Timestamps are in GMT/BST.
[0:02] <TheSeeker> sorry, I was away getting food, which is good when you've neglected to eat in a couple days.
[0:25] * m[a]zzanet (~mazzanet@mazzanet.user) Quit (Client Quit)
[0:27] * mazzanet (~mazzanet@mazzanet.user) has joined #freenet
[0:52] * hobbs (~andrew@hobbs.user) has joined #freenet
[0:52] <hobbs> Is it me, or is sourceforge having issues right now? :)
[1:07] * maelztrom (~maelztrom@65-102-101-17.albq.qwest.net) has left #freenet
[1:09] * albo (~albo@pcp05055555pcs.sntafe01.nm.comcast.net) has joined #freenet
[1:10] <albo> the freenet-latest.tgz downloads at snails pace. help
[1:10] <hobbs> albo: looks like it's an issue on sourceforge's server, near as I can tell
[1:10] <hobbs> albo: but the rest of the site downloads just fine, so I'm not sure what's up
[1:50] * sdogi_ (~java@84-50-18-244-dsl.kjj.estpak.ee) has joined #freenet
[2:02] * sdogi (~java@84-50-18-37-dsl.kjj.estpak.ee) Quit (Read error: 110 (Connection timed out))
[2:21] * mazzanet (~mazzanet@mazzanet.user) Quit (Client Quit)
[2:22] * mazzanet (~mazzanet@mazzanet.user) has joined #freenet
[2:45] <KenMan> toad_: I've done it. I've incubated specialization! Rather than dropping DS keys with LRU, I'm dropping the 'least specialized key' - that with the greatest span from the next-left and next-right neighbor keys.
[2:48] <KenMan> It didn't notably improve success. Rather, it increased the key capacity of the network 10 fold (this is traction anyway)! Average HTL for success went UP, but, this is a breakthrough (for me anyway).
[2:48] <KenMan> With most sims I've tried, success was basically a function of redundancy :(
[2:49] <KenMan> Now I'm going to blend a completely arbitrary rule - "drop the least specialized key from the tail 1/3 of the LRU queue."
[2:49] <KenMan> I have to examine the contents of the stores for each node before I declare this new wrinkle to be a victory.
[2:50] <KenMan> DS contents are still determined by the peers. But remember, a hostile node is free to manipulate DS in any way it sees fit!
[2:51] <KenMan> So I don't consider this new tactic to be completely invalid...
[2:53] <KenMan> but, i suppose, you might. Anyway, this is something I consider worth investigating.
[2:54] <KenMan> I realize our current interpretation says "the set of routes=the potential DS", and as the potential DS is vastly greater than the actual DS, it *should* have a much greater degree of influence.
[2:58] <KenMan> So perhaps you should focus on finding a way to 'specialize' our set of routes. In theory, it should happen - most new connections will be obtained by those nodes providing the highest levels of success.
[2:59] <KenMan> assuming the pSuccess per peer is not level across the set of peers
[3:02] <KenMan> My current estimator remembers the hits from (and inserts to) each peer. It then takes the average distance of the five closest known successful keys, to produce the estimate.
[3:02] <KenMan> I think this is a fair approximation of the preNGR model.
[3:03] <KenMan> I'm storing the last 45 known keys (for each route), so 5/45 are used to produce the estimate. And it seems to be working.
[3:04] <KenMan> i'll make time to work on it some more tomorrow, anyway.
[3:07] <KenMan> s/obtained by those nodes/obtained via those routes/
[3:10] * sdogi (~java@84-50-19-113-dsl.kjj.estpak.ee) has joined #freenet
[3:21] * sdogi_ (~java@84-50-18-244-dsl.kjj.estpak.ee) Quit (Read error: 110 (Connection timed out))
[3:33] * Toast (~Toast@82-34-170-133.cable.ubr03.maid.blueyonder.co.uk) has joined #freenet
[3:33] * Ash-Fox (HAL-9000@aah236.neoplus.adsl.tpnet.pl) Quit (Client Quit)
[3:49] * sdogi_ (~java@84-50-19-236-dsl.kjj.estpak.ee) has joined #freenet
[3:49] * Toast (~Toast@82-34-170-133.cable.ubr03.maid.blueyonder.co.uk) Quit ()
[4:01] * hobbs (~andrew@hobbs.user) Quit ("Leaving")
[4:02] * sdogi (~java@84-50-19-113-dsl.kjj.estpak.ee) Quit (Read error: 110 (Connection timed out))
[4:56] * tyesinclair (~tyesincla@170.252.64.1) has joined #freenet
[4:57] * tyesinclair (~tyesincla@170.252.64.1) has left #freenet
[5:06] * Hory (~Miranda@82.76.81.56) has joined #FreeNet
[5:46] * toad_ (toad@toad-with-underline.active.supporter.pdpc) Quit (Remote closed the connection)
[6:54] * Hory (~Miranda@82.76.81.56) Quit ("CyberLore.net - Recommendations on the best games, freeware and websites.")
[7:00] * toad_ (toad@toad-with-underline.active.supporter.pdpc) has joined #freenet
[7:00] * ChanServ sets mode +o toad_
[7:07] * toad_ (toad@toad-with-underline.active.supporter.pdpc) Quit (Remote closed the connection)
[7:20] * d-ArkAngel (~Rob@cpc2-midd2-5-0-cust16.midd.cable.ntl.com) has joined #freenet
[7:43] * toad_ (toad@toad-with-underline.active.supporter.pdpc) has joined #freenet
[7:43] * ChanServ sets mode +o toad_
[7:58] * sdogi (~java@84-50-18-103-dsl.kjj.estpak.ee) has joined #freenet
[8:00] * d-ArkAngel (~Rob@cpc2-midd2-5-0-cust16.midd.cable.ntl.com) Quit (Read error: 54 (Connection reset by peer))
[8:10] * sdogi_ (~java@84-50-19-236-dsl.kjj.estpak.ee) Quit (Read error: 110 (Connection timed out))
[8:17] * intrus0 (~t3rra008@113-24-108.adsl.cust.tie.cl) has joined #freenet
[8:26] * locksy (~locksy@202.155.174.254) has joined #freenet
[8:32] * intrus0 (~t3rra008@113-24-108.adsl.cust.tie.cl) Quit ("chao! jotes rq xD")
[8:34] * d-ArkAngel (~Robert@213-131-104-86.onyx.net) has joined #freenet
[8:40] * robilad (~topic@mpiat2305.ag2.mpi-sb.mpg.de) has joined #freenet
[8:49] * robilad (~topic@mpiat2305.ag2.mpi-sb.mpg.de) Quit (Remote closed the connection)
[8:49] * FallingBuzzard (~srademach@207.152.112.129) has joined #freenet
[8:50] * robilad (~topic@mpiat2305.ag2.mpi-sb.mpg.de) has joined #freenet
[8:53] * Naked (naked@naked.iki.fi) has joined #freenet
[9:23] * Hory (~Miranda@82.76.81.56) has joined #FreeNet
[9:29] * Ash-Fox (HAL-9000@aah236.neoplus.adsl.tpnet.pl) has joined #FreeNET
[9:44] * locksy (~locksy@202.155.174.254) has left #freenet
[9:48] * locksy (~locksy@mrtg.sisgroup.com.au) has joined #freenet
[9:55] * toad__ (toad@82-32-17-1.cable.ubr03.azte.blueyonder.co.uk) has joined #freenet
[9:55] * locksy is leaving
[9:55] * dystopia (admin@gaia-wan0.ipv6.freeshell.bofx.net) Quit (zelazny.freenode.net irc.freenode.net)
[9:55] * Hory (~Miranda@82.76.81.56) Quit (zelazny.freenode.net irc.freenode.net)
[9:55] * toad_ (toad@toad-with-underline.active.supporter.pdpc) Quit (zelazny.freenode.net irc.freenode.net)
[9:55] * MrNaughty (MrNaughty@d199-126-25-30.abhsia.telus.net) Quit (zelazny.freenode.net irc.freenode.net)
[9:55] * Redb3ard (oylerjm@c-24-125-12-101.va.client2.attbi.com) Quit (zelazny.freenode.net irc.freenode.net)
[9:55] * KenMan (~dogzilla@pcp403292pcs.mntcrm01.md.comcast.net) Quit (zelazny.freenode.net irc.freenode.net)
[9:55] * locksy is now known as locksy_AFK
[9:58] * Hory (~Miranda@82.76.81.56) has joined #freenet
[9:58] * MrNaughty (MrNaughty@d199-126-25-30.abhsia.telus.net) has joined #freenet
[9:58] * Redb3ard (oylerjm@c-24-125-12-101.va.client2.attbi.com) has joined #freenet
[9:58] * KenMan (~dogzilla@pcp403292pcs.mntcrm01.md.comcast.net) has joined #freenet
[9:58] * dystopia (admin@gaia-wan0.ipv6.freeshell.bofx.net) has joined #freenet
[10:07] * locksy_AFK is now known as locksy
[10:07] * locksy is back
[10:09] * locksy is now known as locksy_AFK
[10:13] * locksy_AFK is now known as locksy
[10:17] * goatee (~goatee@ip216-239-84-60.vif.net) Quit ("Leaving")
[10:26] * goatee (~goatee@ip216-239-84-60.vif.net) has joined #freenet
[10:39] * robilad (~topic@mpiat2305.ag2.mpi-sb.mpg.de) Quit ("ChatZilla 0.9.52B [Mozilla rv:1.6/1]")
[10:39] * Redb3ard (oylerjm@c-24-125-12-101.va.client2.attbi.com) has left #freenet
[10:40] * mazzanet (~mazzanet@mazzanet.user) Quit (Client Quit)
[10:41] * mazzanet (~mazzanet@mazzanet.user) has joined #freenet
[10:50] * locksy (~locksy@mrtg.sisgroup.com.au) Quit ("Leaving IRC - dircproxy 1.0.5")
[10:50] * locksy (~locksy@mrtg.sisgroup.com.au) has joined #freenet
[10:51] * robilad (~topic@mpiat2305.ag2.mpi-sb.mpg.de) has joined #freenet
[10:52] * locksy is leaving
[10:52] * locksy is now known as locksy_AFK
[11:01] * guido^pe (~unknown@dsl-082-082-158-058.arcor-ip.net) has joined #freenet
[11:01] * locksy_AFK is now known as locksy
[11:01] * locksy is back
[11:03] * locksy is leaving
[11:03] * locksy is now known as locksy_AFK
[11:13] <toad__> hi
[11:13] * toad__ is now known as toad_
[11:13] <d-ArkAngel> hi toad
[11:13] <toad_> hi d-ArkAngel
[11:14] <toad_> <mule> anybody aware whether fcp needs client state over a sequence of connections? -- i can tell you categorically that it does not
[11:14] <d-ArkAngel> my server is finaly back up, tho I've had to reset your password, I'll mail you a copy of the new password when I get home this evening.
[11:15] <toad_> <hobbs> Is it me, or is sourceforge having issues right now? :) -- YES :<
[11:15] <d-ArkAngel> it's been updated to FC3, so I don't have java installed yet, tho I'll do that for you this evening too.
[11:15] <toad_> d-ArkAngel: did the .ssh survive? and did the server's ID change?
[11:15] <d-ArkAngel> new id.
[11:15] <d-ArkAngel> hdd failure.
[11:15] <d-ArkAngel> I managed to save the /home partition tho
[11:16] <toad_> [07:46] <KenMan> toad_: I've done it. I've incubated specialization! Rather than dropping DS keys with LRU, I'm dropping the 'least specialized key' - that with the greatest span from the next-left and next-right neighbor keys.
[11:16] <toad_> [07:49] <KenMan> It didn't notably improve success. Rather, it increased the key capacity of the network 10 fold (this is traction anyway)! Average HTL for success went UP, but, this is a breakthrough (for me anyway).
[11:16] <toad_> hmmm
[11:16] <d-ArkAngel> so I've got your files on my laptop at the moment, tho I had to bzip them.
[11:16] <toad_> KenMan: you are wrong. I tried dropping keys according to estimates, that is more or less the same thing
[11:16] <toad_> it means no keys will ever be dropped
[11:16] <toad_> and the whole network collapses because it can't drop unused content
[11:18] <d-ArkAngel> toad_ I wonder if you could combine the two methods in some beneficial way....
[11:18] <TheSeeker> hmm... "least specialized key not requested in at least X" ?
[11:18] <toad_> d-ArkAngel: perhaps
[11:18] <toad_> some combination of time/frequency of access and estimator
[11:19] <toad_> difficult to come up with something that isn't totally alchemical
[11:20] <TheSeeker> least requested key at least X away from our specialization point?
[11:20] * locksy_AFK is now known as locksy
[11:20] * locksy is back
[11:20] <toad_> also LRU is very nearly provable
[11:21] <toad_> LRU for connections is known to have some very good properties
[11:21] <d-ArkAngel> toad how about something really really simple, like alternating between LRU and furthest from specilisation.
[11:21] * Hory (~Miranda@82.76.81.56) Quit (Read error: 60 (Operation timed out))
[11:21] <d-ArkAngel> on a per drop basis...
[11:21] * mazzanet (~mazzanet@mazzanet.user) Quit (Client Quit)
[11:22] <d-ArkAngel> that way you clean up all the nodes on the edges of your specilisation, which will tend to narrow the store, but the LRU will tend to thin it...
[11:22] <d-ArkAngel> tho that probably doesn't so much deal with the problem as make it less predictable...
[11:23] * mazzanet (~mazzanet@mazzanet.user) has joined #freenet
[11:23] <toad_> it's all alchemy
[11:23] <d-ArkAngel> indeed
[11:24] <d-ArkAngel> I'm afraid I'm not any where near the level of mathmatics to look at provable solutions, so I'm an alchemist at heart :-)
[11:25] <d-ArkAngel> mind you alchemy isn't always a bad thing :-) it's just not provably a good thing :-)
[11:25] <d-ArkAngel> lol
[11:26] <toad_> ugh, JVM bugs affecting first time users... not good!
[11:26] <d-ArkAngel> sorry I'll stop muttering now :-)
[11:26] <d-ArkAngel> bugs with which jvm?
[11:26] <toad_> d-ArkAngel: there's a nice poster on that on one of the freesites... ian and told teach alchemy: a rough guide to winging it
[11:27] <KenMan> d-ArkAngel: yeah, all that muttering almost gave me hope.
[11:28] <d-ArkAngel> it's funny really, we're in the same situation as physicists really, we're just smashing lumps of matter together (our simulations) and the mathmaticians can tell us what it all means afterwards :-)
[11:29] <KenMan> my alchemical blend of LRU + "self-specializing" did not produce an alloy :( so smash, smash again :)
[11:29] <TheSeeker> I see alternating least used and least specialized as being far less alchemical than making up a more complex formula for when something should eb dropped and why...
[11:30] <TheSeeker> KenMan: are you attempting to try these ideas as they are being spoken? o.O
[11:32] <KenMan> oh come on, no one is THAT good ;)
[11:32] <d-ArkAngel> I'm sure soneone is :-) lucky bastard!
[11:33] <KenMan> and I'm just making it all up. I'm not really doing anything, just trying to make it sound like I am.
[11:33] <d-ArkAngel> aren't we all?
[11:33] <d-ArkAngel> :-)
[11:33] <KenMan> d-ArkAngel: we need bigger matter smashers, is all...
[11:34] <d-ArkAngel> there is no freenet it's all in the mind :-)
[11:34] <toad_> the Large Hadron Collider, perhaps...
[11:34] * d-ArkAngel feels guilty that he's not woked on the parallel sim for so long.
[11:34] * locksy is leaving
[11:34] * locksy is now known as locksy_AFK
[11:34] <d-ArkAngel> /ignore locksy
[11:35] * locksy_AFK is now known as locksy
[11:35] * locksy is back
[11:45] <KenMan> actually toad, i was hoping that my suggested experiment was generic enough that you might some day try it out. It is not precisely the same thing as dropping the key with the worst estimate.
[11:48] <KenMan> But i must figure out if it has grown something useful or not. I believe it is 'learning' , but it may just be some elaborate quirk :(
[11:50] <KenMan> 450 nodes, RT of 12, 10 hops max, 200 DS ... I'm looking at success up to 60% , average HTL=6.7 for resolved keys, yet the average redundancy factor is ONLY 1.5 !
[11:51] <KenMan> The network is storing 70K unique keys in only 90K storage slots, and it is finding more than half of them !!!
[11:52] <KenMan> okay, less than 1.5 ... 70K/90K = 1.28
[11:52] <TheSeeker> o.o
[11:53] <KenMan> 90K/70K ~= 1.28
[11:53] <toad_> KenMan: that's quite impressive yes
[11:53] <TheSeeker> that tried to go over my head, but just clipped me. ow *.o
[11:53] <toad_> is that with ongoing data churn?
[11:53] <toad_> in other words, can it deal with having to throw out old data?
[11:55] <toad_> oh, does anyone want to implement UP&P? I have no router to do it with..
[11:55] <toad_> and there's a REALLY nice library just been posted a link to on devl for it
[11:55] <toad_> I'll help as far as the freenet end goes
[11:55] <toad_> but I can't debug it without the necessary hardware
[11:55] <toad_> which I won't buy without personal approval from ian for the project refunding it :)
[11:57] <KenMan> toad - yeah, i'm doing 400 insert ops per 10K request ops. But now I have to get out the microscope and figure out what is happening.
[11:59] <KenMan> It almost sounds too good to be true. For instance, it could be (primarily) dropping those keys which are the newest insertions... thus making the stores very sludgey. I just don't know yet. But that is a plausible explanation.
[12:01] <KenMan> although there is pCaching, so that makes it even harder to guess what is going on.
[12:04] <KenMan> my money is on the sludgieness concept, but time will tell.
[12:05] <KenMan> still, even if it is sludging, there could be a way to combine "LRU with Glue(tm)" in a useful way
[12:07] * TLF (~francisco@58.Red-81-40-113.pooles.rima-tde.net) has joined #freenet
[12:08] <TheSeeker> GLUE(tm) ?
[12:10] * d-ArkAngel (~Robert@213-131-104-86.onyx.net) has left #freenet
[12:14] * Sugadude (~Sugadude@static-64-115-210-23.isp.broadviewnet.net) Quit (Remote closed the connection)
[12:15] * Sugadude (~Sugadude@216.17.101.36) has joined #freenet
[12:15] * robilad (~topic@mpiat2305.ag2.mpi-sb.mpg.de) has left #freenet
[12:16] <toad_> KenMan: huh?
[12:16] <toad_> KenMan: what exactly are you simulating?
[12:16] * Hory (~Miranda@82.76.81.56) has joined #FreeNet
[12:17] * toad_ will have to go out soon for a night of chatting to anarchists and geeks about life, the universe, wireless networking, and extorting money out of large charities and governments
[12:17] <toad_> and refurb'd PCs
[12:27] <KenMan> toad_: I can assure you that you would scoff at what I am doing, because it is different than what you are doing.
[12:27] <KenMan> It is supposed to be a reasonable representation of FreeNet routing with hops and caching and some form of peer-estimation and...
[12:28] <toad_> KenMan: it sounds interesting :)
[12:28] <toad_> if it can be tweaked to work significantly better than what we do now, then post a summary and some stats
[12:28] <KenMan> well, hopefully it becomes something useful. But, for now, that is all it is. "Interesting..."
[12:29] * Ash-Fox (HAL-9000@aah236.neoplus.adsl.tpnet.pl) Quit (Read error: 104 (Connection reset by peer))
[12:31] <KenMan> It will be a simple matter to apply this one concept to the java sim you built (before you broke) and see if it can help. Have you started rebuilding after the fire ?
[12:31] <toad_> http://news.zdnet.co.uk/software/developer/0,39020387,39179365,00.htm
[12:31] <toad_> hmmm
[12:31] <toad_> KenMan: I have started to build a whole new simulator
[12:31] <toad_> I haven't finished it yet
[12:31] <toad_> but now I must go
[12:32] <KenMan> cool, try not to filch off the old sim too much.
[12:32] <toad_> back tomorrow probably, maybe late tonight; talking to geeks and anarchists (www.bristolwireless.net, currently offline due to defacement i believe)
[12:32] <KenMan> have a good time !
[12:33] <toad_> well more vandalism than defacement; some idiot was quite thorough
[12:33] * goatee (~goatee@ip216-239-84-60.vif.net) Quit (zelazny.freenode.net irc.freenode.net)
[12:33] * Gunnar (~gunnar@34.80-203-209.nextgentel.com) Quit (zelazny.freenode.net irc.freenode.net)
[12:34] * goatee (~goatee@ip216-239-84-60.vif.net) has joined #freenet
[12:34] * Gunnar (~gunnar@34.80-203-209.nextgentel.com) has joined #freenet
[12:54] * Hadaka (naked@naked.iki.fi) Quit (Ping timeout: 14400 seconds)
[12:54] * Naked is now known as Hadaka
[13:49] * plixed (~plixed@i3ED62453.versanet.de) Quit (Read error: 104 (Connection reset by peer))
[13:53] * d-ArkAngel (~Rob@cpc2-midd2-5-0-cust16.midd.cable.ntl.com) has joined #freenet
[13:53] * d-ArkAngel (~Rob@cpc2-midd2-5-0-cust16.midd.cable.ntl.com) has left #freenet
[13:56] * Ash-Fox (HAL-9000@aah236.neoplus.adsl.tpnet.pl) has joined #FreeNET
[14:13] <KenMan> aaagh! well, FunkySim is demonstrating emergent behavior. Unfortunately, that behavior may be "the network as a whole converges on a single point in keyspace" :o
[14:13] <KenMan> so, time to get 'back to the old drawing board' :)
[14:14] <KenMan> picture a puppet, with the same number of strings as there are peers in the routing table. Who can pull the hardest ?
[14:15] <KenMan> time to start generating graphics (no animations yet!) and see what's really going on...
[14:16] * pittaman (~mistery_b@d51A410CA.kabel.telenet.be) has joined #freenet
[14:24] <KenMan> well, perhaps each node is converging on a unique point, but that is only a single unique point per node. Still, that beats a network which could only store 'A' keys ;)
[14:40] <KenMan> a bug, a boo. In my datastores too. Looks like I failed to adequately implement LRU.
[14:45] * T3chni (~T3chni@reshall-137-55.OIT.EDU) has joined #freenet
[14:47] <KenMan> the RAD fad makes me mad. It says bugs are to be had.
[15:02] * Redb3ard_ (oylerjm@c-24-125-12-101.va.client2.attbi.com) has joined #freenet
[15:04] * sdogi (~java@84-50-18-103-dsl.kjj.estpak.ee) Quit (Read error: 104 (Connection reset by peer))
[15:06] * sdogi (~java@84-50-17-197-dsl.kjj.estpak.ee) has joined #freenet
[15:17] * TLF (~francisco@58.Red-81-40-113.pooles.rima-tde.net) Quit ("Adi?s a todos y a todas, y que les vaya bien. || http://www.vitaliano.esp.cc")
[15:20] * nextgens (~nextgens@d80-170-86-232.cust.tele2.fr) has joined #freenet
[15:22] * TLF (~francisco@1.Red-81-40-113.pooles.rima-tde.net) has joined #freenet
[15:28] * Redb3ard_ (oylerjm@c-24-125-12-101.va.client2.attbi.com) Quit ("ircII EPIC4-2.1.1 -- Are we there yet?")
[15:30] <i2p_iip> <mule> i'm trying to send fcp traffic over a mix network to a number of freenet nodes and that seems to fail
[15:31] <i2p_iip> <mule> it works as long as fcp traffic is sent to one node only. however if it is addressed to more than one in sequence, it seems to fail
[15:31] <i2p_iip> <mule> anybody aware whether fcp needs client state over a sequence of connections?
[15:31] * pittaman (~mistery_b@d51A410CA.kabel.telenet.be) Quit (Client Quit)
[15:56] * TLF (~francisco@1.Red-81-40-113.pooles.rima-tde.net) Quit (Read error: 104 (Connection reset by peer))
[15:56] * jay (~jcl@198.38.10.200) has joined #freenet
[16:10] * Ash-Fox (HAL-9000@aah236.neoplus.adsl.tpnet.pl) Quit (Nick collision from services.)
[16:10] * Ash-Fox (HAL-9000@aba127.neoplus.adsl.tpnet.pl) has joined #FreeNET
[16:13] * TLF (~francisco@113.Red-81-40-113.pooles.rima-tde.net) has joined #freenet
[16:13] * jay (~jcl@198.38.10.200) Quit ()
[16:20] <nextgens> toad_> are you here ?
[16:22] * nextgens has read the ML and wonders about UP&P
[16:23] <nextgens> Will freenet be UP&P compliant before 0.6?
[16:24] * Ash-Fox (HAL-9000@aba127.neoplus.adsl.tpnet.pl) Quit (Nick collision from services.)
[16:24] * Ash-Fox (HAL-9000@abm207.neoplus.adsl.tpnet.pl) has joined #FreeNET
[16:25] <KenMan> <toad_> back tomorrow probably, maybe late tonight; talking to geeks and anarchists
[16:25] <KenMan> nextgens: try him again in about 3 hours...
[16:26] <nextgens> KenMan> will you worry about UP&P?
[16:26] <KenMan> not really. I mean, it matters, but the project seems to have bigger issues right now.
[16:27] <KenMan> I don't know, I'm not the one to implement it in any case. It may be gruesomely relevant at this time. I just don't know.
[16:27] <nextgens> KenMan> doesn't misconfigured nodes hurt the network?
[16:28] <KenMan> possibly. I don't have any way to measure the impact.
[16:28] <nextgens> ok, thx for having replied :)
[16:28] <KenMan> I am more concerned that the project may shrivel up and die if we don't fix the basics (good routing=success)
[16:29] <KenMan> adding UP&P may only permit more users to experience ~1% success, which isn't really all that good.
[16:30] <KenMan> I am extremely confident that it would NOT enhance pSuccess in any significant manner.
[16:31] <nextgens> do you think 'rotating' seednode could help (less overload on these nodes) and lead to a higher success ?
[16:32] <KenMan> yeah, it could. But it is unlikely to make the scale of improvement that is needed. Any progress is appreciated, but going from ~1% success to ~2% is not that helpful.
[16:33] <nextgens> :-D I agree
[16:34] <KenMan> Either something useful comes from toad doing simulations, or it doesn't. But that's where I choose to invest my limited hope-reserves these days.
[16:36] <KenMan> one thing is for certain - while sims are NOT real-world measurements, he gets much faster feedback on any changes he makes.
[16:36] <KenMan> So in theory it should allow learning & progress to occur faster...
[16:40] <KenMan> cool - now that I fixed the LRU snafu in my sim, I can dial in the degree of LRU -vs- Glue . Node DSs are developing spec faster, and so far have more than one point of spec.
[16:40] <KenMan> Ultimately I expect each one will converge to their own unique single spec point. I (still, years later) don't know if I like this.
[16:42] * nextgens (~nextgens@d80-170-86-232.cust.tele2.fr) Quit (Remote closed the connection)
[16:43] <KenMan> convergence of the DS for a single node around a single keyspace point is happening faster than I expected.
[16:43] * nextgens (~nextgens@d80-170-85-120.cust.tele2.fr) has joined #freenet
[17:06] <KenMan> yes, i've got something good happening here. I like what I'm seeing. Clear spec point(s) on each node, high success, hmm: what's missing ?
[17:07] <nextgens> KenMan> you're dreaming :-P
[17:07] <KenMan> ( I'm running sims in the land of "maybe" )
[17:08] <KenMan> it's not an actual node's performance.
[17:08] <KenMan> it's simply a model of what might be possible , is all
[17:09] <KenMan> what i need now is complete logging of every event, and tools to inspect all the data. Which is not unreasonable.
[17:12] <KenMan> and I'm still trying to derive a measure of specialization for a set of points. The best I've come up with in over a year's time is...
[17:13] <KenMan> take the standard deviation over the distances between each neighboring point. It is almost too simple.
[17:14] <KenMan> high stddev indicates higher specialization (or clustering) , i think
[17:16] <KenMan> right now I have 454x12x10x200 , with ~66% success and average key redundancy of 9 (10150/90800)
[17:16] <KenMan> average successful HTL is 5.2 (actual) hops , not counting the originating node as a hop
[17:16] * Markaci (~Markaci@CPE00508de53d8f-CM0011aea10c8a.cpe.net.cable.rogers.com) Quit ("Chaos, Panic, Disorder; My work here is done.")
[17:23] * T3chni (~T3chni@reshall-137-55.OIT.EDU) Quit (Read error: 104 (Connection reset by peer))
[17:28] * mazzanet (~mazzanet@mazzanet.user) Quit (Client Quit)
[17:28] * mazzanet (~mazzanet@mazzanet.user) has joined #freenet
[17:32] * FallingBuzzard (~srademach@207.152.112.129) Quit (Read error: 104 (Connection reset by peer))
[17:43] * Hory (~Miranda@82.76.81.56) Quit (Read error: 54 (Connection reset by peer))
[17:49] * TLF (~francisco@113.Red-81-40-113.pooles.rima-tde.net) Quit ("Adi?s a todos y a todas, y que les vaya bien. || http://www.vitaliano.esp.cc")
[17:56] * nextgens (~nextgens@d80-170-85-120.cust.tele2.fr) Quit (Remote closed the connection)
[17:57] * nextgens (~nextgens@d213-103-2-83.cust.tele2.fr) has joined #freenet
[18:01] * bluephile (bluephile@69-160-215-63.clvdoh.adelphia.net) has joined #freenet
[18:27] * mazzanet (~mazzanet@mazzanet.user) Quit (Client Quit)
[18:35] * Hory (~Miranda@82.76.81.56) has joined #FreeNet
[18:39] * linagee_ (~linagee@netblock-66-245-229-97.dslextreme.com) has joined #freenet
[18:56] * linagee (~linagee@netblock-66-245-229-97.dslextreme.com) Quit (Read error: 110 (Connection timed out))
[19:14] * nextgens (~nextgens@d213-103-2-83.cust.tele2.fr) Quit (""ya mum quits IRC"")
[19:32] * verl (verl@h152n2fls33o877.telia.com) Quit ()
[19:35] * verl (verl@h152n2fls33o877.telia.com) has joined #freenet
[19:51] * Hory (~Miranda@82.76.81.56) Quit ("CyberLore.net - Recommendations on the best games, freeware and websites.")
[19:54] <KenMan> cool, i can watch a datastore in real-time now. Cheap 'animation' at work:
[19:55] <KenMan> while true; do {fetch file}; {produce barcode image}; display -delay 3 {image}; done
[19:55] <KenMan> now i'm itching to produce a shareable, genuine animation file...
[19:56] <KenMan> everyone uses AVI thanks to windows, but I want to do .mng ... there is a mozilla/ffox plugin for it.
[20:49] * toad_ (toad@82-32-17-1.cable.ubr03.azte.blueyonder.co.uk) Quit (Remote closed the connection)
[20:56] * Redb3ard_ (oylerjm@c-24-125-12-101.va.client2.attbi.com) has joined #freenet
[20:56] * Redb3ard_ (oylerjm@c-24-125-12-101.va.client2.attbi.com) Quit (Remote closed the connection)
[21:27] * toad_ (toad@toad-with-underline.active.supporter.pdpc) has joined #freenet
[21:27] * ChanServ sets mode +o toad_
[21:48] <KenMan> well, now i am suspicious of things. Perhaps we are not using sufficient datastore sizes. Somehow I am not comfortable with anything *but* LRU for DS maintenance.
[21:48] * toad_ (toad@toad-with-underline.active.supporter.pdpc) Quit (Remote closed the connection)
[21:51] <KenMan> I am forcing some specialization to happen by trimming the DS selectively. And for stores of only 200 items, it appears to boost success... but i still don't understand it.
[22:06] <KenMan> argh! the sim has converged on a favored subset of keyspace, meaning key retention over time is a function of key value :(
[22:07] <KenMan> there appear to be two dominant points, but 4 random nodes exhibit roughly the same spec...
[22:28] * toad_ (toad@toad-with-underline.active.supporter.pdpc) has joined #freenet
[22:28] * ChanServ sets mode +o toad_
[22:30] <KenMan> what is up with the amphibious computer ? must be losin' power or something...
[22:41] <KenMan> Richard Angwin says things are calm... hmm!
[22:42] * jay (jcl@ool-18bf6dac.dyn.optonline.net) has joined #freenet
[22:42] <KenMan> hayyy-O !
[22:43] <jay> whaddup
[22:46] * kevlAway (~kevloral@CZ1-RAS-8-u-0179.du.onolab.com) Quit ("Peace and Protection 4.22")
[23:02] <TheSeeker> KenMan: use ogm? :D
[23:03] * guido^pe (~unknown@dsl-082-082-158-058.arcor-ip.net) Quit (Remote closed the connection)
[23:05] <jay> ogm?
[23:24] <TheSeeker> a container format that allows for streaming, and multiple audio/video/subtitle channels... much more flexible than the AVI container format.
[23:25] <TheSeeker> whereas mng is just an animated png ...
[23:26] <TheSeeker> I suppose mgn would work OK for the barcodes thing, but you could still get better (though lossy) compression with a video codec like xvid
[23:32] * mazzanet (~mazzanet@mazzanet.user) has joined #freenet
[23:43] <mikeeusa> hi
These logs were automatically created by Jay Oliveri with his gimp hapi on irc.freenode.net.