.hccnet.nl) Quit ()
[7:07] * sanity (~ian@81-178-89-164.dsl.pipex.com) Quit (Read error: 104 (Connection reset by peer))
[7:09] * Iakin3 (~nb@212.105.104.162) has joined #freenet
[7:11] <iip_i2p> <Newsbyte2> some would argue (ask the bible-belt dudes) that counts for sex too. The point is rather the amount of pleasure you derive of it on itself.
[7:11] <hobx_> Ok, if you're brain is wired to release endorphines when you get moderated, then I can concede.
[7:12] <iip_i2p> <Newsbyte2> ;-)
[7:12] <hobx_> But as far as I can tell, the only joy I get out of being moderated up is that you are recognized as having written something smarter than most comments, and more people will read it.
[7:14] <iip_i2p> <Newsbyte2> but you could have a point too, after all even Kant found that sex was getting boring and not, ultimately, giving him the satisfaction he was looking after
[7:14] <hobx_> So the sex analogy kind of works here too. There is the purely physical pleasure of sex, which is the same whether you are fucking a slut or non-slut, but there is also the ego boost of being selected by the girl, which does dimish if she sleeps with anyone.
[7:14] <hobx_> Did Kant get laid a lot?
[7:15] <iip_i2p> <Newsbyte2> good question...the moment I send it, I was wondering if it wasn't Nietzshe. I should look it up.
[7:15] <hobx_> Being moderated up gives no physical pleasure. Only the latter applies, and that has been devalued by slashdot drowning in mod points.
[7:15] <hobx_> I don't think Nietzsche go laid much.
[7:16] <hobx_> He was a misogynist who only got along with his sister.
[7:16] * sanity (~ian@81-178-89-164.dsl.pipex.com) has joined #freenet
[7:16] * ChanServ sets mode +o sanity
[7:16] <iip_i2p> <Newsbyte2> nah, you're mistaken in that. The weirdest things can give the sort of pleasure that sex gives (to some extend, at least).
[7:17] <iip_i2p> <Newsbyte2> even mondane things, like eating exquisite food
[7:17] <iip_i2p> <Newsbyte2> not that that is mondane for a lot of ppl ;-)
[7:18] <hobx_> Yes, but not being moderated up.
[7:18] <hobx_> Slashdot today is like the special olympics of debate. Everybody is a winner!
[7:19] <iip_i2p> <Newsbyte2> I don't think that's quite true, though I understand the feelings
[7:19] <iip_i2p> <Newsbyte2> but even now, the +5 are, in general, always better then the -1 :-)
[7:20] <hobx_> not +5 funny
[7:20] <iip_i2p> <Newsbyte2> ah well, 'funny' is a far more subjective interpretation
[7:21] * |ux (~|ux@82-37-17-24.cable.ubr03.wolv.blueyonder.co.uk) has joined #freenet
[7:21] <hobx_> But "I for one would like to welcome our ____ overlords" is always funny...
[7:22] <iip_i2p> <Newsbyte2> for instance, I think I'm quite funny, most of the time. the un-funny freenetters without much sense of humor, however, may find I'm not :-)
[7:22] <iip_i2p> <Newsbyte2> well, yes, one would think so. Though, I believe I've been rated funny AND unfunny, with that one
[7:23] <hobx_> I was being sarcastic...
[7:23] <iip_i2p> <Newsbyte2> it is true a joke loses it's appeal a bit, after hearing it 500 times ;-). Then again, it maybe the 'known-pun' that gets funny...
[7:24] <iip_i2p> <Newsbyte2> nonsense, hobx, that's a prime example why 'funny' is relative
[7:24] <iip_i2p> <Newsbyte2> the first time I came to /. and saw the 'I, for one, welcome' I was rofling
[7:25] <hobx_> In soviet russia, relative is funny
[7:26] <hobx_> newsbyte: The first time I heard it was funny too (on the Simpsons, in it's original context)
[7:26] <iip_i2p> <Newsbyte2> then again, I'm hardly roffling with that one anymore, exept when it's very well worded or there's an extra twist.
[7:27] <iip_i2p> <Newsbyte2> point being, humor differs and it's like taste; you can't objectively debate the intrinsical 'worth' of it
[7:27] <hobx_> yeah yeah
[7:28] <iip_i2p> <Newsbyte2> with 'insighfull' and the rest, that's less the case, and , I must say, sometimes, there a real beauties under those posts. (though granted, some seem a bit overrated to me as well)
[7:29] <hobx_> yes, but they drown.
[7:29] <hobx_> In the Doom 3 release post, Carmack posted, which should have been lifted to the top.
[7:29] <hobx_> But even sorting by mod point, it hardly makes it onto the first page.
[7:29] <iip_i2p> <Newsbyte2> because he's Carmack? ;-)
[7:29] <hobx_> yes
[7:29] <hobx_> it deserved +5
[7:30] <hobx_> But the 20 posts in front of it didn't
[7:30] <iip_i2p> <Newsbyte2> maybe that's more to do with the 3-day limit
[7:31] <iip_i2p> <Newsbyte2> even I had to waste my last point to an inferior post, because I kept waiting for a better post, almost too long
[7:31] <hobx_> No.
[7:31] <hobx_> Because it isn't like good posts aren't being moderated up.
[7:31] <hobx_> There are simply too many mod points in circulation.
[7:31] <hobx_> I only moderated down nowadays.
[7:32] <iip_i2p> <Newsbyte2> ah, so you were the one that stole my points! ;-)
[7:32] <mazzanet> heh
[7:34] <hobx_> There were 57 +4 or +5 posts in that story that end up before john carmack's post when sorting by moderation.
[7:36] <iip_i2p> <Newsbyte2> ah well...I don't think it's that dramatic. I have a healthy reluctance for the 'in my days, it was better' thingy. The only thing that really annoys me sometimes, are posts that get, indeed, mod up, but are completely wrong, because they SOUD like they know what they are talking about. For non-experts, it becomes difficult to know if it's valuable or not, then.
[7:37] <iip_i2p> <Newsbyte2> a prob that can't really be resolved, unless you equip the mods only of teams of known experts that only may moderate on their field of expertise
[7:38] <iip_i2p> <Newsbyte2> which isn't the goal of /. and would be difficult to implemnt, unless on a specialised forum
[7:39] <hobx_> Slashdot has moved from a respresentative system where a small amount of people were selected as moderated and then held accountable for it, to one where everybody gets mod points and then aren't accountable for how they are used.
[7:39] <hobx_> That was dumb.
[7:39] <iip_i2p> <Newsbyte2> but, like, take the solar sail thingy...solar sails are NOT meant to be driven by 'solar wind'. It *sounds* good and logical...if you don't know squat about it, but it just is not right.
[7:40] * guido^pe (~unknown@dsl-082-082-156-230.arcor-ip.net) Quit (Read error: 110 (Connection timed out))
[7:41] <hobx_> Nobody in that story pointed out that while solar sails can accelerate for an infinite amount of time, because the amount of energy you receive dminishes by inverse square, you can only get a finite amount of acceleration out of it.
[7:41] <toad_> so what?
[7:42] <toad_> start at earth you can get a fair speed up
[7:42] <iip_i2p> <Newsbyte2> I have to concur with toad on that one: so what?
[7:42] <hobx_> newsbyte: Hoist the sails, catch the solar wind, and sail the seven galaxies!
[7:42] <iip_i2p> <Newsbyte2> to reach the stars, you don't need infinite acceleration
[7:42] <toad_> anyone done the math on that?
[7:42] <hobx_> because people were comparing it carrying a fixed amount of fuel, which also gives a finite amount of acceleration.
[7:43] <iip_i2p> <Newsbyte2> loads of ppl have, toad :-)
[7:43] <iip_i2p> <mazzanet> bring bring bring bring bring banana phone
[7:43] <iip_i2p> <Newsbyte2> it's not like freenet!;-)
[7:43] <hobx_> So saying "Solar winds are better than fuel because you can accelerate forever" is stupid
[7:44] <iip_i2p> <Newsbyte2> well, they are better, in that they are far more effecient
[7:44] <hobx_> different argument though
[7:44] <hobx_> not a big deal
[7:45] <hobx_> just something I thought about while reading it.
[7:45] <toad_> reaching the stars, with current physics, means one of 3 things: 1. solar sails. 2. bussard ramjets. 3. beam the craft energy
[7:45] <hobx_> (and was frustrated because I am banned from posting=
[7:45] <hobx_> 4. cryogenics
[7:45] <iip_i2p> <Newsbyte2> myes..well, even the acceleration can easily be added (though not infinitely ;-)
[7:45] <toad_> hobx_: no, you need to GET there
[7:45] <toad_> hobx_: I'm talking even with a probe
[7:45] <iip_i2p> <Newsbyte2> if one uses lasers
[7:46] <hobx_> toad_: Yeah but if you have cryogenics you can use a chemical rocket.
[7:46] <toad_> hobx_: no, you cam
[7:46] <toad_> can't
[7:46] <iip_i2p> <Newsbyte2> toad, point 2 is unvalid
[7:46] <hobx_> The voyagers are heading for the stars
[7:46] <toad_> you can't even use fusion
[7:46] <hobx_> from what I read
[7:46] <toad_> in fact, you can't use antimatter drive
[7:47] <toad_> if you want to get there in less than a century and carry your own fuel
[7:47] <toad_> all your own fuel
[7:47] <iip_i2p> <Newsbyte2> bussard ramjets have been analysed more closely, and found they wouldn't work
[7:47] <hobx_> but why less than a century?
[7:47] <toad_> well, voyager will get there in what, a million years?
[7:47] <toad_> Newsbyte2: really? url?
[7:48] <iip_i2p> <Newsbyte2> lol. Not everything I read comes from the Net ;-)
[7:48] <hobx_> The speed of light is 300,000 km/s. At 1 km/s you can get to the nearest stars in about one and half million years.
[7:49] <hobx_> But chemical spacecraft using slingshot effects could probably go a hundred times faster.
[7:49] <toad_> hobx_: voyager is about 15km/sec
[7:49] <toad_> the Breakthrough Physics Programme does the math
[7:49] <iip_i2p> <Newsbyte2> when the density of hydrogen atoms becomes dense enough to become usefull for the combustion, it is also dense enough to cause a severe drain on the captivator, reducing the forward motion to nothing.
[7:50] <hobx_> so it should be there in about 100,000 years
[7:50] <hobx_> So why isn't cryogenics and chemical rockets and option?
[7:51] <hobx_> (exceot that Voyager isn't heading for AC of course)
[7:51] <iip_i2p> <Newsbyte2> that was the reason bussard ramjets won't work, IIRC
[7:51] <hobx_> What about warp drives?
[7:52] <iip_i2p> <Newsbyte2> hobx: chemical rockets will take ages. Even frozen solid, it's always better to keep the freezingperiod shorther then longer, IMHO
[7:52] <hobx_> In "Singularity Sky" they a ship with a black hole in the middle that propelled it by space time warping.
[7:53] <iip_i2p> <Newsbyte2> thus, maybe it would be feasable, but why the heck not use at least fusion-reactors, then, instead of chemicals? ;-)
[7:53] <hobx_> I'm just saying
[7:53] <toad_> http://members.nova.org/~sol/station/interste.htm
[7:53] <toad_> |Under those constraints, a single-stage rocket would need to have a specific impulse (Isp) of over four million seconds with over 97 percent of its total mass comprised as fuel.
[7:53] <hobx_> toad said that cryogenics wasn't sufficient.
[7:53] <toad_> 4Msec ISP => antimatter annihilation
[7:54] <toad_> this is to get to alpha centauri in 40 years
[7:54] <iip_i2p> <Newsbyte2> well...on itself, it's doubtfull they would, exept as a pure theorethical possibility
[7:54] <iip_i2p> <Newsbyte2> read my explanation why bussard ramjets won't work, toad?
[7:55] <hobx_> But you never said anything about 40 years.
[7:55] <hobx_> I don't see why it can't take thousands of years as long as you can keep people frozen.
[7:55] <hobx_> or why even that. Keep frozen DNA on the ship. Then clone and incubate humans when you have reached the destination.
[7:56] <hobx_> That way we could send out thousands of ships, and only those that found suitable destinations would begin populating.
[7:56] <iip_i2p> <Newsbyte2> ermm...you'll need to add something to bring the little lads up too, hob ;-)
[7:57] <hobx_> Newsbyte: I think robotics capable of that is not beyond possible.
[7:57] <iip_i2p> <Newsbyte2> in an infinite universe, everything is possible
[7:57] <hobx_> I'm not talking AI, just something capable of feeding a baby.
[7:58] <hobx_> But I thinking doable in our lifetimes.
[7:58] <iip_i2p> <Newsbyte2> that said, if we have that kind of technology, we can just go as unmortal machines ourselves, no doubt
[7:58] <iip_i2p> <Newsbyte2> oh, that, so you are talking realistically?
[7:58] <iip_i2p> <Newsbyte2> then I have to dissapoint you
[7:59] <hobx_> Why?
[7:59] <iip_i2p> <Newsbyte2> experiments have already been made by purely mechanical 'upbringing' machines (with apes), the results were not encouraging, to use an euphemism
[8:00] <hobx_> true.
[8:00] <hobx_> That is by far the most difficult problem, but I'm gonna live for another 50 or 60 years.
[8:00] <iip_i2p> <Newsbyte2> it seems that 'just feeding a baby' *really* is not enough
[8:01] * greycat (~wooledg@192.35.79.70) has joined #freenet
[8:01] <hobx_> I think though that in some ways humans are easier then animals in this sense. Because humans, even children, can understand that something on a video screen is something real, animals can't.
[8:01] <iip_i2p> <Newsbyte2> I think that, if you want robotic mommies, they are going to be as good as real ones, or at least able to pretend they are, which WOULD involve a considerable AI, indeed
[8:01] <hobx_> But yeah, it is very difficult.
[8:02] <hobx_> Still our best shot at populating space IMHO. We ought to be working on it.
[8:02] <hobx_> Real AI will never happen.
[8:02] <toad_> lol. are preschools, schools, etc? :)
[8:03] <iip_i2p> <Newsbyte2> I disagree. I don't think it is a viable solution, unless you have a true AI, and thus, it's rather a long term issue
[8:03] <toad_> hehehehehe
[8:03] <toad_> hobx_ is now going to invoke the Emperor's New Clothes
[8:03] <toad_> maybe he could even explain it to me; I've never "got" it
[8:04] <hobx_> in what context?
[8:04] <toad_> <hobx_> Real AI will never happen.
[8:04] <hobx_> well, that is just a guess of course.
[8:04] <iip_i2p> <Newsbyte2> whether (older) kids can see a video for what it is, is irrelevant, they NEED a mommy/daddy (real upbringing figure)
[8:05] <hobx_> But to make it more defendable, I will say this: Real AI would require so much knowledge about the human mind and what makes us ourselves, that before we have it long term cryogenics will be perfectly possible.
[8:05] <iip_i2p> <Newsbyte2> I would rather says it's unevitable
[8:06] <toad_> interesting. do you think people would be willing to PAY FOR a mission that wouldn't be done for thousands of years?
[8:06] <iip_i2p> <Newsbyte2> that's something else. I agree with your last part
[8:06] <toad_> will it? has there been ANY progress in cryo over the last century?
[8:06] <hobx_> well, I would.
[8:06] <hobx_> Can't speak for other people.
[8:07] <hobx_> toad_: There has been zero in AI as well.
[8:07] * toad_ would pay for a skyhook :)
[8:07] <iip_i2p> <Newsbyte2> I see cryogenics happen before AI. It doesn't mean AI won't happen, though. In fact, it's in the natural order of (human) things, that, sooner or (probably) later, we'll create one
[8:07] <toad_> hobx_: there has been a lot in AI
[8:07] <toad_> there has been nil in cryo
[8:07] <hobx_> No, there have been a algorithms developed that people call AI.
[8:07] <toad_> yeah
[8:08] <iip_i2p> <Newsbyte2> in fact, I have a theory that says that humans can create ANYTHING they set their mind too, if it's not countrary to physical laws
[8:08] <hobx_> We can't even make a robot as smart as a housefly.
[8:08] <iip_i2p> <Newsbyte2> and making an AI isn't
[8:08] <toad_> there is no actual human level AI
[8:08] <toad_> so what?
[8:08] <toad_> there has been progress on several different ways of mimicking intelligence, and most of them have found applications
[8:08] <hobx_> toad_: I would settle for Dodo level AI.
[8:08] <iip_i2p> <Newsbyte2> actually, we've made robots that are as smart as coakroaches ;-)
[8:09] <toad_> hobx_: I was under the impression that we had made robots as smart as small dumb insects
[8:09] <toad_> no, not quite, actually, iirc...
[8:09] <toad_> robotic vision is coming along rather nicely
[8:09] <iip_i2p> <Newsbyte2> you are both wrong, there has been slight progress in both area's
[8:09] <toad_> Newsbyte2: there has been progress in cryo? What? When?
[8:10] <hobx_> toad_: I don't believe that those algorithms have anything to do with real AI.
[8:10] * |ux (~|ux@82-37-17-24.cable.ubr03.wolv.blueyonder.co.uk) Quit (Read error: 110 (Connection timed out))
[8:10] <toad_> hobx_: hrrm
[8:10] <toad_> why?
[8:10] <toad_> I mean neural nets, not necessarily in their current form, will be useful for simulating neurons, right?
[8:11] <hobx_> I don't know. I don't think so.
[8:11] <toad_> obviously not backprop nets
[8:11] <toad_> hobx_: what is to stop us simulating the brain cell by cell?
[8:11] <toad_> it might take a century to know enough and have the computing power.. but?
[8:11] <iip_i2p> <Newsbyte2> the latest progress I'm aware of, included small animals that were frozen solid and revived. The progress was made because they let them 'underchill' (or what's the word in english) first, before freezing them.
[8:11] <hobx_> I think intelligence is fundamentally tied to human consciousness, which can't be replicated deterministically.
[8:12] <toad_> hobx_: it uses atoms and quarks and molecules
[8:12] <hobx_> The simple fact is that my brain can't be simulated by code, because I'm hear and I'm not in the code.
[8:12] <toad_> it can be simulated by atoms and quarks and molecules
[8:12] <toad_> unless you argue that it is a protrusion of divinity
[8:12] <hobx_> toad_: Yes. But not in software.
[8:13] <toad_> hobx_: not even in quantum software ?
[8:13] <hobx_> No, not divine as such.
[8:13] <toad_> in the rest of the body you can build reasonable sims on a cell by cell level
[8:13] <hobx_> toad_: Not sure. I don't know enough about quantum computing etc.
[8:13] <toad_> there's early work on this
[8:14] <iip_i2p> <Newsbyte2> toad: read about the bussard-thingy?
[8:14] <toad_> but if there are wierd quantum effects in the brain, they can probably be simulated now, it'd just take crazy computing power
[8:14] <toad_> certainly with quantum computers... anyway
[8:14] <toad_> Newsbyte2: hmm?
[8:14] <toad_> Newsbyte2: the page said nobody had solved the engineering challenges
[8:14] <iip_i2p> <Newsbyte2> what?
[8:14] <toad_> Newsbyte2: not that somebody had shown it's extremely implausible
[8:15] <toad_> Newsbyte2: what have you read, exactly?
[8:15] <iip_i2p> <Newsbyte2> ermm...I already told you, twice
[8:15] <toad_> sorry tell me again :)
[8:16] <iip_i2p> <Newsbyte2> they calculated that, when the density of hydrogen atoms becomes dense enough to become usefull for the combustion, it is also dense enough to cause a severe drain on the captivator, reducing the forward motion to nothing.
[8:16] <toad_> what do you mean?
[8:17] <toad_> you capture stuff, and shoot it out the back, at exactly equal rate, right?
[8:17] <iip_i2p> <Newsbyte2> what do you mean, what do I mean? that's what they said. :-)
[8:17] <toad_> what exactly causes the drain?
[8:18] <iip_i2p> <Newsbyte2> before they can capture it, there must be enough to capture, when there is, it also cause a drain..o wait
[8:18] <toad_> when you shoot it out the back it has a lot more energy
[8:18] <iip_i2p> <Newsbyte2> my english :-)
[8:18] <toad_> because you have a fusion reactor going on in the middle... ah
[8:18] <toad_> shit
[8:19] <toad_> I see
[8:19] <hobx_> I don't
[8:19] <toad_> fusion does mass to energy
[8:19] <toad_> so you lose some mass in the middle
[8:19] <iip_i2p> <Newsbyte2> i think the correct word is: friction ;-)
[8:19] <toad_> do you lose enough mass in the middle to make up for it?
[8:19] <toad_> i mean to make what goes out the back less than what comes in?
[8:19] <toad_> Newsbyte2: ahhhh
[8:20] <toad_> Newsbyte2: so it's a matter of collection being difficult?
[8:20] <iip_i2p> <Newsbyte2> it's not about the conversion, it's about the friction
[8:20] <toad_> Newsbyte2: so the problem is that some of the matter will bounce off and not be captured?
[8:21] <toad_> but will still slow us down?
[8:21] <iip_i2p> <Newsbyte2> indeed. first, they recalculated the original diameter necessary for a bussard
[8:21] <hobx_> Because certainly every hydrogen atom releases more energy from fusion then is needed to accelerate it when collected.
[8:21] <toad_> how much of a problem is it that some matter isn'
[8:21] <toad_> t captured?
[8:21] <iip_i2p> <Newsbyte2> it showed that it would need to be ten times wider then originally envisaged, if you wanted to fly with it through 'empty' space as well
[8:21] <toad_> how much can be reasonably captured?
[8:22] <iip_i2p> <Newsbyte2> noooo
[8:22] <hobx_> Don't these sound like practical problems?
[8:22] <toad_> as a proportion of what enters?
[8:22] <iip_i2p> <Newsbyte2> it's not what doesn't get captured, it is what IS being captured
[8:22] <toad_> Newsbyte2: eh?
[8:22] <toad_> how can what is captured be a problem?
[8:23] <toad_> I mean, if you can fuse most of it, and then accelerate it to 99% of the speed of light on the way out, what's the problem?
[8:23] <hobx_> The energy from each hydrogen atom would be equal to <energy released from fusion> - <energy needed to accelerate atom>
[8:23] <iip_i2p> <Newsbyte2> what the 'mouth' captures, is tunneled inwards, to the engine, this causes friction; they calculated that, when enough density is met to make the engine work, it is already causing a lot more friction
[8:24] <hobx_> At some velocity the later has been greater than the first
[8:24] <toad_> Newsbyte2: I still don't get it
[8:24] <iip_i2p> <Newsbyte2> what, exactly, don't you get?
[8:24] <toad_> why friction is a problem at all
[8:24] <toad_> are you saying friction is so high that the particles don't actually reach the reactor?
[8:25] <toad_> also what materials are they assuming?
[8:26] <iip_i2p> <Newsbyte2> no, I'm saying that, when enough particals reach the engine so that it could start working, the capturing of those particles make so much friction, that the forward movement is null
[8:26] <toad_> Newsbyte2: ummm
[8:26] <toad_> Newsbyte2: of course we lose momentum in capture
[8:27] <hobx_> it is very odd to talk about friction on a particle level people
[8:27] <toad_> the point is, we gain momentum when we shoot them out the other end
[8:27] <toad_> hobx_: very
[8:27] <hobx_> Particles don't have friction
[8:27] <toad_> hobx_: considering we don't really know what friction IS
[8:27] <hobx_> That was why I was trying to talk about relative kinetic energy
[8:27] <toad_> and we gain a lot of momentum when we shoot them out the back
[8:27] <iip_i2p> <Newsbyte2> when they are catured, then they have
[8:27] <toad_> because we have some fusion going on in the middle
[8:27] <hobx_> toad_: Really? Isn't it just elecrtostatic forces between molecules?
[8:27] <toad_> they leave faster than they enter
[8:28] <toad_> hobx_: nope, I read some cool work on it in Scientific American ages ago
[8:28] <toad_> Newsbyte2: if you can find the original paper, mail me a scan
[8:29] <toad_> otherwise, I'll continue to believe that it might be possible :)
[8:30] * toad_ wonders if inertia modification is possible... there's some nice experimental results, just no theoretical basis :)... you'd still need a hell of a lot of energy so i suppose it doesn't make much difference in practice
[8:30] <toad_> anyway bbl
[8:30] <iip_i2p> <Newsbyte2> look at it this way, toad: bussards need a 'sea' of fuel to wade through, to get forward motion. It has a 'mouth' that captures (some of) the water it wades through; that causes a drag/friction; the faster they will go, the more friction. they calculated that, in order to gain a seizable fraction of the speed of light, the friction would outweigh the forward movement by a large factor.
[8:30] <toad_> Newsbyte2: I agree that the friction would be significant
[8:30] <iip_i2p> <Newsbyte2> it was a book about spacetravel. I think it was one of my local library.
[8:31] <toad_> but if as you say it's all about stuff that DOES get captured, then I don't see why the exhaust can't be even bigger
[8:31] <toad_> if it's about stuff that bounces off then maybe i can see your point
[8:31] <toad_> you could argue that only 1% of incoming matter gets captured
[8:31] <iip_i2p> <Newsbyte2> ermm..you DO realise you need a minimal density?
[8:31] <toad_> in which case you're stuffed
[8:32] <toad_> :)
[8:32] <iip_i2p> <Newsbyte2> if you use a bigger exhaust, you need more fuel, thus, you need to capture more, thus your mouth has to be bigger, thus the drag becomes worse
[8:32] <Redb3ard> ah
[8:33] <toad_> Newsbyte2: I can't see how the drag on a hydrogen atom can be higher than the massive amount of energy released on its fusion
[8:33] <Redb3ard> so all you need is an infinite-sized exhaust, with infinite fuel
[8:33] <hobx_> One could argue that it would happen if the craft ends up pushing a cloud of hydrogen ahead of it.
[8:33] <hobx_> But I doubt that would happen.
[8:33] <iip_i2p> <Newsbyte2> toad...because you need a MASSIVE amount of hydrogens for the engine to work
[8:33] <toad_> so?
[8:34] <iip_i2p> <Newsbyte2> what, so? So, it creates a massive drag
[8:34] <hobx_> You people aren't arguing logically.
[8:34] <toad_> if you fuse all of them, then expel all of them (on average), then what's the problem? I don't see why n billion moles of hydrogen atoms can exert more friction than the KE they'd gain from being fused
[8:34] <hobx_> Stop and see what you agree on.
[8:35] <toad_> well, I've got to go get breakfast
[8:35] <toad_> after that I'm going out
[8:35] <Redb3ard> dude, what are we trying to accomplish here?
[8:35] <toad_> then when I get back I'll work on Freenet :)
[8:35] <Redb3ard> i mean, if you already have fusion...
[8:35] <hobx_> 1) Drag caused by the hydrogen collected is because the hydrogen has mean velocity 0 compared to the ship, and needs to be accelerated when it enters the collector?
[8:35] <hobx_> yes+
[8:35] <iip_i2p> <Newsbyte2> toad, I'm not getting this. It's not that complicated to grasp
[8:35] <Redb3ard> and its not enough energy for you, then step it up to M/AM
[8:35] <toad_> hobx_: I thought it was caused on entry?
[8:36] <Redb3ard> and if you need reaction mass... scooping it up as you go along just wont cut it
[8:36] <iip_i2p> <Newsbyte2> the drag you experience, comes from the hydrogen captivated
[8:36] <toad_> Redb3ard: even with antimatter annihilation it still takes ages to get anywhere, and you need to carry a LOT of fuel
[8:36] <hobx_> toad_: In what way?
[8:36] <Redb3ard> no, the fuel is M and AM
[8:36] <toad_> Newsbyte2: so is it incurred in the process of capture, or the process of expulsion?
[8:36] <toad_> Redb3ard: http://members.nova.org/~sol/station/interste.htm
[8:36] <Redb3ard> reaction mass is what you shoot out the ass of the ship, to make use of the energy
[8:37] <iip_i2p> <Newsbyte2> captured
[8:37] <toad_> Redb3ard: if you can get 4Msec isp from am, it'll still take 40 years to get to proxima, with 97% of your mass as fuel (which probably isn't possible with antimatter)
[8:37] <hobx_> What do you shoot out the ass of an antimatter ship?
[8:37] <Redb3ard> there are workable fission plans, that could put a probe at proxima in what, 90 years?
[8:37] <hobx_> Because when you put M and AM together they annihilate right?
[8:37] <toad_> Redb3ard: there are?
[8:38] <Redb3ard> yeh
[8:38] <toad_> not in 90 years...
[8:38] <Redb3ard> of course, assembling 600,000 50kiloton nukes in orbit would have the UN in a tizzy
[8:38] <toad_> sadly i can't find my original source, it was on the Breakthrough Physics Program, explained...
[8:38] <toad_> Redb3ard: ahh, orion drive
[8:38] <toad_> i see
[8:38] <toad_> that's cheating :)
[8:38] <Redb3ard> 300,000 of them for acceleration
[8:38] <toad_> but it counts as energy beaming
[8:39] <toad_> hobx_: umm, you were saying?
[8:39] <Redb3ard> of course, 180 years before we get to see the cool hubble-telescope-like pictures
[8:39] <Redb3ard> bah
[8:39] <Redb3ard> 94, duh
[8:39] <Redb3ard> still, i wont be alive
[8:39] <hobx_> toad_: About what?
[8:39] <toad_> hobx_: what do we have in common?
[8:39] <toad_> re bussard ramjets?
[8:39] <Redb3ard> only 180, if the thing returns with film, haha
[8:39] <iip_i2p> <Newsbyte2> heh
[8:40] <hobx_> toad_: NM. You are ten minutes lagged again. I meant that you and newsbyte weren't reasoning, you were just repeating the same things.
[8:40] <toad_> 1. friction happens because the matter enters, its velocity goes to zero.
[8:40] <hobx_> relative to the ship?
[8:40] <toad_> yup
[8:40] <hobx_> yeah
[8:41] <toad_> 2. we gain momentum by fusing the matter and shooting it out the back at the entry rate.
[8:41] <toad_> Newsbyte2: agreed?
[8:41] <hobx_> yupp
[8:41] <Redb3ard> yeh, but the numbers arent just there
[8:41] <iip_i2p> <Newsbyte2> indeed
[8:41] <Redb3ard> maybe if the universe were filled with slush-solid H
[8:41] <iip_i2p> <Newsbyte2> that's the general theory
[8:41] <toad_> well it's about the numbers
[8:41] * Hory (~Miranda@81.196.25.110) Quit (Read error: 54 (Connection reset by peer))
[8:42] <toad_> can the friction outweigh the fusion?
[8:42] <toad_> Newsbyte2 has apparently read that it can
[8:42] <Redb3ard> besides, if you have fusion already, then taking some H with you isnt such a big deal
[8:42] <toad_> perhaps that is right, but I'd need to read it from somebody who knows, with the maths
[8:42] <toad_> Redb3ard: no, it is
[8:42] <toad_> Redb3ard: it'll take centuries
[8:42] <toad_> read http://members.nova.org/~sol/station/interste.htm
[8:42] <hobx_> bbl
[8:43] <toad_> Redb3ard: 97% mass as fuel, 40 year trip to alpha centauri, needs 4Msec isp
[8:43] <iip_i2p> <Newsbyte2> it is, in those area's that have enough hydrogen per sqaure mile to be able to provide enough density that the engine might work (fusing)
[8:43] <Redb3ard> i think oneway trips to the near stars of 100-200 years is thoroughly feasible
[8:43] <toad_> 4Msec ISP can only be achieved by antimatter
[8:43] <Redb3ard> and at that, taking fuel with you isnt any big deal
[8:43] <toad_> Redb3ard: well, if you believe that, that's your problem. NASA doesn't.
[8:43] <Redb3ard> dude, 40 years?
[8:43] <Redb3ard> why not 200
[8:44] <iip_i2p> <Newsbyte2> I'm not sure enmore, but I thought the original 'mouth' was supposed to be 2 miles, is that right, toad?
[8:44] <Redb3ard> that drops it quite a bit
[8:44] <Redb3ard> 40 years triptime is "lets impress congress" numbers
[8:44] <Redb3ard> and antimatter is only a energy storage solution
[8:44] <Redb3ard> not energy production
[8:45] <Redb3ard> what, we might be able to manufacture micrgrams of it someday?
[8:45] <Redb3ard> we need some sort of anti-matter breeder reaction
[8:45] <toad_> |Imagine using conventional chemical rockets for the journey and we want to carry a payload in the order of size of a bus. Even if we say that we will not need to slow down and allow 900 years to make the journey the propellant required would still be greater than the entire mass in the universe! Clearly the amount of propellant required is an extremely important factor.
[8:46] <Redb3ard> so far, my attempts to modify matter symetry with my mind powers have been failures
[8:46] <Redb3ard> well, other than accidentally setting the cat on fire
[8:46] <toad_> argh, it doesn't do the math for anything else
[8:46] * greycat hisses
[8:46] <Redb3ard> pyrokinesis is both a blessing and a curse
[8:47] <Redb3ard> 40 years is way too optimistic
[8:47] <Redb3ard> start setting it back a bit, though less than 1000 years, and fusion becomes doable
[8:47] <Redb3ard> though hell, we cant even do fusion, really
[8:47] <Redb3ard> you think a production level tokamak is something you cna put on an unmanned spaceship?
[8:47] <toad_> http://www.grc.nasa.gov/WWW/PAO/images/warp/warp06.gif
[8:48] <toad_> that's what i was thinking of
[8:48] * fqdfqa (~bla@195.25.40.253) has joined #freenet
[8:48] <toad_> canister 10 tons
[8:48] <fqdfqa> hi everybody
[8:48] <Redb3ard> ESC fusion is the only real way to fuse
[8:48] <Redb3ard> 10 tons seems reasonable
[8:48] <iip_i2p> <Newsbyte2> An optimistic estimate would be that only 1% of the hydrogen would be actually usable as fuel. So in fact much of the propulsive power would be used up slogging through a soup of useless hydrogen.
[8:48] <Redb3ard> if im sending a probe to another star for the first time, i want to cram everything on it i can think of
[8:49] <Redb3ard> with secondary and tertiary backups
[8:49] <toad_> http://www.grc.nasa.gov/WWW/PAO/html/warp/scales.htm
[8:49] <toad_> Redb3ard: that's a one way trip though
[8:49] <toad_> it doesn't stop
[8:49] <toad_> it takes a lot more fuel to stop
[8:49] <toad_> which means you have to have more fuel to propel the first lot :)
[8:49] <toad_> anyway bbl
[8:51] <fqdfqa> I've got that : Conections open (Inbound/Outbound/Limit) 63 (0/63/200)
[8:51] <fqdfqa> is it normal ?
[8:51] <fqdfqa> to have 0 in inbound ?
[8:51] <greycat> sounds like your NAT is interfering with inbound connections
[8:51] <fqdfqa> yop but frost is working good
[8:51] <iip_i2p> <Newsbyte2> yes we can, we already made fusion for a picosecond, or something
[8:51] <fqdfqa> and freesites too
[8:52] <Redb3ard> bah
[8:52] <greycat> having just outbound connections can work out, you just take a bit longer to build up your connections
[8:52] <Redb3ard> everytime someone smacks their hands together, its possible that 2 H atoms will fuse
[8:52] <Redb3ard> thats not the same thing as fusion
[8:52] <Redb3ard> we're talking sustainable fusion here, and to date, no one has made any real progress past what we could do in the 60s
[8:52] <iip_i2p> <Newsbyte2> me has to repeat what he said :-)
[8:52] <toad_> Redb3ard: so we can do it in 900 years with m/am annihilation... if we don't want to stop. that's not bad.
[8:53] <toad_> :)
[8:53] <toad_> Redb3ard: really? they CLAIM they have...
[8:53] <Redb3ard> toad, until we have m/am breeder reactors...
[8:53] <Redb3ard> toad, read through it
[8:53] <Redb3ard> even the tokamak is a joke
[8:53] <toad_> Redb3ard: umm, right now, it costs a hell of a lot more energy to make am than you get when you burn it
[8:53] <toad_> Redb3ard: they can make it self-sustaining
[8:53] <Redb3ard> though the neutron count is *alot* higher than the first fusion experiments...
[8:53] <toad_> did it a few years ago
[8:54] <Redb3ard> they're just using more electricity to even get that
[8:54] <toad_> it's not hot enough to produce more energy than it takes to start it
[8:54] <Redb3ard> not all that close to breakeven at all
[8:54] <Redb3ard> screw ignition
[8:54] * toad_ prefers the apple topology
[8:54] <Redb3ard> we'll fire up coal plants, if thats what it takes to start a fusion reactor
[8:54] <Redb3ard> breakeven energies are the key
[8:54] <iip_i2p> <Newsbyte2> yesyes, but it WAS fusion, red, even if it wasn't yet economical viable :-)
[8:54] <toad_> which is probably going to be used in the latest incarnation of ITER
[8:55] <toad_> anyway fusion sucks, terrestrially
[8:55] <Redb3ard> iip, you're using terms that confuse
[8:55] <toad_> just build a decent storage system, and a skyhook, and use solar
[8:55] <Redb3ard> any fusion past breakeven is economically viable, id think
[8:55] <iip_i2p> <Newsbyte2> iter makes a good chance of actually making it self-sustaining
[8:55] <toad_> even if you don't have a skyhook, if you can store it, wind will provide sufficient power for the UK and the USA at least
[8:55] <toad_> :)
[8:55] <Redb3ard> breakeven is where its putting out enough heat to sustain the reaction
[8:56] <Redb3ard> ignition, is the energy it takes to start it
[8:56] <toad_> but fusion would be incredibly useful in space
[8:56] <toad_> like fission
[8:56] <iip_i2p> <Newsbyte2> at least, if they aren't going to be cheap about it
[8:56] <Redb3ard> breakeven isnt that much closer today, than it was 10 years ago
[8:56] <toad_> now i've really gotta go :)
[8:56] <Redb3ard> ouch
[8:56] <Redb3ard> skyhook?
[8:56] <iip_i2p> <Newsbyte2> has it been decided yet where they are going to put it? Last time I checked, they were debating France or Japan
[8:57] <Redb3ard> like i want them beaming gigawatt microwaves down to earth
[8:57] <Redb3ard> especially with a home simpson manning the recieving station
[8:57] <Redb3ard> iip, the big fusion plant is scary
[8:57] <Redb3ard> they need 2 meters of liquid lithium shielding
[8:58] <Redb3ard> its unclear whether theyll go much past breakeven
[8:58] <iip_i2p> <Newsbyte2> that's a beloved SF-thingy and dreamstuff of unrealistic ninkenpoops. ppl will not appraciate microwave-sats above their heads
[8:58] <iip_i2p> <Newsbyte2> has it been decided yet where they are going to put it? Last time I checked, they were debating France or Japan <--- ITER, I mean
[8:58] <Redb3ard> when we have the space elevator, and they can transfer the electrical power down with wires, ill think about it
[8:58] <Redb3ard> france
[8:58] <Redb3ard> all the EU people are getting snooty
[8:58] <hobx_> The problem is you can't really place somewhere remote without loosing a lot of power to transferal.
[8:59] <toad_> hobx_: that's not a big problem
[8:59] <greycat> we just need superconductors, then.
[8:59] <toad_> superconducting wires are only 3-5x more expensive than normal ones
[8:59] <iip_i2p> <Newsbyte2> snooty? We've lot's of catching up to do with the USA, still, rest assured, red ;-)
[8:59] <toad_> they're being laid
[8:59] <toad_> right now
[8:59] <hobx_> toad_: But nobody uses superconducting for power transferal currently, right?
[8:59] <iip_i2p> <Newsbyte2> in being snooty, that is ;-p
[8:59] <toad_> they have much lower losses (no losses on DC, but less losses on AC)... either way they need some power to make nitro of course
[8:59] <toad_> hobx_: wrong, they do
[9:00] <toad_> it's being deployed
[9:00] <hobx_> really? /me curious
[9:00] <toad_> in domestic areas
[9:00] <toad_> somewhere in the US
[9:00] <toad_> also some japanese i think
[9:00] <toad_> you need a coolant generator every 500 meters iirc
[9:00] <toad_> that puts the cost up a bit
[9:00] <hobx_> I made a ytrium barium copper oxide superconductor in middle school. It didn't work.
[9:01] <toad_> but it's practical, it carries a hell of a lot of current, and it's rapidly getting cheaper
[9:01] <iip_i2p> <Newsbyte2> take an example at that, toad! Even superconducting wires are being laid!
[9:01] <iip_i2p> <Newsbyte2> lol
[9:01] * Troublemaker (egal@pD9E3A6B7.dip.t-dialin.net) has joined #freenet
[9:01] <Redb3ard> you guys are missing the point
[9:01] <toad_> much higher power density
[9:01] <Redb3ard> having the fusion reactor remote makes no difference...
[9:01] <Redb3ard> there may not be all that much power to transfer from it
[9:01] <hobx_> Byt skyhook remote thus...
[9:02] <toad_> Redb3ard: we were talking about downlink from satellite, actually
[9:02] <Redb3ard> they project power output just above breakeven
[9:02] <Redb3ard> oh
[9:02] <Redb3ard> power from a satellite is bad business
[9:02] <toad_> IMHO we won't be able to do it with wires
[9:02] <toad_> not for a loooong time
[9:02] <toad_> Redb3ard: why?
[9:02] <hobx_> Power isn't a big problem.
[9:02] <iip_i2p> <Newsbyte2> I concur. At least in the USA they have been being laid for some years, now.
[9:02] <Redb3ard> evne a space elevator doesnt really fix it
[9:02] <hobx_> Fission is here, safe, and works.
[9:02] <toad_> Redb3ard: if you have a space elevator, it's cheap
[9:02] <Redb3ard> 22,000 miles to geosync
[9:02] <toad_> ROFL hobx_
[9:03] <Redb3ard> toad, figure copper wire 1/8" DIA
[9:03] <Redb3ard> 22,000 miles of it
[9:03] <Redb3ard> its too heavy
[9:03] <toad_> Redb3ard: exactly
[9:03] <toad_> it's FAR too heavy
[9:03] <toad_> you have to beam it
[9:03] <Redb3ard> and 1/8" cant transfer all that much juice anyway
[9:03] <toad_> and that doesn't scare me more than fission
[9:03] <toad_> and it's certainly better than oil!
[9:04] * Troublemaker (egal@pD9E3A6B7.dip.t-dialin.net) Quit ("L'amore ? pi? pericoloso che tutta la vita")
[9:04] <Redb3ard> ok, lets forget the fact that if its knocked slightly out of alignment, you slag an elementary school 17 miles away from the recieving station
[9:04] <toad_> but as i said, wind is sufficient if you can store electricity on a decent scale
[9:04] <hobx_> superstition and propaganda.
[9:04] <hobx_> That is the only problem with fission.
[9:04] * fqdfqa (~bla@195.25.40.253) has left #freenet
[9:04] <Redb3ard> what do you think that will do to local weather patterns?
[9:04] <toad_> lol hobx_ , yeah, whatever you say
[9:04] <Redb3ard> wind, same problem
[9:04] <Redb3ard> it doesnt scale
[9:04] * toad_ not getting into THAT one
[9:04] <toad_> Redb3ard: it doesn't?
[9:04] <hobx_> http://argument.independent.co.uk/commentators/story.jsp?story=524230
[9:04] <Redb3ard> put up windmills everywhere, and thats just a different flavor of ecological disaster
[9:05] <toad_> Redb3ard: it takes a very small amount of energy out of the system
[9:05] <toad_> that's like saying solar has ecological impact
[9:05] <hobx_> Whatever you want, silly hippy.
[9:05] <toad_> it takes a tiny amount, even within the turbine
[9:05] <Redb3ard> dude, hydroelectric only takes a small amount of energy out of the system
[9:05] <toad_> Redb3ard: hydo has obvious problems
[9:05] <Redb3ard> and it still manages to ruin alot of things
[9:05] <toad_> yes
[9:05] <Redb3ard> well, then wind has non-obvious problems
[9:05] <toad_> because of the fact that you have to build a dam
[9:05] <iip_i2p> <Newsbyte2> well...I have my doubts about the optimistic greenies too. ,I actually don't think 'green' energyproduction is going to help much to elevate the energyburden in the short and mid-long term
[9:05] <Redb3ard> give me fission
[9:06] <hobx_> But damns are pretty cool.
[9:06] <Redb3ard> give me fusion
[9:06] <toad_> and completely change a large part of the landscape
[9:06] <hobx_> damms
[9:06] <hobx_> whatever
[9:06] <Redb3ard> anything else is hamsters on a treadmill
[9:06] <hobx_> fuck the landscape
[9:06] <hobx_> the landscape is fucking luxury
[9:06] <toad_> Redb3ard: wind is almost economical now
[9:06] <toad_> if we could store it it will be economical
[9:06] <toad_> probably without subsidy
[9:06] <toad_> fusion is a pipe dream
[9:07] <iip_i2p> <Newsbyte2> I mean, geez. In my country, they took ages, dozens of complaints, huge amounts of money, to build some windmills...that barely represent 0.2% of energy production
[9:07] <Redb3ard> energy storage alone, would do alot to alleviate problems in our grid
[9:07] <toad_> and fission is not only unsafe, it's politically infeasible
[9:07] <Redb3ard> fission isnt unsafe at all
[9:07] <iip_i2p> <Newsbyte2> I just don't see windturbines as a viable replacement, at least here
[9:07] <Redb3ard> not even close
[9:07] <Redb3ard> the 1950s reactor designs are unsafe
[9:07] <Redb3ard> which is all that was ever built
[9:07] <toad_> Redb3ard: I'll remind you of that when terrorists blow up hansard :)
[9:07] <toad_> or sellafield
[9:07] <Redb3ard> the new IFR designs, even blowing it up wouldnt do alot
[9:08] <toad_> yeah, that's what they said last time
[9:08] <toad_> but i'm not interested
[9:08] <Redb3ard> they cant aersolize it without the water coolant systems
[9:08] <toad_> I'm going to get some food
[9:08] <Redb3ard> then you arent intersted in technology
[9:08] <toad_> then I'll come back and do some actual WORK
[9:08] <toad_> Redb3ard: no
[9:08] <iip_i2p> <Newsbyte2> I think cheap solarcells would do it better then windturbines, even
[9:08] <Redb3ard> wind, solar....
[9:09] <toad_> I'm not interested in any energy source that costs more than renewables do when you include decomissioning
[9:09] <Redb3ard> both cause unforseen problems
[9:09] <toad_> AND that produces PLUTONIUM
[9:09] <toad_> which produces clearly foreseen problems
[9:09] <Redb3ard> IFR doesnt produce weapons grade fissile material
[9:09] <toad_> Redb3ard: you can blow up reactor grade uranium
[9:09] <iip_i2p> <Newsbyte2> let's not exagerate
[9:09] <Redb3ard> actually, it takes it in the opposite direction, makes it even harder to process
[9:09] <toad_> Redb3ard: well, fortunately, I have the public on my side
[9:10] <Redb3ard> again, IFR cant produce weapons grade stuff
[9:10] <toad_> :)
[9:10] <toad_> so I can just ignore you :)
[9:10] <toad_> bye
[9:10] <Redb3ard> haha
[9:10] <Redb3ard> yes, having the public on your side, thats a good thing
[9:10] <iip_i2p> <Newsbyte2> once upon a time, I was all against nuclear energy too. But scientific research has shown it isn't all that bad, compered to other energysources, whetever the greens may say.
[9:11] <toad_> Newsbyte2: yeah, like all the scientific research in the 60s, 70s, 80s... and the work that showed lead was harmless to kids.. etc etc
[9:11] <iip_i2p> <Newsbyte2> indeed, red. Especially when there is a mob, ready to lynch somebody ;-)
[9:11] <iip_i2p> <Newsbyte2> then it's ESPECIALLY good if they are on your side :-))
[9:11] <Redb3ard> the new fission designs are about as safe as anything gets
[9:12] <Redb3ard> they could produce alot of energy, for a long time, before we'd even have to mine more fuel
[9:12] <Redb3ard> and if you're worried about something blowing them up, guard it out the wazoo
[9:12] <iip_i2p> <Newsbyte2> I would like to remind you, toad, that while resaerch is sometimes unperfect, it ,was equally science that showed lead etc. could be dangerous
[9:12] <Redb3ard> have F-16s in the air, ready to down anything that comes close
[9:13] <hobx_> anyone have a clever name for a castle with a lot of towers?
[9:13] <greycat> "the castle"
[9:13] <Redb3ard> but instead, we'll put windmills everywhere, watching all sorts of borderline endangered birds be chopped into giblets
[9:13] <iip_i2p> <Newsbyte2> towermansion?
[9:13] <hobx_> clever
[9:13] <Redb3ard> watch all sorts of weather anomalies develop
[9:14] <Redb3ard> and in 50 years, the eco-hippies will be crying again
[9:14] * Hory (~Miranda@81.196.25.110) has joined #FreeNet
[9:14] <iip_i2p> <Newsbyte2> I feel I have to agree with you, red. Strange. Are you sure you're from the USA?
[9:14] <Redb3ard> not by choice
[9:14] <hobx_> redb3ard is right. Only nuclear energy is ecological.
[9:15] <Redb3ard> you act like everyone moved here as an adult
[9:15] <iip_i2p> <Newsbyte2> well...let's make that fusion
[9:15] <Redb3ard> nuclear energy can be dirty too, if people do dumb things
[9:15] <hobx_> I would move to america as an adult.
[9:15] <Redb3ard> like try to build 10,000 warheads, or build reactors with 1950s tech
[9:16] <iip_i2p> <Newsbyte2> there ARE some probs with nuclear energy today, after all. even if greens exagerate it.
[9:16] <Redb3ard> but all this radiation people rally against... where do they think it comes from?
[9:16] <iip_i2p> <Newsbyte2> hobx; you mean you're not an adult?
[9:16] <hobx_> That's like saying bathrooms are dirty because you can crap on the floor.
[9:16] <Redb3ard> once we mind the fuel, the radiation fairy sprinkles it, and it starts emitting harmful invisible rays?
[9:16] <iip_i2p> <Newsbyte2> that would explain why you would wanna go ;-)
[9:17] <hobx_> Newsbyte: No, I mean that other circumstances have kept me from moving yet.
[9:17] <Redb3ard> the act of using nuclear energy just concentrates what was already there
[9:17] <hobx_> Whether I'm adult or not is subject to debate.
[9:17] <iip_i2p> <Newsbyte2> pfff...each his choice. I couldn't be moved to go to the US even if they presented me a green card and a mercedes on top
[9:18] <hobx_> America has it's issues. Name the country that doesn't.
[9:18] <hobx_> I would rather live there than in this hellhole, that is for sure.
[9:19] <Redb3ard> whats so bad about sweden?
[9:19] <hobx_> Um, shit. where to start.
[9:21] <iip_i2p> <Newsbyte> ok: andorra
[9:22] <hobx_> Sweden is only good for moochers and leechers looking to live off the backs of others.
[9:22] <Redb3ard> cool
[9:22] <Redb3ard> i think ill go there
[9:22] <hobx_> The whole country is just a reenactment of "Atlas Shrugged"
[9:22] <hobx_> Redb3ard: If that is your thing, I recommend it.
[9:23] <hobx_> In Sweden, you can get sick leave from your paternity leave from your unemployment and have all three paid by everybody else.
[9:24] * hirvox (~hirvox@cs181027153.pp.htv.fi) Quit (Connection timed out)
[9:25] <iip_i2p> <Newsbyte> ah, I see.hobx is like nobody :-)
[9:25] <Redb3ard> actually, id be happy with a job
[9:26] <hobx_> like nobody?
[9:26] <iip_i2p> <Newsbyte> indeed
[9:27] <hobx_> which means?
[9:27] * greycat is also confused by that comment
[9:27] <iip_i2p> <Newsbyte> what? with 'sanity' at least some understood it was a handle... ;-)
[9:28] <greycat> 09:46 There is no such nick nobody
[9:28] <hobx_> I think Newsbyte lost it.
[9:28] <iip_i2p> <Newsbyte> did I say he was here, no?
[9:28] <iip_i2p> <Newsbyte> now?
[9:29] <toad_> <Redb3ard> but instead, we'll put windmills everywhere, watching all sorts of borderline endangered birds be chopped into giblets - I still can't believe that anyone could believe that windmills are harmful and nuclear reactors aren't. it just seems hilarious to me :)
[9:29] * theArthur (~arthur@unix1.ic.ucsc.edu) has joined #freenet
[9:29] <iip_i2p> <Newsbyte> I'm making perfectly sense. Seen your comment about sweden, you are like 'nobody'. He always says the same things too.
[9:30] <toad_> <Redb3ard> and in 50 years, the eco-hippies will be crying again - no big deal, the immediate problem is that 99% of our energy comes from fossil fuels, deal with that first, then worry about the implausible consequences of taking a tiny amount of energy out of the winds
[9:30] <greycat> has anyone but Newsbyte ever seen this alleged "person"?
[9:31] <hobx_> toad_: Is that 99% a POYA statistic?
[9:31] <theArthur> i only ever get a connection to a single node and its always the same node? could sombody point me to a seednode that works?
[9:31] <hobx_> toad_: We have a way to deal with it. It is called fission.
[9:31] <hobx_> http://argument.independent.co.uk/commentators/story.jsp?story=524230
[9:31] <toad_> <Redb3ard> whats so bad about sweden? - not a lot, from what I've heard :)
[9:31] <toad_> cheap bandwidth!
[9:31] <iip_i2p> <Newsbyte> toad, dude, you sound like me, now, years ago, when I thought the greens told nothing but the truth :-)
[9:32] <toad_> right, hobx_'s problem is he's a right wing libertarian
[9:32] * theArthur looks at the topic for this channel???
[9:32] <toad_> if so, he's better off in the US
[9:32] <greycat> theArthur: what build number?
[9:32] <theArthur> 5090
[9:32] <iip_i2p> <Newsbyte> fact is, nuclear energy has it's drawbacks, but has also a lot of benefits. All by all, when looked at objectively, it's not a bad choice, for the near and mid-term future.
[9:33] <theArthur> i think that is the new one
[9:33] <greycat> theArthur: and the seednodes from freenetproject.org/snapshots/ didn't work?
[9:33] <hobx_> toad_'s problem is that he is a leftist moocher who believes in living off the backs of others.
[9:33] <theArthur> it seems so
[9:33] <theArthur> could sombody give me a seed to there node?
[9:33] <toad_> Newsbyte2, hobx_, toad_, please take it to #freenet-chat or #freenet-politics - op bit or not, we've got a real user here :)
[9:34] <theArthur> and what is wrong with the seeds from the project?
[9:34] <iip_i2p> <Newsbyte> true, true, but then again, you seem an elitist capitalistic pig ;-p
[9:34] <toad_> theArthur: stable or unstable? what happens with the official seeds?
[9:34] <theArthur> stable?
[9:34] <theArthur> should i get unstable?
[9:34] <toad_> theArthur: no, stable
[9:34] <theArthur> what is the difference
[9:34] <theArthur> (other than one being more "stable")
[9:34] <toad_> what happens if you try to reseed with the current stable seednodes?
[9:34] <iip_i2p> <Newsbyte> ermm...if I'm not mistaken, you talked a lot of chitchat here too, toad
[9:35] <toad_> Newsbyte: sure, but I've stopped now
[9:35] <toad_> and I have an op bit too
[9:35] <toad_> and we have a "customer" :)
[9:35] <iip_i2p> <Newsbyte> besides, the chan is dead otherwise, anyway
[9:35] <iip_i2p> <Newsbyte> me shouts:ppl have any probs with freenet they wanna discuss right now?!
[9:35] <toad_> theArthur: well, there are 638 nodes in the seednodes
[9:36] * tyesinclair (~tyesincla@170.252.64.1) has joined #freenet
[9:36] <toad_> theArthur: what happens when you reseed? do you get the OOMs?
[9:36] <greycat> theArthur: linux or windows?
[9:36] <theArthur> if i restart freenet will it reseed?
[9:36] <theArthur> linux
[9:36] * tyesinclair (~tyesincla@170.252.64.1) has left #freenet
[9:36] <iip_i2p> <Newsbyte> ofcourse, arthur must be the spoil :-/
[9:36] <toad_> theArthur: if you stop freenet, touch seednodes.ref, start freenet
[9:36] <greycat> theArthur: if you update seednodes.ref and restart, yes.
[9:36] <iip_i2p> <Newsbyte> ;-)
[9:36] <toad_> update it is even better
[9:36] <toad_> ls -l seednodes.ref ?
[9:37] <theArthur> i was thinking that seeds were chosen randomly from the seednodes when started
[9:37] <toad_> theArthur: they are, when you reseed
[9:37] <toad_> not otherwise
[9:37] <iip_i2p> <Newsbyte> ah well...time for grandia anyway. btw, toad, is my SCIP ok?
[9:37] <iip_i2p> <Newsbyte> ah well...time for grandia anyway. btw, toad, is my SCIP ok?
[9:37] <toad_> also how long has your node been up?
[9:37] <toad_> Newsbyte: I'm supposed to check it?
[9:37] <theArthur> 1 arthur arthur 26215373 Aug 10 07:53 seednodes.ref
[9:37] <toad_> theArthur: that's the right one
[9:37] <iip_i2p> <Newsbyte> well, if not you, who else? :-)
[9:38] <theArthur> the date was very wrong before i touched it
[9:38] <theArthur> perhaps this will cause things to behave better
[9:38] <greycat> probably
[9:38] <theArthur> restarting freenet now
[9:38] <toad_> Newsbyte: I have as little chance of convincing you and hobx that nuclear power is dirty as you have of convincing me it's clean. so ceasefire, and take it to somewhere where it isn't quite so offtopic.
[9:39] <theArthur> java.lang.OutOfMemoryError
[9:39] <theArthur> Caught, in Main:
[9:39] <theArthur> thats new
[9:39] <theArthur> i have lots of swap
[9:40] <greycat> how much RAM?
[9:40] <toad_> theArthur: okay
[9:40] <toad_> edit your start-freenet.sh
[9:40] <theArthur> ok
[9:40] <toad_> change -Xmx128M to -Xmx160M
[9:40] <toad_> stop freenet
[9:40] <toad_> touch seednodes.ref again
[9:40] <toad_> start it
[9:40] <theArthur> ty
[9:41] <jokern> toad_ : Getting some of these now : 15:56:37 Type wrong on reporting: transfer rate should be time
[9:42] <iip_i2p> <Newsbyte> my point wasn't that it's clean. If hobx says that, I disagree. What I'm saying is, it has it's benefits too, and on the short run, it's one of the best choices. On the long run, I agree we should move to something else, like fusion.
[9:43] <toad_> Newsbyte: so does carbon sequestration, in the short run. I suggest that the best way you can move forward is by arranging for a gateway from iip to #freenet-chat :)
[9:43] <theArthur> greycat: 128+256
[9:43] <toad_> bbl, going for walk
[9:43] <theArthur> still bailing on out of mem
[9:43] * greycat was going to suggest raising -Xmx but toad beat me to it
[9:44] <theArthur> what dose -Xmx stand for?
[9:45] <greycat> it's the amount of memory the Java runtime allocates for something
[9:45] <theArthur> there are 2 lines that use it
[9:45] <theArthur> i have now changed them both
[9:46] <hobx_> What did you do to make the node require hundreds of megs?
[9:46] <hobx_> Don't you think a refactoring is in order or something.
[9:46] <theArthur> i hve no idea
[9:46] <theArthur> i have everything compleatly stock
[9:46] <hobx_> theArthur: Not you. The coders.
[9:46] <theArthur> except for a 2 gb store
[9:46] <theArthur> im willing to accept this as my fault
[9:46] <hobx_> it isn't.
[9:47] <theArthur> it ALWAY comes down to that through some long chain of interactions
[9:47] * hobx_ feels so bad for all the poor users who keep thinking the problem is with them.
[9:47] <greycat> freenet has a lot of issues still.
[9:47] <hobx_> theArthur: Not here. It is our fault.
[9:47] <theArthur> this is my 3rd annual test
[9:47] <iip_i2p> <Newsbyte> Last thing: only, being realistic, there is no way in hell (at least in my country) that we can change all our nuclear energyproduction into 'clean' energy. It's downright impossible, whether we use solar, wind, a combination, whatever: at most we can reach 10 - 20% of current production. which means, we will have to build A LOT of classical energyhouses, which in turn will mean a HUGE addition of CO2 and the like, and we're already way ab
[9:48] <theArthur> i check every year to see if things have changed
[9:48] <theArthur> maby its the 2nd time is hard
[9:48] <hobx_> theArthur: Me too. Better luck next year.
[9:50] <theArthur> perhaps i will have more time some day and can help
[9:51] <iip_i2p> <Newsbyte> above what's allowed according to kyoto. so, you tell me: is it that bad to keep our nuclear engines going, untill a viable solution is found?
[9:52] <greycat> theArthur: freenet actually worked a lot better a month ago. You came in during one of the "interesting time".
[9:52] <greycat> +s
[9:53] <theArthur> is the information stored on my system encrypted like it used to be?
[9:53] <theArthur> my check last year indicated something changed about that
[9:54] <hobx_> you certainly won't see any plaintext on your harddisk
[9:54] <greycat> it used to be "encrypted" twice: one by the nature of freenet (you need the key to decrypt it), and once by a private node encryption key. That second layer is no longer in place.
[9:54] <theArthur> ok
[9:55] <theArthur> so if somone posts something illegal in my area and it happens to be sitting on my system when BB come in i go to jail
[9:55] <theArthur> ie how is it encrypted by the nature of freenet
[9:55] <iip_i2p> <Newsbyte> also, arthur, have you had a look at www.freenethelp.org, yet?
[9:55] <hobx_> theArthur: That would have happened anyways.
[9:59] * guido^pe (~unknown@dsl-213-023-238-112.arcor-ip.net) has joined #freenet
[10:06] * unknown_ (~unknown@dsl-082-082-158-185.arcor-ip.net) has joined #freenet
[10:11] * hirvox (~hirvox@cs181027153.pp.htv.fi) has joined #freenet
[10:13] * thelema is abck for a while
[10:14] <thelema> theArthur: files are encrypted before they're inserted into the network.
[10:14] * lolo-laptop (~lostlogic@68.251.84.226) has joined #freenet
[10:15] * ShaunMacPherson (~ShaunMac@Sudbury-HSE-ppp3980474.sympatico.ca) has left #freenet
[10:15] <thelema> the decryption key is part of the URI (or can be derived from the URI)
[10:16] * theArthur needs a more fundamental understanding
[10:16] <theArthur> so far i hve managed to get one thumbnail image to load
[10:16] <theArthur> am i to expect better?
[10:19] <theArthur> what dose "backed off" and "cleanly rejected" indicate?
[10:19] <sanity> theArthur: give it time, it will improve
[10:19] * guido^pe (~unknown@dsl-213-023-238-112.arcor-ip.net) Quit (Read error: 110 (Connection timed out))
[10:19] <theArthur> im assuming that cleanly rejected indicates that that "direction" wont take me to the data
[10:19] <sanity> theArthur: when you send a node a message it will say "don't send me another message for X seconds" - during this time the node is said to be "backed off"
[10:19] <theArthur> ok
[10:20] <sanity> cleanly rejected means that the node couldn't serve your request for some reason (perhaps it is starting up or seriously overloaded)
[10:20] <theArthur> not as good as i was hopeing for
[10:20] <sanity> theArthur: it takes a few hours to get good, it has to collect data about other nodes in the network to help it make effective routing decisions
[10:21] <theArthur> the cleanly revected thing that it
[10:21] <theArthur> i can immagin
[10:21] <theArthur> im still trying to get my mind around the overall structure
[10:21] <thelema> it takes a fw hours for your node to build connections to enough nodes to not have RNFs caused by the high mRIs we have.
[10:21] <sanity> theArthur: its complicated :-)
[10:21] <theArthur> ?mRI?
[10:21] <theArthur> ?RNF?
[10:22] <sanity> the: see http://freenetproject.org/index.php?page=ngrouting, it might help
[10:22] <sanity> mRI is the time your node most back-off for after sending a request
[10:22] <sanity> RNF means Route Not Found, your node couldn't find another node that wasn't backed off
[10:22] <theArthur> ok
[10:23] <sanity> RNFs are common while your node gets assimilated into the network
[10:23] <theArthur> so now i go fencing for a few hours...
[10:23] <sanity> theArthur: i assume you mean the legal kind of fencing :-)
[10:23] <theArthur> if i have to restart it saves it connections in seednodes.ref?
[10:23] <sanity> theArthur: ie. not selling stolen goods
[10:23] <theArthur> is that a correct assumption
[10:23] <sanity> theArthur: no, it saves them somewhere else
[10:24] <theArthur> yea, stabbling people with stolen swords ;)
[10:24] <sanity> theArthur: seednodes.ref is only used the first time your node runs
[10:24] <theArthur> is there some way i can give my list of nodes to sombody
[10:24] <sanity> theArthur: yes
[10:24] <sanity> theArthur: use the distribution servlet
[10:24] <theArthur> to help them get started with out waiting for the really bussy seeds
[10:25] <theArthur> so i run that and put ip on on my website
[10:25] <sanity> theArthur: you can also download them from the web interface, but i can't recall the URL - perhaps someone else can help you with that
[10:25] <thelema> theArthur: you run that and give the address it gives you to some people.
[10:25] <theArthur> im off now, go stab ppl with stolen goods
[10:25] <sanity> theArthur: happy perforation
[10:30] * joltrushsoon (josh@mir.stevecole.org) has joined #freenet
[10:40] * thelema is now known as thelema|away
[10:58] * interrupt (~chatzilla@64.122.23.213) Quit ("Chatzilla 0.9.64b [Mozilla rv:1.7/20040707]")
[11:08] <joltrushsoon> do requests and inserts have any kind of priority?
[11:10] <toad_> hi ppl
[11:10] <toad_> especially sanity :)
[11:11] <toad_> <hobx_> What did you do to make the node require hundreds of megs?
[11:11] <toad_> <hobx_> Don't you think a refactoring is in order or something.
[11:11] <toad_> partly, writing it in java :)
[11:11] <toad_> apart from that, profiling is in order
[11:12] * toad_ ignores Newsbyte, take it to chat
[11:14] * toad_ thinks fencing sounds like fun, but barely has enough time for aikido and walking :)
[11:14] <toad_> joltrushsoon: priority over what?
[11:14] <joltrushsoon> other requests
[11:14] <toad_> okay, now to fix the f*cking memory usage on seeding problem
[11:14] <joltrushsoon> like, some kind of priority setting
[11:14] <toad_> joltrushsoon: do requests have priority over requests?
[11:14] <toad_> ah no
[11:15] <joltrushsoon> ok, thanks
[11:15] <toad_> a) it would give far too much away to attackers and nodes, b) it'd be abused - everyone would set it to max for their reqs
[11:15] <joltrushsoon> sure
[11:16] * Zorix (Brandon@fl-64-45-226-120.dyn.sprint-hsd.net) Quit (Read error: 110 (Connection timed out))
[11:17] <toad_> hmmm
[11:17] <toad_> first, check what is in my stable that shouldn't be...
[11:17] <joltrushsoon> erm, just for my clarity, what would it give to attackers?
[11:17] <toad_> joltrushsoon: request types, probably
[11:18] <toad_> also we'd probably need separate estimators for different priority requests
[11:18] <toad_> it'd be messy
[11:18] <toad_> joltrushsoon: you minimize the amount of info given away
[11:18] * thelema|away wonders what would happen if we changed all the bigints in the estimator code to doubles (or even just floats)
[11:18] * thelema|away is now known as thelema
[11:18] <toad_> joltrushsoon: for similar reasons we'd like to fix the key sizes eventually and use splitfiles for everything
[11:18] <thelema> I really don't think the last 20 bytes of accuracy are that important...
[11:18] <toad_> thelema: floats? eeeewww
[11:19] <toad_> you do know that floats are 19 bits accuracy, right?
[11:19] <thelema> bits.
[11:20] <thelema> I think it's sufficient for estimator points to be seperated by a minimum of 1/512000th of the keyspace
[11:21] <thelema> all the extra bytes of accuracy give us is lots of extra work and the ability for points to be really really close.
[11:21] <thelema> once we've divided the keyspace into 512,000 pieces (19 bits of accuracy), I don't think we need to be any more accurate for estimators.
[11:22] <toad_> i don't like artificially limiting freenet's scalability just to obtain 30% less cpu usage or something
[11:23] <thelema> 16 point KSEs don't limit freenet's scalability?
[11:24] <toad_> thelema: not if freenet routing works
[11:24] <thelema> *estimator*
[11:24] <thelema> and the goal wasn't to save CPU usage, it was to save memory
[11:25] <thelema> floats take up lots less space (1/23rd the space)
[11:25] <thelema> and we use a lot of them per node. (16 * 5?) * 200
[11:25] <thelema> 16,000 bigints
[11:26] <thelema> we already convert to a double before doing interpolation...
[11:27] <thelema> why not store as a double?
[11:28] * Hory (~Miranda@81.196.25.110) Quit (Connection timed out)
[11:29] <toad_> why is eclipse so slow today?
[11:30] <thelema> phase of the moon?
[11:30] <greycat> The Moon is Waning Crescent (25% of Full)
[11:31] <toad_> hmmm
[11:32] <toad_> format seems to be doing irritating things
[11:32] <toad_> it changes d.registerBinomial(...) to d.\nregisterBinomial(\n...
[11:32] <toad_> even though d.registerBinomial( would easily fit on one line
[11:32] <toad_> how to fix?
[11:35] * mazzanet (~mazzanet@mazzanet.registered) Quit (Read error: 104 (Connection reset by peer))
[11:35] <thelema> minAllowed is a double; BucketStores using as much precision as you're giving them will be rejected as invalid anyway.
[11:36] <toad_> thelema: from network yes
[11:36] <toad_> but they can evolve to use that much accuracy
[11:36] * mazzanet (~mazzanet@mazzanet.registered) has joined #freenet
[11:39] <thelema> I really doubt that'll happen much because of there only being 16 buckets.
[11:41] * TLF (francisco@36.Red-81-40-113.pooles.rima-tde.net) has joined #freenet
[11:44] <thelema> but we'll see the difference in effectiveness once I'm done with DoubleBucketStore
[11:48] <toad_> ugh
[11:48] <toad_> Eclipse is losing files, and refusing to apply new files from patches
[11:49] <toad_> oh and vim takes literally a minute to load
[11:49] <toad_> every single time
[11:50] <toad_> okay, worked around it
[11:50] <toad_> (close the project and open it)
[11:50] <toad_> anyone'd think frustration for the users costs less than programmer hours to fix the bugs :)
[11:55] <toad_> of course it's poetic justice as i've been avoiding fixing the seednodes bug for the last two days :)
[11:55] <KenMan> argh - 60187. averageOutgoingMRI made a nice smooth curve since startup , from useable to this level of backoff :
[11:55] <KenMan> http://www.dragg.net/users/pennywitt/ysam/ysam4.gif
[11:56] <toad_> KenMan: umm, can you show me the original data?
[11:56] * cehteh (foobar@cehteh.homeunix.org) has joined #freenet
[11:56] <thelema> nice backoff level.
[11:56] * cehteh (foobar@cehteh.homeunix.org) has left #freenet
[11:56] <toad_> e.g. as a graph?
[11:57] <KenMan> yeah, hold on pardner ...
[11:58] <toad_> PRI Aug 10, 2004 5:16:05 PM (freenet.node.Main, main, NORMAL): Memory usage now:
[11:58] <toad_> 46911328
[11:58] <toad_> PRI Aug 10, 2004 5:16:14 PM (freenet.node.Main, main, NORMAL): Memory usage afte
[11:58] <toad_> r reading seeds: 178426496
[11:58] <toad_> woah
[11:59] <toad_> and then a lot of EstimatorFormatException's
[11:59] <KenMan> http://mywebpages.comcast.net/jkcorson/FreeNet/60187c.png
[12:01] <toad_> hmmm
[12:01] <toad_> is that with frost running?
[12:01] <toad_> imho that's because of feedback between high outRIs and low incomingrequests
[12:02] <toad_> which is the fundamental problem with the current propagation code - outgoing reqs / incoming reqs gets rather high if incoming reqs is low
[12:03] <KenMan> no, i shut down frost around 2:00 , surprise surprise...
[12:03] <toad_> Memory usage now:
[12:03] <toad_> 47772840
[12:03] <toad_> Memory usage afte
[12:03] <toad_> r GC: 45794504
[12:03] <toad_> Memory usage afte
[12:03] <toad_> r reading seeds: 173080384
[12:03] <toad_> Memory usage afte
[12:03] <toad_> r GC: 172671624
[12:03] <toad_> Memory usage afte
[12:03] <toad_> r seeding: 172677192
[12:03] <toad_> Memory usage afte
[12:03] <toad_> r routingStore checkpoint: 172762248
[12:03] <toad_> Memory usage afte
[12:03] <toad_> r left seeding block: 172765872
[12:03] <toad_> Memory usage afte
[12:03] <toad_> r GC: 172674264
[12:04] <toad_> WHAT?!
[12:05] <KenMan> important disk buffers that java caches (for the OS) or some such nonsense
[12:05] <toad_> no
[12:05] <toad_> wrong scope for the variable
[12:05] <toad_> so it didn't get GC'd
[12:05] <KenMan> :)
[12:06] <toad_> that's an obscene amount of memory...
[12:06] * toad_ thinks we could improve on that just by keeping them as strings
[12:06] <toad_> Memory usage after left seeding block: 155666160
[12:06] <toad_> Memory usage after GC: 29599656
[12:07] <toad_> okay, that's more what i expected
[12:08] <KenMan> 7 peers on newunstable. Toad will you mind if I defect back to newstable, to collect numbers ??
[12:08] <toad_> first off, don't seed unless we need to
[12:08] <toad_> KenMan: i think newunstable has proved that the current rate limiting propagation code sucks
[12:09] <toad_> KenMan: if you have any better ideas for how to do the rate limiting load propagation...
[12:09] <toad_> tell me them after i've fixed the seednodes
[12:09] <KenMan> well, yeah. And I can't contribute to the net until 2007...
[12:09] <thelema> outgoing reqs stays high even when incoming reqs goes low?
[12:09] <thelema> isn't that just a measuring problem?
[12:09] <KenMan> request interval (if you are looking at the graph)
[12:12] * bluephile (bluephile@69-160-205-193.clvdoh.adelphia.net) has joined #freenet
[12:12] <KenMan> thelema ? what are you referring to ?
[12:13] <KenMan> hey bluephile - moderate time no see...
[12:13] <bluephile> hey KenMan
[12:13] <bluephile> what's up
[12:13] <toad_> thelema: perhaps
[12:14] <toad_> but there were SOME requests
[12:14] <toad_> so it's not just "we only report the average from the old values which are close together"
[12:14] <toad_> KenMan: defect for now, later we need to discuss
[12:15] <KenMan> yeah. will do.
[12:15] <toad_> KenMan: it'd be nice to graph the variables that it uses to calculate the loafd
[12:15] <toad_> load
[12:15] <toad_> from the General infolet
[12:15] <toad_> propagation load
[12:17] <KenMan> ahh... the x / y values . gotcha
[12:22] <toad_> also graph the actual incoming reqs
[12:22] <toad_> maybe y is inaccurate
[12:37] <toad_> bbiab
[12:38] * toad_ is about half way to finishing the fix
[12:41] * Zorix (Brandon@fl-64-45-226-120.dyn.sprint-hsd.net) has joined #freenet
[12:43] * TLF (francisco@36.Red-81-40-113.pooles.rima-tde.net) Quit ("http://www.vitaliano.esp.cc || Adi?s a todos y a todas y que les vaya bien.")
[12:45] * salahx (salahx@sc1-24.217.174.147.charter-stl.com) has joined #freenet
[12:55] <thelema> toad_: is there a reason we write out both center and leftEdge to disk (in BucketStore)?
[13:09] * thelema commits FloatBucketStore.java
[13:15] * Hory (~Miranda@81.196.25.110) has joined #FreeNet
[13:23] <thelema> brb
[13:24] * jay (jcl@ool-18bf6dac.dyn.optonline.net) has joined #freenet
[13:26] <Redb3ard> so toad_, what do you think would be possible anon-network-wise, with something like an ansible?
[13:28] * ticoli (~ticoli@179.254.39-62.rev.gaoland.net) has joined #freenet
[13:34] <thelema> Redb3ard: FTL communication is assumed in our network design.
[13:34] <leex-work> whats toad_ bbl like some hrs
[13:35] * leex-work is now known as leex-going-home
[13:35] <leex-going-home> :)
[13:35] <leex-going-home> i got 3 nods running now
[13:35] <leex-going-home> nodes
[13:42] * TLF (francisco@36.Red-81-40-113.pooles.rima-tde.net) has joined #freenet
[13:46] <Redb3ard> i was thinking more along the lines of the wireless through lead walls quality
[13:46] <Redb3ard> think about it, there is no good way to know where the other transciever is
[13:47] <leex-going-home> heh
[13:47] <leex-going-home> brb
[13:47] <leex-going-home> going home
[13:47] <leex-going-home> trying to
[13:52] * moskau23 (~Miranda@dsl-213-023-250-039.arcor-ip.net) Quit (Read error: 104 (Connection reset by peer))
[13:53] * TLF (francisco@36.Red-81-40-113.pooles.rima-tde.net) Quit ("http://www.vitaliano.esp.cc || Adi?s a todos y a todas y que les vaya bien.")
[14:20] * Lage (~Miranda@pD9EAF531.dip.t-dialin.net) has joined #freenet
[14:27] * Lage (~Miranda@pD9EAF531.dip.t-dialin.net) has left #freenet
[14:28] * Lage (~Miranda@pD9EAF531.dip.t-dialin.net) has joined #freenet
[14:28] * salahx (salahx@sc1-24.217.174.147.charter-stl.com) Quit ()
[14:34] <Iakin> Hmmm..
[14:34] <Iakin> - Core.logger.log(
[14:34] <Iakin> + Core.logger
[14:34] <Iakin> + .log(
[14:35] <Iakin> ...that was from your latest commit toad
[14:35] * sanity (~ian@81-178-89-164.dsl.pipex.com) Quit (Read error: 104 (Connection reset by peer))
[14:35] <thelema> Iakin: yeah, format was being mean to him.
[14:36] * sanity (~ian@81-178-89-164.dsl.pipex.com) has joined #freenet
[14:36] * ChanServ sets mode +o sanity
[14:39] * Zorix (Brandon@fl-64-45-226-120.dyn.sprint-hsd.net) Quit (Read error: 110 (Connection timed out))
[14:45] * TLF (francisco@226.Red-81-40-113.pooles.rima-tde.net) has joined #freenet
[14:46] * spaetz (~spaetz@217-162-197-136.dclient.hispeed.ch) has joined #freenet
[14:48] <Iakin> Cool thelema.. That alternate bucketstore can be really interesting to see the impact of
[14:51] <thelema> Iakin: it's not enabled in the code; I encourage people to make the few-line change to replace BucketStore with FloatBucketStore (hopefully it's only a few lines), and try it out as a vriant build.
[14:52] <thelema> actually, I'm going to need to write some more code; it's not compatible fieldset-wise with other nodes, so no xferring estimator data when you use it.
[14:52] * Superfan (francisco@226.Red-81-40-113.pooles.rima-tde.net) has joined #freenet
[14:53] * Superfan (francisco@226.Red-81-40-113.pooles.rima-tde.net) Quit (Client Quit)
[14:54] <thelema> I'll be back later tonight (6-7 hours)
[14:54] * thelema is now known as thelema|away
[14:58] * sanity (~ian@81-178-89-164.dsl.pipex.com) Quit (Read error: 104 (Connection reset by peer))
[15:03] * TheIdiot (~x@cc21438-a.groni1.gr.home.nl) has joined #freenet
[15:04] <TheIdiot> hello everybody
[15:04] <TheIdiot> i fucked up my node, could someone help me
[15:04] * thelema|away is now known as thelema
[15:04] * thelema isn't away as I thought I'd be
[15:05] <TheIdiot> see i think the seednodes.ref is too big, i get "java virtual machine out of memory errors"
[15:05] <TheIdiot> but i got over 500 mb memory!
[15:08] * sanity (~ian@81-178-89-164.dsl.pipex.com) has joined #freenet
[15:08] * ChanServ sets mode +o sanity
[15:09] * ticoli (~ticoli@179.254.39-62.rev.gaoland.net) Quit ("Leaving")
[15:12] <thelema> TheIdiot: java won't use all of your memory by default.
[15:13] <thelema> change -Xmx128M to -Xmx160M in your start-freenet.sh
[15:14] <TheIdiot> wow thanx
[15:19] <TheIdiot> i just set the table.. and my cat he just vomited neatly on one of the plates
[15:19] <TheIdiot> its like an entree you know
[15:21] <TheIdiot> hes not castrated btw
[15:21] <TheIdiot> i see little black kittens all over the neighbourhood
[15:22] <TheIdiot> the other day he dragged this rotten mole into the backyard, must have been dead for weeks, crawling with maggots
[15:25] <TheIdiot> er, i dont seem to have a start-freenet.sjh
[15:25] <TheIdiot> sh
[15:25] <TheIdiot> is that a linux thing
[15:25] <leex-going-home> or freenet.ini
[15:26] <thelema> yes. are you a windows user?
[15:26] <TheIdiot> yes.
[15:26] * leex-going-home is now known as leex-at-home
[15:26] <thelema> is there a -Xmx128M in your freenet.ini?
[15:27] <TheIdiot> no
[15:27] <leex-at-home> i look
[15:27] <leex-at-home> no 128mbs
[15:27] <leex-at-home> in freenet.ini
[15:28] <TheIdiot> hmm i cant be the only one with this problem
[15:30] <TheIdiot> how do you tell the VM to request more memory
[15:30] <leex-at-home> there be lots if thay update there nodes
[15:30] <leex-at-home> 1 sec i give you an small one
[15:30] <jokern> look for a file named flaunch.ini
[15:31] <jokern> Put in JavaMem=256M or whatever you need
[15:31] <leex-at-home> cool
[15:32] <leex-at-home> 256m or mb
[15:32] <jokern> M
[15:32] * JustMe_ (JustMe_@cs6669187-74.houston.rr.com) has joined #freenet
[15:32] <leex-at-home> (i get the same error with 1gb of ram heh)
[15:32] <leex-at-home> but this should fix it
[15:33] <jokern> if you are on win :)
[15:33] <TheIdiot> great
[15:33] <TheIdiot> it was set to default
[15:33] <leex-at-home> still the freenet.exe should give better errors
[15:33] <leex-at-home> or even better give errors with out haveing to look at the log
[15:34] <TheIdiot> ah it seems to be working, nice
[15:38] <TheIdiot> its really working, eureka
[15:38] <Redb3ard> so when are you guys gonna roll freenet into a kernel module?
[15:40] <thelema> Redb3ard: never. mjr wrote a similar type program as a kernel module, but freenet is too heavy for a kernel module.
[15:40] <TheIdiot> thanks again and goodbye
[15:40] * TheIdiot (~x@cc21438-a.groni1.gr.home.nl) Quit ()
[15:43] * TLF (francisco@226.Red-81-40-113.pooles.rima-tde.net) Quit ("http://www.vitaliano.esp.cc || Adi?s a todos y a todas y que les vaya bien.")
[15:47] * Lage (~Miranda@pD9EAF531.dip.t-dialin.net) Quit (Read error: 104 (Connection reset by peer))
[15:53] * Zorix (Brandon@fl-64-45-226-120.dyn.sprint-hsd.net) has joined #freenet
[15:59] * jay (jcl@ool-18bf6dac.dyn.optonline.net) Quit ("Client exiting")
[16:16] * salahx (salahx@sc1-24.217.174.147.charter-stl.com) has joined #freenet
[16:16] <theArthur> ahoy
[16:17] <theArthur> how indiciteive of the network load as a whole it the load indicator on fproxy?
[16:19] <theArthur> my node has 57 outgoing connections and 2 incoming? this cant be healthey for the net?
[16:19] <theArthur> is that normal
[16:19] <hirvox> connections are bidirectional
[16:19] <salahx> right
[16:19] <greycat> it's normal for a new node to have lots more outgoing than incoming
[16:19] <salahx> all it means is you intieated the connection instead of the other node
[16:19] <greycat> once the connection has been made, it doesn't matter which direction it's going
[16:19] <theArthur> then there must be some one out thare with lots of incoming?
[16:21] <salahx> WEll those not behind a NAT/firewall (or a peroperly confugured one) will have listenPort open
[16:22] <theArthur> I am running it on a system not NATed, i would like to encourage it to accept incoming connections
[16:22] <theArthur> probibly cant do that?
[16:22] <salahx> it'll so do autmaitcally
[16:22] <salahx> Since you;re recieveing incoming connections, you have it configured properly
[16:22] <theArthur> just need a while to be found
[16:23] <salahx> right
[16:24] * Hory (~Miranda@81.196.25.110) Quit (Read error: 54 (Connection reset by peer))
[16:25] <theArthur> if it wernt for the evils of DSLs bein assymetric, systems like this would work so much bette
[16:26] <salahx> yeah
[16:26] <salahx> But there are signs of hope: http://www.dslreports.com/shownews/51636
[16:27] <theArthur> i have a nice rant on the subject, need to get this working well enough to insert it so it can be properly ignored without wasting anyones time
[16:27] <theArthur> 51636 up?
[16:28] <salahx> eh ?
[16:28] <salahx> tha'ts just the acticle number
[16:28] <theArthur> ahh
[16:28] <theArthur> the size of my window split the line and confused me (part of the vast conspiracy)
[16:28] <salahx> ahh
[16:28] * theArthur reads link
[16:32] <salahx> Freenet will work much ebtter once the mRI problem is fixed
[16:33] <theArthur> thats the issue with nodes always being backed off?
[16:33] <salahx> yes
[16:33] <theArthur> yea thats a tad irratating
[16:34] <theArthur> is the network really that overloaded, or is there a problem with how the delays are calculated
[16:34] <salahx> Freenet, by defiantltion is ALWAYS "overloaded"
[16:34] <theArthur> if that overloaded then can anything be done about the back off times
[16:35] * greycat (~wooledg@192.35.79.70) Quit ("Lick Bush in '04")
[16:35] <salahx> tha'ts wha'ts bring worked on
[16:35] <theArthur> definatly not a simple bug fix
[16:35] <salahx> not that simple, yes
[16:35] <salahx> certianyl not the hardest problem though
[16:36] * Zorix (Brandon@fl-64-45-226-120.dyn.sprint-hsd.net) Quit (Read error: 110 (Connection timed out))
[16:36] <theArthur> when did it show up?
[16:36] <salahx> the problem always existed; it didn't show uyp until maybe 2 months ago
[16:36] <theArthur> (i started my annual freenet checkup today) so i dont have much prespective
[16:38] <thelema> salahx: the problem has been around since the beginning of 0.4
[16:38] <salahx> an "Exterminatal unstable" build has been created to play around with some ideas
[16:38] <salahx> that long ?
[16:39] <thelema> it's just that the way we've analyzed freenet has been blind to this kind of problem until lately.
[16:39] <thelema> lately = last year or so
[16:40] <theArthur> any ideas?
[16:40] <theArthur> im pondering taking time from my other projects to work on this
[16:40] <salahx> Yep, several
[16:41] <theArthur> but the project as a whole seems like it has a very steep learning curve
[16:41] <theArthur> do you have time to explain how the backoffs are calculated?
[16:41] <theArthur> ?link?
[16:42] <salahx> Ufneoruntyl even I'm nto sure on the exact formula
[16:42] <theArthur> the obzious question: why not just subtrace some from the timout?
[16:43] <theArthur> how dellacate is the network in respect to that?
[16:44] <salahx> well essential the backoff is computed so that the node has 100% utilitzation
[16:44] <salahx> There's no point in sending a request to node that will simply say "I'm too busy"; its a waste of bandwith and CPU time
[16:45] <theArthur> is the problem that the backoffs are being exadurated?
[16:45] <thelema> the backoffs aren't being exaggerated.
[16:45] <salahx> oen is they get quite large; the other being the change too much
[16:46] <thelema> nodes just become really busy sometimes and need to tell other nodes 'send your queries elsewhere'
[16:46] <thelema> but at other times, they just need to moderate the number of incoming queries so that it's at a level they can handle.
[16:47] <thelema> At the moment, nodes come up with values for how much load they're under (as a % of full utilization) and use that value to adjust the existing mRI so that the new mRI should result in 100% load
[16:47] <thelema> s/100% load/100% utilization/
[16:48] <theArthur> <assumption> but they are beind calculated so the nodes are underutalized</assumption>
[16:50] <thelema> I can assure you that nodes are quite overutilized; only nodes with *huge* amounts of bandwidth aren't being limited by the bandwidth limiter
[16:53] <thelema> the number of concurrent xfers and the resulting speed of them is another indication of overload
[16:54] <salahx> Yeah, if our node is trying to tranfer 50 1 Meg chunks its going to choke :)
[16:56] <theArthur> until things change i guess all i can do is wait...
[16:56] <theArthur> :(
[16:56] <thelema> it's going to need to not start any more xfers
[16:57] <thelema> at the moment, there's no way to accept (non-trivial) requests without having the chance of needing to perform more data xfers
[17:12] * dERMzzz is now known as dERMOTH
[17:24] * TLF (francisco@148.Red-81-40-116.pooles.rima-tde.net) has joined #freenet
[17:26] * unknown_ (~unknown@dsl-082-082-158-185.arcor-ip.net) Quit ()
[17:32] <iip_i2p> <Newsbyte> nono, arthur
[17:32] <leex-at-home> ?
[17:32] <iip_i2p> <Newsbyte> all you can do is financially support Freenet, in the meantime ;-)
[17:33] * leex-at-home is now known as leex-remote
[17:33] <leex-remote> ic
[17:33] * lolo-laptop (~lostlogic@68.251.84.226) Quit ("Client exiting")
[17:33] <leex-remote> need to read back an little
[17:41] * moskau23 (~Miranda@dsl-213-023-250-039.arcor-ip.net) has joined #freenet
[17:42] * thelema is now known as thelema|away
[17:52] * TLF (francisco@148.Red-81-40-116.pooles.rima-tde.net) Quit ("http://www.vitaliano.esp.cc || Adi?s a todos y a todas y que les vaya bien.")
[17:55] <toad_> <thelema> toad_: is there a reason we write out both center and leftEdge to disk (in BucketStore)? - no. I didn't. Somebody changed the format.
[17:58] * theArthur (~arthur@unix1.ic.ucsc.edu) Quit ("Leaving")
[17:59] * Zorix (Brandon@fl-64-45-226-120.dyn.sprint-hsd.net) has joined #freenet
[18:00] <toad_> hi
[18:06] <leex-remote> lo
[18:11] <leex-remote> i got 10 users on this little network of ares
[18:11] <leex-remote> toad_ :)
[18:13] <toad_> ugh
[18:13] <toad_> horrible bytes vs chars vs ReadInputStream vs String issues...
[18:14] <toad_> I'm going to have to take the data, parse it into strings to find the lines, put them together into pre-fieldset big strings, then turn them back into byte arrays to parse them!
[18:14] <toad_> and all because of the idiocy of the binary presentation operating partially on UTF8 and partially on raw bytes
[18:15] * toad_ supposes he could keep them as bytes and scan for \n's
[18:15] <toad_> but.. yuck
[18:15] <leex-remote> ok ??
[18:15] <toad_> leex-remote: what's a network of ares
[18:15] <toad_> ?
[18:16] * toad_ will have to hack something together to deal with \r\n's too
[18:16] <toad_> this is ugly
[18:17] <toad_> not hard but just horribly ugly
[18:17] <leex-remote> ?
[18:17] <leex-remote> ares ?
[18:17] <toad_> <leex-remote> i got 10 users on this little network of ares
[18:20] * guido^pe (~unknown@dsl-082-082-158-185.arcor-ip.net) has joined #freenet
[18:24] <verl> toad: i think he means 'ours'. also is it meaningfull to stay on it any longer? are you through testing?
[18:24] <leex-remote> ours yes
[18:24] <leex-remote> heh
[18:25] * leex-remote is now known as leex-remote-gett
[18:25] <leex-remote-gett> fs
[18:25] * leex-remote-gett is now known as leex-r-geting-fo
[18:25] <leex-r-geting-fo> fs
[18:25] <leex-r-geting-fo> FOOD
[18:25] <verl> :)
[18:25] <toad_> verl: not sure
[18:25] <toad_> KenMan's graph was pretty alarming
[18:26] <toad_> now this is really annoying
[18:26] <verl> toad: so more change are coming?
[18:26] * toad_ is going to have to do some REALLY messy hacks here
[18:26] <toad_> we ought to be able to support arbitrarily large seednodes files, right?
[18:28] * toad_ hmmm
[18:28] <toad_> the solution is to add a readLineOfBytes method to RIS... and avoid binary presentation for seednodes
[18:29] <toad_> binary presentation for seednodes would actually be rather nice, that's the problem...
[18:29] <toad_> somebody's bound to do it sooner or later
[18:31] <toad_> I suppose I could go through it twice, count them the first time...
[18:32] <toad_> or I could write the CGI script...
[18:34] * toad_ just turns it back into byte[] :(
[18:42] <KenMan> is this complete ? List of query sources: 1) requestors 2) FCP (fproxy&frost) 3) previously routed items that QR'd (looped, RNFed, whatever)
[18:45] * KenMan wonders if the amphibian received KenMan's graphic in the mail yet
[18:58] <toad_> KenMan: hmmm
[18:58] <toad_> 1) remote requestors
[18:58] <toad_> 2) FCP
[18:58] <toad_> 3) internal requestors
[18:58] <toad_> 4) retries
[19:03] <toad_> internal requestors includes fproxy
[19:03] <KenMan> isn't 2 & 3 the same thing ?
[19:03] <toad_> but there are others
[19:03] * weilawei (1000@122.2-177-216.bed.freedialup.org.dialup.G4.Net) has joined #freenet
[19:03] <KenMan> I thought fproxy went through FCP !!
[19:04] <toad_> no
[19:04] <toad_> it can
[19:04] <toad_> but it doesn't
[19:04] <KenMan> hmmm... more to learn :)
[19:05] <KenMan> how about that splitfile interface ? that doesn't go through FCP either ?
[19:05] <toad_> nope
[19:05] <KenMan> :o
[19:05] <toad_> also the fetch-the-alinks-on-startup code
[19:05] <toad_> initial requests from announcements
[19:05] <toad_> and other stuff
[19:06] * toad_ wonders why vim consistently takes ~1 minute to load
[19:06] <weilawei> LMFAO
[19:06] <weilawei> my vim loads in a blink
[19:06] <weilawei> whatd you do to it?!
[19:07] <toad_> i don't know
[19:07] <toad_> perhaps the combination of freenet, eclipse and linux 2.6
[19:07] <weilawei> ouch maybe
[19:08] <weilawei> vim should load instantaneuosly.. it is a comparatively lightweight editor.. but its the best., better than GUI editors even
[19:12] <mikeDOTd> what about emacs? </troll>
[19:13] <weilawei> YECH
[19:13] <weilawei> </reply>
[19:14] <KenMan> don't worry, the amphibian still loves emacs AND vi (he IS an amphibian, after all)
[19:14] * toad_ uses eclipse for coding though
[19:14] <weilawei> lmao
[19:15] <weilawei> eclipse is based on what editor?
[19:15] <toad_> eclipse
[19:15] <KenMan> windows notepad
[19:15] <toad_> heh
[19:15] <weilawei> OOOH wait.. im thinking of a commercial editor based around eclipse
[19:15] <weilawei> DOH
[19:20] * toad_ wonders why it takes so long for the new code to read the seednodes (not to apply them)
[19:21] <weilawei> mayube theres a lot of seednodes?
[19:21] <weilawei> =P
[19:21] * dERMOTH (~dERMOTH@dsl-131-244.aei.ca) Quit ("Client Exiting")
[19:26] <toad_> KenMan: I'm fixing the seednodes thingy
[19:26] <toad_> after that i'll read your mail
[19:30] * weilawei (1000@122.2-177-216.bed.freedialup.org.dialup.G4.Net) Quit ("Leaving")
[19:33] * guido^pe (~unknown@dsl-082-082-158-185.arcor-ip.net) Quit (Remote closed the connection)
[19:35] <toad_> Memory usage aft
[19:35] <toad_> er seeding: 130399024
[19:35] <toad_> hmmmm
[19:35] <toad_> is that low enough?
[19:35] <Redb3ard> hey toad
[19:37] <toad_> hi
[19:39] * leex-r-geting-fo is now known as leex
[19:39] <leex> doad
[19:39] <toad_> hi
[19:40] <leex> lo
[19:41] * greycat (rfc1413@wooledge.org) has joined #freenet
[19:41] <leex> any new ver of your unstable vers
[19:45] <leex> no?
[19:45] <KenMan> no, he has not made a new version yet
[19:45] <leex> :)
[19:45] * salahx is using "experminetal" 60187
[19:45] <salahx> Current estimated load for rate limiting: 3,572.6%. :(
[19:46] <leex> Current estimated load for rate limiting 37.4%
[19:46] <leex> hmm
[19:46] <leex> that good ?
[19:46] <salahx> well your node is underloaded
[19:47] <leex> it should be with only 10 users
[19:47] <salahx> yeah
[19:48] <salahx> Load due to propagation = 3,572.6% = 1147.7309883410226 / 32.125798851198326
[19:48] <toad_> salahx: running frost?
[19:48] <salahx> I was
[19:48] <toad_> you were when?
[19:48] <salahx> until 1 monute ago
[19:49] <toad_> okay, committed 5091...
[19:49] <leex> how many seeds are there now
[19:55] <salahx> Hi have 11 nodes in my RT
[19:56] <leex> hmm i see an 60188 node
[19:58] <leex> you lot quite to night
[19:58] * leex is now known as leex-going-home
[19:58] <jokern> Any other things than mem use fixed in 5091?
[20:00] <KenMan> jokern - i don't think so
[20:00] <jokern> k, then I`ll wait upgrading
[20:01] <jokern> Uptime 2 days 13 minutes :) Thats why
[20:02] <leex-going-home> 1 day 1 hr
[20:02] <leex-going-home> uptime
[20:02] <leex-going-home> (OS) Windows 2000 Enterprise Server, Service Pack 4 (5.0 - 2195), (installed for) 77w 2d 17h 36m, (uptime) 4w 3d 2h 32m 22s, (record) 10w 1d 42m 38s (set on 06.02.2004 23:34 using Win2K)
[20:03] <leex-going-home> probly crip compered to some of the linux users in here
[20:03] <leex-going-home> crap
[20:04] <jokern> :) Even linux has to reboot when upgrading kernel..
[20:04] <KenMan> I built commercial infrastructure apps that depended on W2K to stay up for months at a time, and it did. It all depends on what you run...
[20:05] <KenMan> My apps were simple, and the administrators were grateful for that...
[20:05] <leex-going-home> i did that uptime with out an UPS
[20:05] <leex-going-home> whould of been longer but some n00b program restarted me server
[20:06] <jokern> What does your http://127.0.0.1:8888/servlet/nodestatus/psuccess_data.txt show?
[20:06] <KenMan> ;) those damned b00bs , err, i mean n00bs
[20:06] <leex-going-home> Max: 0.036437247
[20:06] <leex-going-home> Most successful: 5
[20:07] <jokern> Max: 0.039906103 average 0.035 approx
[20:07] <jokern> About the same then
[20:07] <leex-going-home> what that mean then
[20:08] <jokern> Probability of success of an incoming request
[20:09] <leex-going-home> Max: 0.0034227038
[20:09] <leex-going-home> Most successful: 1
[20:09] <leex-going-home> on my other node
[20:09] <leex-going-home> be more numbers if i was part of the full freenet network
[20:10] <jokern> yepp
[20:10] <leex-going-home> 11 nodes ant much (testing#)
[20:12] <leex-going-home> any way i going home and now going sleep
[20:12] * toad_ waits for eclipse to respond...
[20:12] <leex-going-home> ...
[20:13] <leex-going-home> toad_ ner unstable out now ?
[20:13] <leex-going-home> new
[20:13] * leex-going-home is now known as leex-gh-then-sle
[20:13] <toad_> not yet
[20:13] * leex-gh-then-sle is now known as leex-gh-tn-sleep
[20:13] <leex-gh-tn-sleep> go home then sleep :)
[20:14] <toad_> no new stable yet either
[20:14] <leex-gh-tn-sleep> sya tomoro in about 20 hrs
[20:14] <leex-gh-tn-sleep> cya
[20:14] <leex-gh-tn-sleep> work sucks
[20:14] <leex-gh-tn-sleep> ZZzz
[20:18] * JustMe_ (JustMe_@cs6669187-74.houston.rr.com) Quit ()
[20:24] * toad_ revokes the experimental network in 60188...
[20:26] <KenMan> so 60188 is meant for 60165- ?
[20:26] <KenMan> or rather , 60175- ?
[20:26] <KenMan> duh, " 60185- "
[20:29] <toad_> argh
[20:29] * toad_ didn't merge the NPE fix...
[20:32] * JustMe_ (JustMe_@cs6669187-74.houston.rr.com) has joined #freenet
[20:35] * toad_ tests 5092...
[20:37] <leex-gh-tn-sleep> who are the others in this room
[20:37] <Redb3ard> toad, you read orson scott card at all?
[20:37] <KenMan> they are shadowy and mysterious folk, to be sure...
[20:37] <toad_> no
[20:37] * leex-gh-tn-sleep is now known as leex-home-goings
[20:37] <jokern> Alot of sleeping ppl
[20:37] <toad_> not that i know of
[20:38] <Redb3ard> hmm, in one of the books, there is a communication device, the ansible... its claim to fame being that it allows FTL comm
[20:38] <Redb3ard> but also, there is no barrier to the transmission, no distance limitations
[20:38] <greycat> that's an invention from Ursula K. LeGuin's stories.
[20:38] * goatee (~goatee@ip216-239-87-183.vif.net) has joined #freenet
[20:38] <greycat> if Card's using it, it's either plagiarized, or a "tribute"
[20:39] <Redb3ard> i dont believe any scifi authors have it trademarked or anything dude
[20:39] <Redb3ard> the concept was flying around the physics circles as far back as the 1950s
[20:39] <greycat> ah, ok
[20:39] <Redb3ard> anyway, im just trying to figure out how you'd even use something like that
[20:40] <toad_> | I suppose I have to admit to being a little concerned that claiming 5090 seems to work rather well might immediately lead to everything going seriously downhill again. So maybe I should instead just complain about the way 5090 seems to have greatly reduced the instant RNF situation and express my sheer dismay at the increased success rate of recent key tests.
[20:40] <toad_> |At this rate I may actually be left with little to get annoyed about as far a Freenet is concerned ...
[20:40] <Redb3ard> for instance, whether its my network, or freenet... we all have to use the internet as a backbone
[20:40] <toad_> |Nah. It'll never happen.
[20:40] <toad_> hehehehehehe
[20:40] <Redb3ard> cant exactly lay our own fiber cables across the ocean floor
[20:41] <Redb3ard> but better yet, i dont think such devices could be triangulated, or intercepted
[20:41] <toad_> Redb3ard: not yet :)
[20:41] <toad_> well, build one and we will come :)
[20:41] <Redb3ard> and, im trying to come up with a scheme, that if you had 2 of them... you'd not even know who had the other one
[20:41] <Redb3ard> toad, its not so farfetched
[20:41] <Redb3ard> quantum entangle 2 copper atoms
[20:42] <toad_> Redb3ard: not FTL
[20:42] <Redb3ard> put 1 each in a small box that has an gigabit port
[20:42] <greycat> toad_: no, quantum mechanics *does* appear to permit instantaneous communications
[20:42] <toad_> Redb3ard: spooky action at a distance does not convey DATA
[20:42] <Redb3ard> dude, whether or not its FTL is debatable
[20:42] <greycat> toad_: it's "spooky action at a distance" at its finest
[20:42] <Redb3ard> and thats not even the useful part of it
[20:43] <toad_> at least if it does, that's revolutionary in itself, and I'd like a reference on it
[20:43] <Redb3ard> for us, the "wireless at any distance, no way to block or intercept it" is what we'd like
[20:43] <toad_> last I heard although the signal front could travel FTL, the communications data (as opposed to random uncontrollable noise) could not
[20:43] <Redb3ard> toad, so what?
[20:43] <greycat> Redb3ard: the major drawback of these things seems to be that you can only use each pair of atoms *once*, before you have to bring them back together to re-sync them
[20:43] <Redb3ard> i mean really, i wont argue that its faster than light
[20:44] <toad_> hmm, this is quantum teleportation? that requires a solid classical comms channel...
[20:44] <Redb3ard> greycat, theres gotta be a way to cheat though
[20:44] <Redb3ard> use 2 pairs, and have the first transmit enough info to resync the first
[20:44] <Redb3ard> i dunno
[20:44] <toad_> Redb3ard: how do you use it to transmit/receive info?
[20:44] <toad_> i mean, all info even at the quantum level is exchanged via particles, right?
[20:45] <Redb3ard> not even sure of the technical stuff, toad
[20:45] <KenMan> what's wrong with two soup cans and a bit of string ?
[20:45] <greycat> toad_: have an array of 8 atoms, and selectively "set" some of them. That gives you one byte.
[20:45] <Redb3ard> im more concerned whether it could even be put to use, if it did exist
[20:45] <Redb3ard> say i drop one of these in NYC central park, hidden
[20:46] <Redb3ard> then i leave a message on freenet, that theres a QE transciever there, i wont even know who grabs it
[20:46] <Redb3ard> for all they know, i was vacationing, and im back in paris now
[20:46] <toad_> Redb3ard: afaics all communication systems transmit particles
[20:47] <toad_> if you have some new particles we can use, that'd be interesting. e.g. neutrinos would be REALLY cool, there's some work on that.
[20:47] <Redb3ard> neutrinos... well, we could use those even through the earth itself
[20:47] <greycat> Redb3ard: indeed, if you can arrange a nice little demo, I'd be most impressed.
[20:47] <Redb3ard> too bad you need 1 cubic mile of pure water and $1 million worth of sensors to recieve my neutrino transmission
[20:48] <Redb3ard> haha, im more in philosophy mode right now, not engineer mode
[20:48] <Redb3ard> and, im not convinced that QE means there are virtual particles "transmitting" the info
[20:48] <toad_> okay, released it
[20:49] <toad_> Redb3ard: :)
[20:49] <toad_> Redb3ard: it doesn't
[20:49] <toad_> there are always particles at some point
[20:49] <Redb3ard> or that even if there are, that there arent microscopic wormholes taking them where they need to go
[20:49] <KenMan> oh no! now what have you unleashed ? err, i mean, 'released'
[20:49] <toad_> spooky action at a distance doesn't use particles, as far as we know - but it doesn't transmit information either!
[20:50] * KenMan feels the hair on his neck doing something spooky
[20:50] <greycat> toad_: but it does. You either pushed the button or you didn't.
[20:50] <greycat> toad_: that's a bit.
[20:50] * toad_ changes topic to 'Upgrade to 5091 (5089: mandatory reset, 5090: RNF bug fixed!) | Unstable: Upgrade to 60189 | Logs: http://newton.matcmp.ncc.edu/~lockej/freenet/chanlog/ | #freenet-politics and #freenet-chat are available for offtopic discussions'
[20:50] <toad_> greycat: eh?
[20:50] <KenMan> oh, it was just a spider, n/m
[20:50] <toad_> how can you tell that?
[20:50] <Redb3ard> greycat, its more complicated than that
[20:50] <toad_> okay
[20:50] <toad_> now I can look at KenMan's mail
[20:51] <Redb3ard> the only thing you can do for sure, is cause the particles to unentangle. hell, its tough to not cause them to unentangle
[20:51] <Redb3ard> more so, you only learn the state of the other, through your own particle... which doesnt mean alot
[20:51] <Redb3ard> you can sort of violate uncertainty, maybe
[20:51] <Redb3ard> but they argue about that too
[20:52] * leex-home-goings is now known as a-sleeping
[20:52] <greycat> toad_: take 9 entangled pairs, and separate them. You have 9 buttons, and the other guy has 9 lights. You designate the first 8 of them as the data. The 9th means "message transmitted". So you turn on some of the first 8, and press the 9th one to indicate that you're done with your message. When the receiver sees that 9th light go on, he reads the other lights to see which bits are on and which are off.
[20:52] <Redb3ard> what you need, is a way to affect one, without causing it to unentangle
[20:52] * a-sleeping is now known as leex-sleeping
[20:52] <Redb3ard> then, when they do, you can see how the other guy "pushed" his pairmate
[20:52] <toad_> Redb3ard: ummm
[20:52] <toad_> whatever
[20:52] <leex-sleeping> making my head hurt
[20:52] <leex-sleeping> nn
[20:53] <greycat> Redb3ard: now you're way off the deep end.
[20:53] <toad_> if somebody who actually knows about this stuff says he's sent data via spooky action, i'd be amazed
[20:53] <Redb3ard> like i said, im more concerned with the philosophical stuff, the applications... not the details
[20:53] <toad_> leex-sleeping: get 60188, go back to the normal unstable net
[20:53] <leex-sleeping> ok
[20:53] <Redb3ard> greycat, you have the pair of particles
[20:53] <toad_> Redb3ard: I'm concerned with the details because the devil (the fact that it's not implementable e.g.) is always in the details
[20:53] <toad_> :)
[20:53] <Redb3ard> you seperate them
[20:53] <toad_> KenMan: I got your pic
[20:53] <Redb3ard> i can cause mine to disentangle, not much else
[20:54] <greycat> Redb3ard: correct. Continue.
[20:54] <Redb3ard> it may not even be clear to you that they did disentangle
[20:54] <toad_> Redb3ard: can you reliably detect the time of disentanglement?
[20:54] <toad_> or even whether they have disentangled?
[20:54] <toad_> at the receiving end?
[20:54] <Redb3ard> if i try to "push" mine, so that when they disentangle, yours is also "pushed"
[20:54] <toad_> KenMan: i like the pic
[20:54] <Redb3ard> then we could communicate
[20:54] <greycat> I'm presuming that the disentanglement is detectable, but not necessarily time-able.
[20:54] <Redb3ard> but as soon as i try, they disentangle
[20:54] <toad_> I'm not convinced it's detectable without a backchannel
[20:55] <toad_> or even then
[20:55] <Redb3ard> so, for the most part, it just hints at communication
[20:55] <toad_> Redb3ard: if you can send ANYTHING, you can send data
[20:55] <Redb3ard> they can entangle lots of stuff now... no one has managed communication yet
[20:55] <toad_> but I doubt you can
[20:55] <greycat> Redb3ard: well the whole *concept* is that when one of the particles is observed, the other one drops out of the quantum superstate and takes on a definitive value.
[20:55] <Redb3ard> greycat, true
[20:56] <Redb3ard> now, how do you know that the other one has a def. value, without observing it?
[20:56] <leex-sleeping> Max: 0.0034227038
[20:56] <leex-sleeping> Most successful: 1
[20:56] <greycat> How you can detect that your particle has dropped out of super-state without observing it yourself and therefore screwing up the *other* one is not clear to me, however. I'm no physicist.
[20:56] <leex-sleeping> ffs
[20:56] <Redb3ard> and if you observe it, well then it causes the drop
[20:56] <Redb3ard> yeh
[20:56] <greycat> so... get that demo up :)
[20:56] <Redb3ard> i hope someone figures it out though
[20:56] <toad_> KenMan: okay so the problem in your opinion is the local requests which bypass most (not all) of the rate limiting system?
[20:56] <leex-sleeping> toad_ 60190 is laties
[20:56] <Redb3ard> i have some cool schemes on how to use it
[20:56] <toad_> leex-sleeping: yes
[20:57] <toad_> brb
[20:57] <leex-sleeping> (02:16:33) (toad_) leex-sleeping: get 60188, go back to the normal unstable net <<<
[20:57] <leex-sleeping> back to norm any way
[20:57] <toad_> KenMan: is the intention of that diagram to show that local requests are the main problem?
[20:57] <leex-sleeping> i a, now sleeping
[20:58] <toad_> I think there's probably some measurement problem with propagation... but I don't know what it is...
[20:58] <KenMan> naw, just a picture on which to cogitate
[20:58] <toad_> well, why did the experimental network fail?
[20:58] <KenMan> yeah, the transparent pipe is a complaint, but you already understand that component.
[20:59] <toad_> the essence of the problem is this:
[20:59] <toad_> the measurement problem
[20:59] <toad_> if incoming requests goes down, load due to propagation goes down!
[20:59] <toad_> err
[21:00] <toad_> goes UP!
[21:00] <toad_> the problem is that high load will cause incoming requests to go down
[21:00] <KenMan> feedback from RNFs caused a meltdown ? i don't know. I am back in stable, recording times of arrival for incoming data requests.
[21:00] <toad_> also if we send zero requests for a while the outgoing reqs won't change
[21:00] <KenMan> how did you compute propagation load again ?
[21:01] <toad_> propload = outgoing request rate / incoming request rate
[21:01] <toad_> it seems the obvious thing
[21:01] <toad_> but it's catastrophic
[21:01] <KenMan> you don't have to balance ins and outs exactly with high precision, but you DO need to recognize if we are reaching 95% average route backoff
[21:02] <KenMan> and you have to count ALL the requests to be routed (and rerouted)
[21:02] <toad_> well, that's optimizing routing through optimizing backoff
[21:02] <toad_> that causes freezeups
[21:02] <toad_> probably
[21:02] <KenMan> not if it is done properly ;)
[21:02] <KenMan> which i believe IS possible
[21:02] <toad_> how?
[21:03] <KenMan> oh, just fix the left screw, and you are all done.
[21:03] <toad_> which left screw? there are thousands of screws...
[21:03] <KenMan> I'm not gonna spew forth 10,000 words in an attempt to explain something that I don't know everything about.
[21:03] <toad_> uh, so where does that leave ME?!
[21:03] <KenMan> or, that i don't know *enough* about. I don't know the code well enough.
[21:03] <toad_> US even!
[21:04] <KenMan> well, tell me what and where you do rate limiting for starters...
[21:04] <leex-sleeping> lol
[21:04] <toad_> okay
[21:04] <leex-sleeping> fs i still here
[21:04] <toad_> what do you need to know?
[21:04] <KenMan> I know you have TDRAs for individual incoming MRIs
[21:04] <KenMan> sorry, TDRAs for individual outgoing MRIs
[21:05] <toad_> hmmm... i do?
[21:05] <KenMan> i think so, some kind of RA anyway
[21:05] <KenMan> in PeerHandler, isn't it ?
[21:05] <KenMan> this is how you maintain 100 different outgoing MRIs
[21:05] <toad_> we have some sort of RA for the number of reqs received from this particular node
[21:05] <toad_> which we use to calculate the MRI for that node
[21:06] <KenMan> it takes in a 'globalQuota' measurement, which is itself some RA...
[21:06] <toad_> derived from one
[21:06] <KenMan> i'm trying to spot all the related RAs involved with the calculation of outgoing MRIs, and probably I need to learn globalQuota as well.
[21:07] <toad_> number of reqs received from this node is derived from an RA, it's actually an Extrapolatingthingy
[21:07] <toad_> globalQuota is essentially incoming request rate / load
[21:07] <KenMan> cool, so you can say requestor X is pumping at Y qph , then.
[21:07] <toad_> averaged to smooth it
[21:07] <toad_> yeah
[21:07] <toad_> brb
[21:08] <KenMan> right now, i'm trying to finish drawing all the relevant pipes in my visual thought model
[21:08] <KenMan> I think I know them all, and will send you an updated pipechart
[21:09] <greycat> Redb3ard: if you're really in a philosophical mood, I'd suggest you go read LeGuin's "The Dispossessed".
[21:09] <KenMan> do you have any discomfort in my separation of 'Rate Control' and 'NGR' ? any alternative rep ? you agree they function hand-in-hand as a logical multiplexing unit ?
[21:10] <toad_> KenMan: hmm?
[21:10] * KenMan decides to avoid the ongoing discussion related to a specific rate limiting impl tonight :)
[21:10] <toad_> heh
[21:11] <KenMan> refer to my graphical model
[21:11] <toad_> well yeah, they're sort of opposing forces
[21:11] <KenMan> but they affect each other.
[21:11] <toad_> somewhat
[21:11] <KenMan> and have to work together to produce good results
[21:11] <toad_> rate control affects what routes NGRouting can actually route to
[21:12] <KenMan> yes, good observation
[21:12] <toad_> there is little cooperation beyond that, they just do their own thing
[21:13] <KenMan> well, I don't wish to talk in java-code, but the rates for routing live *inside* NGR today...
[21:13] <toad_> well, technically, StandardNodeEstimator has isAvailable()
[21:13] <toad_> but it could as easily be elsewhere
[21:13] <KenMan> and I drew the little array inside the RateControl box...
[21:13] <toad_> it's a self contained unit
[21:13] <KenMan> okay. good.
[21:13] <toad_> it's part of rate control logically
[21:13] <toad_> even if it's implemented in SNE
[21:14] <KenMan> I'm collecting arrival times to see if they are normally distributed about an average, but of course, I am trying to measure something that FRED is trying to manipulate
[21:14] <toad_> the question remains: (how) can we make propagation better?
[21:14] <KenMan> so, I need to choose stretches of time when qpm is relatively level
[21:15] <KenMan> I am confident that we can, just not all in one day...
[21:15] <toad_> well what's the next step?
[21:15] <toad_> it looks to me like the propagation formula i was using is crap
[21:16] <toad_> so what we want instead is to target some backed off fraction
[21:16] <toad_> e.g. 50%
[21:16] <KenMan> so toss it for now, you scrambled it together in a hurry anyway
[21:16] <KenMan> or something.
[21:16] <toad_> KenMan: better than doing nothing, and it produced some good experimental data
[21:16] <KenMan> certainly
[21:17] <KenMan> I think 50% can be achieved with a separate model that exists inside my head, but... translation is the hard part.
[21:17] <toad_> all we need to do is make propagation load = backed off fraction * 2, right?
[21:17] <KenMan> Probably, with an entirely different model, any level can be targetted
[21:17] <KenMan> i think that is an interesting approach.
[21:17] <toad_> maybe
[21:17] <toad_> I doubt it
[21:17] <toad_> imho it will lead to freezedown if you target too low
[21:17] <KenMan> you have to keep the full picture in mind, when asking why freezedown happens...
[21:18] <KenMan> I know it to be partially due to the high rate of RNFs (thus multiplied queries)
[21:18] <KenMan> i want to know how many RNFs come back without a single routing (inside the next hop), how many got routed once, etc.
[21:19] <toad_> I know that retries make very little difference
[21:19] <toad_> one of us is wrong
[21:19] <toad_> I have explained over and over and over why retries make no difference and you continue to ignore me
[21:20] <KenMan> do you agree that a retry requires a single item's worth of outgoing capacity ?
[21:20] <KenMan> on the second try
[21:20] <toad_> yes
[21:20] <KenMan> and again on the third (should there be one)
[21:20] <toad_> but it will balance out
[21:20] <toad_> it doesn't cost any more globally than an extra hop
[21:21] <KenMan> I am tunnel-vision focused on how it affects one node, possible two in a routing pair
[21:21] <KenMan> i don't give a hoot about global at the moment
[21:22] <KenMan> i use "global" to mean the entire network, perhaps you mean something else...
[21:22] <toad_> well, the point is, if we retry on node A, then we don't get to node C later on in the chain
[21:22] <toad_> the books balance
[21:23] <toad_> admittedly the traffic level would be lower if we never retried
[21:23] <toad_> but then so would the psuccess
[21:23] <KenMan> if I RNF on 40% of my queries, i make the assumption that my routes upchuck somewhere around 40% of my outgoings. That is heavy multiplication for a single sFC
[21:23] <KenMan> MY book does NOT balance
[21:23] <toad_> KenMan: okay, so how do we fix it?
[21:23] <toad_> is it queueing?
[21:24] <toad_> without queueing, we are going to get RNFs, if our in traffic equals out out traffic. from simulations.
[21:24] <KenMan> I plan to be able to send out 1000 queries. So I try to take in 1000. But 400 bounce back at me. Now I still have 1000 to route, but I've used up 40% of my out-capacity without results.
[21:24] <toad_> okay
[21:24] <KenMan> we will always get *some* RNFs, but 40% is just way too painful
[21:25] <toad_> well, the sims say 35% with the current model...
[21:25] <toad_> which is suspiciously close
[21:25] <KenMan> good enough
[21:25] <toad_> 25% if I take out the individual-interval-check
[21:25] <KenMan> my question is, of that 35%, did 20% come back due to route exhaustion without making any progress whatsoever ??
[21:25] <toad_> surprising really considering it's a rather crude model
[21:26] <KenMan> most likely coincidence, but maybe not
[21:26] <toad_> KenMan: the simulation is that 25%/35% (depending) are exhausted
[21:27] <KenMan> let me define pure exhaustion versus plain exhaustion - pure exhaustion means I took in a query. Because I had 0 routes open at that instant, I had to send it back immediately.
[21:27] <toad_> KenMan: that's what the simulation was simulating
[21:27] <KenMan> Plain exhaustion means I got to sent it out once or more, but eventually that same query hit 0 routes available and it had to go back as RNF.
[21:27] <toad_> it found 25% were purely exhausted
[21:27] <toad_> it assumed success if the query got sent out
[21:27] <KenMan> ouch. That's really really crummy then.
[21:28] <toad_> 35% if we check the interval
[21:28] <toad_> yup
[21:28] <KenMan> okay. stick with the 25% level.
[21:28] <toad_> this is assuming that the incoming requests level is equal to the total of the outgoing request capacities
[21:28] <KenMan> If 75% get though, I will use flaky math to say each one has a 75% of making it though.
[21:28] <toad_> if you reduce it to half that, then you can make it 5%
[21:28] <toad_> however, that probably means freezeup
[21:28] <KenMan> not just yet.
[21:29] <toad_> KenMan: hmm?
[21:29] <toad_> in any case as a practical matter we can't make in == out
[21:29] <KenMan> what does .75 ^ 20 work out to be ? ie. can any route actually reach a straight forward path of 20 hops, or are they much more likely to be swimming in cycles/circles ?
[21:29] <toad_> as we saw on the testnet
[21:30] <toad_> 0.3% :)
[21:30] <KenMan> we saw that a rapidly constructed algorithm didn't work out. I'm not trying to be mean. Just making a slightly subjective observation.
[21:30] <toad_> KenMan: well, as I said, we need a different load measurement
[21:30] <toad_> if we're going to do it through the load system
[21:31] <toad_> very likely the way forward would be to target a specific level of backoff
[21:31] <toad_> but we will still get a fair number of pure RNFs... depending on the RT size i suppose..
[21:32] <toad_> well if we target 50%, and we have 100 conns, then it's unlikely we'll get a pure RNF, right?
[21:32] <KenMan> i was going to suggest you could use the percent backoff as an input for load levelling... but I am not excited about the prospects of it working out.
[21:32] <KenMan> why do you think 100 are necessary ? this is important...
[21:33] <KenMan> why can't we build a model that works as well at 10 ?
[21:33] <toad_> KenMan: if we have 2, and we target 50%, then we'll get lots of pure RNFs
[21:33] <KenMan> okay, so a new way to state the problem is that we need to have routes coming available at a somewhat regular rate...
[21:33] * ninja_ (~im10ninja@3ffe:bc0:8000:0:0:0:0:3fab) Quit (Remote closed the connection)
[21:34] <toad_> KenMan: huh?
[21:34] * ninja_ (~im10ninja@3ffe:bc0:8000:0:0:0:0:3fab) has joined #freenet
[21:34] <KenMan> routes becoming available
[21:34] <toad_> no
[21:34] <toad_> we need to have a reasonable number of available routes at every specific time when a request comes in
[21:34] <toad_> i.e. at every point in time full stop
[21:35] <toad_> since incoming queries are random
[21:35] <KenMan> yes, and if we consume them at a regular rate, doesn't that imply what i said above
[21:35] <KenMan> here we are butting heads. You are right that they arrive somewhat randomly.
[21:35] <toad_> okay, fair enough
[21:35] <toad_> but it's insufficient
[21:35] <toad_> my criterion is a superset of yours
[21:36] <KenMan> lets identify what things we can control , and what we cant. Clearly route attributes are not under our control, beyond "number of routes"
[21:36] <KenMan> ie, we can't ask for a specific MRI ("level of resource allocation")
[21:37] <toad_> the exact timing of incoming queries is also beyond our control
[21:37] <KenMan> yes and no.
[21:37] <KenMan> the exact timing, yes. the average, however, we CAN control.
[21:37] <toad_> the average number of incoming queries in a period is beyond our control
[21:37] <toad_> errr
[21:37] <KenMan> then the game is completely lost, and we cannot ever win
[21:37] <toad_> the average we can constrain
[21:37] <KenMan> yes.
[21:37] <toad_> we can't control the minimum but we can control the maximum
[21:37] <toad_> ok
[21:38] <KenMan> in fact we could outright control the average incoming, but currently we constrain the incoming
[21:38] <mikeDOTd> An unrelated observation: my Current probability of an inbound request causing a transfer outwards = 2.3%. Doesn't that seem excessivly "good"?
[21:38] <KenMan> mikeDOTd: is retrospect, it far outperforms what we had 6 months ago...
[21:39] <KenMan> :p
[21:39] <toad_> we could perhaps ensure we have hardly any RNFs, by ensuring that there's always at least N, N>0, routes available
[21:39] <toad_> unfortunately as we get pickier, we increase the MRI multiplication factor, and thus the chance of freezedown
[21:39] * ninja_ (~im10ninja@3ffe:bc0:8000:0:0:0:0:3fab) Quit (Nick collision from services.)
[21:39] <KenMan> good line of thought though
[21:40] * ninja_ (~im10ninja@3ffe:bc0:8000:0:0:0:0:3fab) has joined #freenet
[21:40] <KenMan> and you have hit upon one of the biggest factors in all this, NGR doesn't favor all routes equally
[21:40] <KenMan> in fact, it is rather imbalanced
[21:40] <toad_> afaics the only way we can deal with the randomness here is to implement some form of queueing, and if necessary take steps to reduce the MRI
[21:40] <mikeDOTd> KenMan: maybe i misunderstand it's meaning. Doesn't that suggest my node has 97.7% of all data requested of it?
[21:41] <toad_> KenMan: yes but I don't think distributing requests evenly to all routes produces good routing
[21:41] <toad_> mikeDOTd: no
[21:41] <KenMan> no, its meaning is that 2.3% are effective, the rest are failures...
[21:41] <toad_> mikeDOTd: 2.3% of requests are successful and cause a transfer back to the requestor
[21:41] <mikeDOTd> ah, i see. thanks
[21:42] <toad_> KenMan: thus we have two difficult things to do - 1. queueing, 2. fixed small file sizes. :(
[21:42] <toad_> and of course no guarantee of success afterwards :)
[21:42] <KenMan> i agree, NGR doesn't route well if balancing is forced upon it...
[21:42] <KenMan> that's the spirit :) !
[21:43] <KenMan> well, we have lots of intuitive reasons to hope that fixed keys will help...
[21:43] <toad_> we have however solved the main design problems with queueing
[21:43] <toad_> and fixed small file sizes is easy it's just a lot of work to implement transparent FEC and so on
[21:43] <KenMan> queueing is just a technical task, it will take effort to get it working properly. And it affects the "backoff panel of blinking lights" tremendously.
[21:44] <toad_> arguably 3. rewrite FCP subsystem and probably protocol so we constrain local requests within reason
[21:44] <KenMan> you need to order you priorities.
[21:44] <KenMan> ;)
[21:44] <toad_> well okay
[21:44] <toad_> IMHO, queueing won't make much difference on its own unless it can queue for a significant period of time
[21:44] <toad_> e.g. 30 seconds
[21:45] <toad_> this might be acceptable for a 1MB chunk
[21:45] <toad_> it won't be acceptable for a 64kB freesite page
[21:45] <KenMan> and perhaps tackle them one at a time, don't try to roll it altogether into one strungout series of releases
[21:45] <toad_> we can make it work better by reducing the MRI significantly via fixed small file sizes
[21:45] <toad_> 3. is priority 3 IMHO
[21:45] <toad_> i.e. the last thing to do
[21:46] <KenMan> cost / benefit examination for fixed size keys - so/so. For 3. the benefit is likely to be rather higher than you might suspect
[21:46] <toad_> KenMan: I dunno, I think fixed size keys are a pretty solid benefit
[21:46] <toad_> but it will take a while
[21:47] <toad_> queueing is interesting w.r.t. CBA
[21:47] <KenMan> then lets just go there, and come back to this stuff AFTER we have examined the full effects of fixed size keys first.
[21:47] <toad_> arguably since most traffic is 1MB chunks, if we scale the queueing time to the file size, then we may get somewhere
[21:47] <toad_> of course it means that smaller files get suckier routing
[21:48] <KenMan> meaning, implement it and see what happens. Keep the other issues in mind, but don't work on them until fixed keys are a known good improvement.
[21:48] <toad_> which is an effect we really don't want to have to deal with...
[21:48] <KenMan> but they all get better routing, just due to queueing...
[21:48] <toad_> KenMan: small keys don't get routed as well because they're not queued for as long
[21:48] <toad_> that's the argument
[21:48] <KenMan> my argument is that if they get queued at all, they get better routing
[21:49] <toad_> if the big keys get queued, this will result in less RNFs
[21:49] <toad_> so even the small keys will be routed better
[21:49] <toad_> yes
[21:49] <KenMan> okay - queueing, fixed keys, and localReq balancing - please pick ONE. any one. But only ONE.
[21:49] <toad_> well IMHO they are already in priority order above
[21:49] <toad_> queueing is the top job
[21:49] <toad_> fixed size keys can come later
[21:50] <KenMan> fixed keys needs a reset, queueing should not... true ??
[21:50] <toad_> local request balancing even further down the line
[21:50] <toad_> correct
[21:50] <KenMan> so queueing it is...
[21:50] <toad_> fixed keys needs a catastrophic reset
[21:50] <toad_> we need to be in a rather better position with the users to pull that off
[21:50] <toad_> without having expanded too much :)
[21:50] * KenMan wonders if he is helping toad by "managing him" :o
[21:50] <toad_> yes, lets do queueing
[21:51] <toad_> KenMan: if I didn't think you were worth talking to I'd tell you to sod off
[21:51] <KenMan> the best part is you can always say STFU...
[21:51] <KenMan> you won't get that benefit in a 'real' job...
[21:51] <toad_> I can do better than that, actually... I can revoke your CVS perms and kickban you :)
[21:51] <KenMan> so true ...
[21:52] <toad_> now, queueing
[21:52] <toad_> typical 1MB transfer runs at 3kB/sec
[21:52] <KenMan> now, go to bed. Aren't you at all tired ??
[21:52] <toad_> this means it takes 333 secs + 30 secs search time
[21:52] <toad_> so if we allow 20 secs per hop extra search time and assume its found in 10 hops, that increases latency by 200 secs...
[21:52] <toad_> hmmm
[21:52] <KenMan> bb5m
[21:53] <toad_> hmmm
[21:53] <toad_> make it 10 secs per hop
[21:53] <toad_> so 333 + 30 + 100 = 463 versus 363
[21:54] <toad_> that's reasonable imho
[21:54] <toad_> for 512k, you can make it 5 secs a hop
[21:54] <toad_> for 256k, 2.5 secs a hop
[21:54] <toad_> for 128k, 1.25 secs a hop...
[21:55] <toad_> and so on, if it's meaningful at all...
[21:57] <toad_> requestCompletionTime is actually closer to 800 seconds here...
[21:58] <toad_> which is odd considering the transfer rate...
[21:59] <KenMan> will you use a queue per requestor, or a single common queue ?
[22:00] <Zorix> oo 5091
[22:01] <KenMan> wow, this "half the sugar" soda, which still uses sugar, instead of aspatame, is not bad. Sorta diet, but without the nasty phenylketoneurics
[22:01] * KenMan will have to examine the label more closely
[22:02] <toad_> :)
[22:03] <toad_> what was the scheme? We queue requests for N millis, at t=0, they will only go to their first choice, and at t=N, they will go to their worst choice...
[22:03] <toad_> or t>N
[22:03] <KenMan> you know, deciding to reject some things on the queue would be equivalent to hitting route exhaustion - so send back RNFs if any items wind up getting too old (or the queue gets too long)
[22:03] <toad_> N=10 secs for 1MB files, 5 for 512kB, 2.5 for 256kB...
[22:04] <toad_> KenMan: yes, we will still get RNFs
[22:04] <toad_> and the other thing?
[22:04] <KenMan> other thing ?
[22:04] * greycat (rfc1413@wooledge.org) Quit ("Lick Bush in '04")
[22:04] <toad_> of all the available requests to send to a node, we should send the one with the best estimate
[22:05] <toad_> if there are multiple nodes available, we should start by sending the {request,node} pair with the best estimate
[22:05] <toad_> then update which are available, and repeat until no more nodes are available or no requests are queued
[22:05] <toad_> that seems just about implementable
[22:05] <KenMan> sounds good. You will have a matrix of available routes, and available queries...
[22:05] <toad_> KenMan: yup
[22:05] <toad_> all the values are precomputed and we knock out a column at a time
[22:06] <KenMan> nice
[22:06] <toad_> probably by setting all the values down it to infinity :)
[22:06] <KenMan> except we will do 10x the routing calculations. Thank goodness for Iakin's blessing !!
[22:06] <toad_> KenMan: no
[22:06] <toad_> we already do all the routing calculations involved!
[22:06] <toad_> we don't need to update them
[22:07] <KenMan> oh, good point. I think I see what you mean.
[22:07] <toad_> we just use the existing ones on the NGRouting
[22:07] <KenMan> each request gets a Routing object, right ?
[22:07] <toad_> calculated at route() time
[22:07] <toad_> yup
[22:07] <toad_> full of Estimates for every node in the RT
[22:07] <KenMan> gotcha :)
[22:07] <toad_> which get deleted when they're used
[22:08] <toad_> every so often we should probably send a QueryRestarted back, but with the queue times so short it might not be necessary
[22:08] <toad_> speaking of which...
[22:08] <toad_> what IS acceptable overhead?
[22:08] <KenMan> you will set an age limit on the queue then ? and are you speaking of a single common queue ?
[22:09] <KenMan> or one per requestor ?
[22:09] <toad_> KenMan: common queue
[22:09] <toad_> between all requests and all nodes
[22:09] <toad_> now, age limit...
[22:09] <KenMan> you suggested that longer queue times might be acceptable for larger keys (fixed keys will negate that work later on)
[22:09] <toad_> if it's queued for too long, we want to RNF it, right?
[22:09] <KenMan> yes
[22:10] <toad_> so we'll have to have 2 phases
[22:10] <toad_> phase 1, we are becoming less picky
[22:10] <toad_> phase 2, we are waiting to be timed out
[22:10] <toad_> or do we only have phase 1, and if our worst choice isn't available at the final instant, tough, RNF?
[22:11] <KenMan> just to get inside your thoughts, what general frame of max queue time are you thinking ? 5 seconds ? 30 seconds ? more ?
[22:11] <toad_> KenMan: probably about 10 seconds on big files
[22:11] <toad_> or maybe 5, not sure
[22:11] <toad_> hmmm
[22:11] <toad_> well I was suggesting a ramp-up time of 10 seconds above
[22:12] <toad_> if we say that we RNF if having reached the bottom of the ramp, we still can't be routed to ANYONE?
[22:12] <toad_> or do we want a period during which we are queued in case somebody becomes available?
[22:12] <toad_> but we'll accept being routed to anyone ?
[22:12] <toad_> we can't accept being routed to anyone from the starting point because that would mess routing up...
[22:13] <toad_> or would it?
[22:13] <toad_> actually, perhaps, on average, it wouldn't
[22:13] <toad_> do you see what i'm saying?
[22:13] <KenMan> assume that the worst 25% (or some amount) of routes are not even worth trying. This conserves capacity
[22:13] <toad_> if BadNode is the only node available, and we're the only request queued, then we get sent to BadNode immediately
[22:13] <KenMan> but, depending on NGR, it may freeze out (tend to ignore) certain routes :(
[22:13] <toad_> but more often we will have several requests to send...
[22:14] <toad_> oh I see what the problem is
[22:14] <toad_> how about, we extend this matrix of nodes against requests
[22:14] <toad_> to include nodes that will be available soon (within the lifetime of the corresponding requests)
[22:14] <KenMan> routes against requests, you mean ?
[22:15] <KenMan> yes
[22:15] <toad_> we then pick the best requests as above, but including the nodes that Will Soon Be Available, and knocking out the involved nodes and requests
[22:15] <KenMan> you are allowed to see into the future, of course, with a little extra coding :)
[22:15] <toad_> so that then, the top choices for the Soon Available nodes will get their requests even though they have to wait
[22:16] <KenMan> that would preserve fairness in respect to the order in which they arrive, i suppose.
[22:16] <toad_> so that the best requests don't get sent immediately to crappy nodes
[22:16] <toad_> KenMan: no, I'm completely ignoring that in favour of sending good nodes good requests
[22:16] <KenMan> in other words, you get a request best suited to route F. You assign it to soon-to-be-open route F, yet then a better request comes in.
[22:16] <KenMan> Which do you pick ?
[22:17] <KenMan> and how must you code ?
[22:17] <toad_> KenMan: hmm?
[22:17] <KenMan> assume that only one can make it down the F pipe
[22:17] <toad_> okay
[22:17] <toad_> I get a request best suited to soon-to-be-available F
[22:17] <toad_> it's also the top request for now-available B
[22:17] <toad_> and C
[22:17] <toad_> and the second top for D
[22:18] <toad_> at that point, I keep it, allocating it to F, so I knock out the row and the column - the node and the request
[22:18] <toad_> we then send the second best choices to B and C
[22:18] <toad_> and the first to D
[22:18] <toad_> so now everyone's backed off
[22:18] <toad_> then a better request comes in for F
[22:18] <toad_> then F becomes available
[22:18] <KenMan> oh man, this makes my head hurt, but you seem to know what you mean ;)
[22:19] <toad_> we send the better request to F, because F is a good node and gets its first choice
[22:19] <toad_> our previous request is still queued
[22:19] <toad_> it can't possibly be sent to F because F's MRI is 300 seconds or so
[22:19] <toad_> so when A becomes available, we send it to A
[22:19] <toad_> KenMan: you follow, more or less?
[22:20] * KenMan actually followed that path, with multiple readings...
[22:20] <toad_> KenMan: does it make sense?
[22:20] <KenMan> i think so, I am confused about your matrix. Explain it just a little more.
[22:20] <toad_> okay
[22:20] <KenMan> how you use it. Routes versus queries
[22:21] <toad_> rows are nodes which either are available now, or will soon be available
[22:21] <toad_> columns are requests
[22:21] <toad_> we find the point on the matrix with the lowest estimate
[22:21] <toad_> we allocate it
[22:21] <toad_> i.e. if the node is available right now, we send the request
[22:21] <KenMan> row headings are requests, and column headings are routes, right ?
[22:21] <toad_> whether or not that is the case, we blank out both the row and the column
[22:21] <toad_> KenMan: yes
[22:21] <toad_> now we repeat
[22:22] <KenMan> let me scribble it on paper...
[22:22] <toad_> until we run out of available nodes or requests
[22:22] <toad_> this is a convergence of a whole bunch of ideas that have been floating around for a while...
[22:23] <toad_> some cells will not be occupied because they've been knocked out as above
[22:23] <toad_> they get a value of Double.+infinity
[22:24] <KenMan> okay, I am with you, and am glad I don't have to translate it into software :o
[22:24] <toad_> all the estimates are immediately and freely available from the Routing objects
[22:24] <toad_> KenMan: :)
[22:24] <KenMan> are the assignments kept in a separate list
[22:24] <toad_> KenMan: assignments?
[22:24] <toad_> yeah, the code will be fun
[22:24] <KenMan> the 'allocations' of Q1 to R5
[22:24] <toad_> sometimes I can get a bit blocked on major architectural stuff :)
[22:24] <KenMan> queries and routes
[22:24] <toad_> KenMan: no
[22:25] <toad_> immediately on deciding to send Q1 to R5, you:
[22:25] <KenMan> the contents of the matrix are estimates, right ?
[22:25] <toad_> - Send it!
[22:25] <toad_> (or at least arrange to send it)
[22:25] <toad_> - blank out all entries on R5 and Q1
[22:25] <KenMan> oh, so when you make an assignment, you inifintize out the other related cells .
[22:25] <toad_> not the intersection, the union
[22:25] <toad_> right
[22:26] <toad_> KenMan: before I implement it of course I'll have to summarize it for the list and for ian :)
[22:26] <KenMan> you realize, this enables you to hit the 100% thing we spoke of last night, don't you ?
[22:26] <KenMan> the time gap between backoffs
[22:26] <toad_> 100% thing?
[22:27] <KenMan> you will see some violations that are off by tiny amounts of time, due to transport effects
[22:27] * toad_ does not know whether it would enable us to propagate load for a given backoff target
[22:27] <toad_> KenMan: ahhh
[22:27] <KenMan> it allows you to utilize a route for 100% of the time, thus, some routes are never 'available' , as they are already occupied before they become available :)
[22:28] <toad_> well, there will be some delays in practice
[22:28] <toad_> due to sending overhead and so on
[22:28] <KenMan> well, throw it at the list ... and go to sleep :)
[22:28] <toad_> but I'll set it to use the average only and not the individual difference as well
[22:28] <toad_> before all this
[22:28] <toad_> okay
[22:28] <KenMan> or, go to sleep, and then throw a polished version at the list
[22:28] <toad_> I'm going to log this conversation
[22:28] <toad_> and then go to sleep
[22:28] <toad_> and then tomorrow write a polished version for the list
[22:30] <toad_> bbl zzz
[22:30] <KenMan> night
[22:30] <KenMan> and thanks
[22:30] * KenMan acknowledges that toad does all the real work :)
[22:30] <toad_> thanks for all your input
[22:31] * toad_ thinks queueing is probably a couple week job..
[22:31] <toad_> anyway we'll see
[22:31] <toad_> might be a month
[22:31] <KenMan> yup
[22:31] <toad_> bbl zzz
[22:32] <KenMan> the beauty of it is that toad can code and work against an existing network...
[22:32] <KenMan> requestors will just see RNFs and DNFs and maybe even some DF's , just like now...
[22:34] <Zorix> looks promising
[22:40] * verl agrees
[22:41] * verl changes topic to 'Upgrade to 5091 (5089: mandatory reset, 5090: RNF bug fixed!) | Unstable: Upgrade to 60190 | Logs: http://newton.matcmp.ncc.edu/~lockej/freenet/chanlog/ | #freenet-politics and #freenet-chat are available for offtopic discussions'
[22:45] * Hadaka (naked@naked.iki.fi) Quit (Read error: 110 (Connection timed out))
[22:48] <toad_> KenMan: something i just wrote down in my TODO to prevent me pondering it all night: We could include local requests in the queue, and give them very very long timeouts
[22:48] <toad_> bbl
[22:48] * KenMan grins
[22:50] * Hamled|Erp (~hamled@pool-68-162-123-225.phil.east.verizon.net) has joined #freenet
[22:51] <KenMan> my stable node has begun to settle (stabilize) after only 3 hours. Sometimes it is faster, sometimes it takes longer. Now to get some more peers...
[22:52] <KenMan> i set max to 80, but I only have 50 so far...
[22:53] <KenMan> Connections open (Inbound/Outbound/Limit) 50 (18/32/80) - normally I have more incoming than outgoing, which, i think, means that other nodes like my node.
[22:53] <Hamled|Erp> anyone here speak spanish?
[22:54] <KenMan> nadada
[22:55] <KenMan> kevloral used to, but he isn't around anymore... lesse, who else comes from that region ??
[22:56] <KenMan> there was at least one other spaniard, or someone from some close by island/territory ...
[22:56] <KenMan> who do the spaniards quibble with all the time, over fishing rights or somesuch nonsense ?
[22:57] * Hamled (~hamled@pool-68-163-62-23.phil.east.verizon.net) Quit (Read error: 60 (Operation timed out))
[22:58] <KenMan> portugal maybe ? oh well, Hamled left. Oh, no he didn't...
[22:58] <Hamled|Erp> eh
[22:58] <KenMan> I think someone portugeuse used to come around in this channel
[22:58] <Hamled|Erp> well... portuguese is a bit different...
[22:58] <Hamled|Erp> but it might work
[22:59] <KenMan> at any rate, i can't help in that language :( I'm just a stupid english-speaking, globe-dominating, american :o
[23:00] <KenMan> Hamled - can you speak english ?
[23:01] * KenMan couldn't resist making the joke !!
[23:03] <KenMan> ahh, it was someone from the canary islands. Yes. That was the other spanish-speaking person I've seen in here before.
[23:49] * Ash-Fox (Hal-9000@abd81.neoplus.adsl.tpnet.pl) Quit (Nick collision from services.)
[23:49] * goatee (~goatee@ip216-239-87-183.vif.net) Quit (Read error: 104 (Connection reset by peer))
[23:49] * salahx (salahx@sc1-24.217.174.147.charter-stl.com) Quit ()
[23:49] * Ash-Foxeus (Hal-9000@abd124.neoplus.adsl.tpnet.pl) has joined #FreeNET
[23:57] <hirvox> does anyone have any refs for build 60190?
These logs were automatically created by Jay Oliveri with his gimp hapi on irc.freenode.net.
