Timestamps are in GMT/BST.
[0:08] * mazzanet (~mazzanet@mazzanet.user) Quit (Client Quit)
[0:08] * yonkeltron (~yonkeltro@pcp04665066pcs.wilog501.pa.comcast.net) Quit (Remote closed the connection)
[0:09] * yonkeltron (~yonkeltro@pcp04665066pcs.wilog501.pa.comcast.net) has joined #freenet
[0:10] * mazzanet (~mazzanet@mazzanet.user) has joined #freenet
[0:18] <KenMan> helloo mazzanet !
[1:26] * MrNaughty (MrNaughty@d199-126-25-30.abhsia.telus.net) Quit ("\(^_^)/' No Soliciting!!! Unless you have legs way, way up and really, really big tits....")
[1:27] * MrNaughty (MrNaughty@d199-126-25-30.abhsia.telus.net) has joined #freenet
[2:17] * mazzanet (~mazzanet@mazzanet.user) Quit (Client Quit)
[2:17] * mazzanet (~mazzanet@mazzanet.user) has joined #freenet
[2:41] <KenMan> howdy mazzanet !
[2:47] <KenMan> is everything in full green bloom down your way ?
[2:49] * hobx_ (~chatzilla@ankh.math.chalmers.se) Quit (zelazny.freenode.net irc.freenode.net)
[2:55] * hobx_ (~chatzilla@ankh.math.chalmers.se) has joined #freenet
[2:59] * sanity (~ian@81-178-106-165.dsl.pipex.com) has joined #freenet
[3:51] * sanity (~ian@81-178-106-165.dsl.pipex.com) Quit ()
[4:24] * TLF (~francisco@36.Red-81-40-113.pooles.rima-tde.net) has joined #freenet
[4:38] * kevloral (~kevloral@CZ1-RAS-8-u-0179.du.onolab.com) has joined #freenet
[4:39] <kevloral> g'morning all
[4:58] * Elly|afk (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Read error: 110 (Connection timed out))
[5:05] * kevloral (~kevloral@CZ1-RAS-8-u-0179.du.onolab.com) Quit ("Peace and Protection 4.22")
[5:20] * kevloral (~kevloral@CZ1-RAS-8-u-0179.du.onolab.com) has joined #freenet
[5:40] * hobx_ (~chatzilla@ankh.math.chalmers.se) Quit (Remote closed the connection)
[5:56] <TheSeeker> freenet crashed. do the devs have any use for hotspot error dumps?
[6:01] <TheSeeker> I guess that would fall under "java crashed" instead of "freenet crashed" ... as there was no indication of a crash in the freenet log...
[6:14] * Ash-Fox (HAL-9000@ada87.neoplus.adsl.tpnet.pl) Quit (Nick collision from services.)
[6:14] * Ash-Fox (HAL-9000@adk220.neoplus.adsl.tpnet.pl) has joined #FreeNET
[6:19] <TheSeeker> there has to be an accoutning error... I should let freenet run without using frost myself to see if it does the same thing...
[6:22] <TheSeeker> here's an idea: does Frenet include keys that the node is currently recieving but haven't been saved to disk as being included in the datastore?
[6:22] <TheSeeker> Could there be a bug where when a key fails to download but would be cached it's counted as in the store each time it attempts?
[6:28] <TheSeeker> hmm, java just crashed again *compares the two logs*
[6:30] <TheSeeker> maybe it doesn't like 200 nodes max...
[6:31] * Hory (~Miranda@82.76.81.56) has joined #FreeNet
[6:33] * moskau23 (~Miranda@dsl-082-082-239-116.arcor-ip.net) has joined #freenet
[6:36] <kevloral> does anyone in here have a lexpark printer? http://www.engadget.com/entry/3508496627337139/
[6:36] <kevloral> s/lexpark/lexmark
[6:56] <TheSeeker> hrm, I wonder what FUQID's "aborted by freenet" error really means?
[7:00] <TheSeeker> hmmm, I have 20 outbound connections, and am sending at 10000 bytes per second so if I go on the assumption that on average I'm only going to use 500 B/s for transmitting data, half of my nodes are sending and the other half recieving, have a 30000 B/s bandwiddth limit, and want to save some % for routing ... hmm... I should be able to handle ... 100 rt nodes, huh.
[7:00] <TheSeeker> (20 outbound transfers, not connections)
[7:01] <TheSeeker> that seems to be a bit below my desired level of performance though...
[7:01] * sanity (~ian@81-178-106-165.dsl.pipex.com) has joined #freenet
[7:02] <TheSeeker> if freenet was good at self-organizing, would the slow nodes move to the outside fringes of the network?
[7:04] <TheSeeker> I would think that a minimum expected speed per node should be 1K/s, which means I should probably not be conencted to more than 60 nodes with a 30K/s upstream limit... is this a valid assumption? and: can Freenet work well when connected to fewer nodes, even if the average speed between nodes is faster?
[7:05] <TheSeeker> KenMan: when you wake up, have you tried running simulations based off of small rt numbers vs large ones? (and is transfer speed at all worked into your sim?)
[7:16] * moskau23 (~Miranda@dsl-082-082-239-116.arcor-ip.net) Quit ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
[7:18] * sanity (~ian@81-178-106-165.dsl.pipex.com) Quit ()
[7:21] * sanity (~ian@81-178-106-165.dsl.pipex.com) has joined #freenet
[8:03] * TLF (~francisco@36.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")
[8:04] <TheSeeker> I think I see an interesting problem with a full datastore...
[8:08] <TheSeeker> while the datastore is filling, keys of all sizes will be stored... but what about when the datastore is full? When a key is set to be cached, does it bump off the LRU key or the LRU key of equal size? if I et a frost message that my store wants to keep, will it bump off a 1 MB key? what happens when the store is full, and a 1 MB key wants to be stored but it bumps off a 4k key? does the 1 MB key get stored anyway?
[8:25] <TheSeeker> Space used by temp files 36,556 KiB
[8:25] <TheSeeker> temp dir reports 16 MB used
[8:30] * Sugadude (~Sugadude@pakastelohi.cypherpunks.to) Quit (Remote closed the connection)
[8:31] <TheSeeker> ok, this might be useful *restarts node* going to save the data on key size distrobution at startup then every hour or so save the data again with the number of extra keys noted
[8:32] * MANCOM (MANCOM@p548101E5.dip.t-dialin.net) has joined #freenet
[8:41] * Hory` (~Miranda@82.76.81.56) has joined #FreeNet
[8:52] * Sugadude (~Sugadude@82.143.204.119) has joined #freenet
[8:54] * Hory (~Miranda@82.76.81.56) Quit (Read error: 110 (Connection timed out))
[8:59] * Gunnar (~gunnar@34.80-203-209.nextgentel.com) Quit (Remote closed the connection)
[9:03] <TheSeeker> I somehow have 80 1 MB keys mroe than I did half an hour ago with 51 MiB transferred ... hmm... are the 1 MB keys files that are between 512K and 1M, or is there padding to ensure that all files are of a static size relative to key size?
[9:27] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:27] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:28] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:28] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:32] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:32] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:32] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:32] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:34] <KenMan> TheSeeker - padding ensures that most files are within a few bytes of a static power of two size.
[9:35] <KenMan> just count the number of unique file sizes in the store subdir, excluding the temp directory.
[9:36] <KenMan> it should only be around 45 or so. (actually it SHOULD be 10, but it's not)
[9:36] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:36] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:37] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:37] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:38] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:38] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:38] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:38] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:41] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:41] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:41] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:41] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:42] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:42] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:42] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:42] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:47] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:47] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:47] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:47] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:47] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:47] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:48] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:48] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:48] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:48] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:49] * guido^pe (~unknown@dsl-213-023-233-049.arcor-ip.net) has joined #freenet
[9:51] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:51] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:52] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:52] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Nick collision from services.)
[9:53] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[9:54] <TheSeeker> gack
[9:55] <TheSeeker> KenMan: did you see my earlier question re: RT conenctions?
[9:58] * unknown_ (~unknown@dsl-213-023-235-120.arcor-ip.net) has joined #freenet
[9:58] <Elly> Yikes.
[9:58] <Elly> Why do I keep getting killed by services?
[10:08] * guido^pe (~unknown@dsl-213-023-233-049.arcor-ip.net) Quit (Read error: 110 (Connection timed out))
[10:10] <TheSeeker> forget to identify?
[10:11] <Elly> no, my script identifies on connect
[10:11] <Elly> and it happens instantly when I connect
[10:12] <Elly> [09:54:27] * You were killed by services. (Nick collision)
[10:12] <TheSeeker> yeah..
[10:12] <TheSeeker> [06:53.53] * Elly has quit IRC (Nick collision from services.)
[10:20] <Elly> So umm.
[10:20] <Elly> Is there any way to prevent that?
[10:20] <TheSeeker> *shrug*
[10:20] <Elly> it seems to happened whenever I reconnect automatically
[10:26] * Elly is now known as Elly|afk
[10:31] * pittaman (~mistery_b@d51A410CA.kabel.telenet.be) has joined #freenet
[10:50] <Ribs> Elly|afk: Sounds like your script is making you ghost yourself
[10:59] * kevloral is now known as kevlAway
[11:43] * m[a]zzanet (~mazzanet@mazzanet.user) has joined #freenet
[11:57] * mazzanet (~mazzanet@mazzanet.user) Quit (Client Quit)
[12:18] * unknown_ (~unknown@dsl-213-023-235-120.arcor-ip.net) Quit (Read error: 110 (Connection timed out))
[12:49] * kevlAway is now known as kevloral
[12:49] <kevloral> re
[12:51] * pittaman (~mistery_b@d51A410CA.kabel.telenet.be) Quit (Read error: 104 (Connection reset by peer))
[12:53] * pittaman (~mistery_b@d51A410CA.kabel.telenet.be) has joined #freenet
[12:59] <TheSeeker> I'm up to over 2k more keys on the histograms than what's in my store...
[13:00] <TheSeeker> in 4.5 hours
[13:02] <TheSeeker> inbound requests: keys: 131561
[13:02] <TheSeeker> successful requests: keys: 2130
[13:02] <TheSeeker> :(
[13:04] <TheSeeker> http://theseeker.boucneme.net:8888/servlet/nodestatus/inbound_request_histogram.txt didn't make sense until I looked at: http://127.0.0.1:8888/servlet/nodestatus/psuccess_data_detail.txt
[13:07] <TheSeeker> all of 7# are generally very pathetic 0.6% to 2%, but 71 is a slightly less pathetic but statistically much higher 5%
[13:22] * Elly|afk is now known as Elly
[13:30] * Element155 (~your@host13-56.pool80183.interbusiness.it) has joined #freenet
[13:33] <Element155> hello, i'm new in freenet use, where can i find some keys?
[13:33] <Element155> or How
[13:34] <i2p_iip> <Sonax> Element155: Have you tried the pages linked from the web-interface? Or Frost?
[13:36] <Element155> ye, but when i chose a link it doesn't work
[13:36] <kevloral> Element155: how long has your node been running?
[13:36] <i2p_iip> <Sonax> How long have you been running freenet, and how long do you give the links.
[13:36] <i2p_iip> <Sonax> What kevloral said :)
[13:37] <kevloral> sonax: stop reading my mind! ;-)
[13:37] <Element155> approx 1 hour
[13:38] <kevloral> Element155: give it at least 48 hours. Freenet performance at the very beginning is horrible.
[13:38] <i2p_iip> <Sonax> kevloral: You beat me to it...
[13:38] <Element155> but 48 hours without stops?
[13:39] <i2p_iip> <Sonax> Element155: Maby run frost while you wait to help your node integrate.
[13:39] <kevloral> Element155: that would be desirable, yes.
[13:39] <Element155> well
[13:40] <Element155> and excuseme, what is the "Load 21%" in the web-interface?
[13:41] <kevloral> Element155: that is an estimation of the current load of your node. If you are interested, you can see how it is calculated on the "general information" link.
[13:42] <Element155> thanks a lot.
[13:42] <kevloral> Element155: you are welcome.
[13:43] <Element155> bye, see you.
[13:44] * Element155 (~your@host13-56.pool80183.interbusiness.it) Quit ()
[13:55] * hobx_ (~chatzilla@ankh.math.chalmers.se) has joined #freenet
[14:04] * SpiKee (~SpiKee@ABordeaux-251-1-37-184.w82-125.abo.wanadoo.fr) has joined #freenet
[14:21] <SpiKee> Hi
[14:22] <SpiKee> Any french around ?
[14:33] <Ribs> SpiKee: I doubt it.
[14:44] <Elly> Not I.
[15:02] * sdogi (~java@84-50-18-71-dsl.kjj.estpak.ee) has joined #freenet
[15:14] * sdogi_ (~java@84-50-20-241-dsl.kjj.estpak.ee) Quit (Read error: 110 (Connection timed out))
[15:19] * Elly is now known as Elly|afk
[15:20] * Hirvox (~hirvox@cs181124095.pp.htv.fi) has joined #freenet
[15:46] <SpiKee> omg, freenet is really a good project ..
[15:48] * goatee (~goatee@ip216-239-84-60.vif.net) Quit (zelazny.freenode.net irc.freenode.net)
[15:49] <i2p_iip> <Michelle> Freenet is against the natural order of things, for shame.
[15:50] * goatee (~goatee@ip216-239-84-60.vif.net) has joined #freenet
[15:57] <TheSeeker> the natural order of things is entropy... Freenet is a huge waste of resources, and thus promotes entropy... :D
[16:00] * SpiKee (~SpiKee@ABordeaux-251-1-37-184.w82-125.abo.wanadoo.fr) Quit ("Leaving")
[16:07] <mikeeusa> hi
[16:07] <mikeeusa> <i2p_iip> <Michelle> Freenet is against the natural order of things, for shame.
[16:07] <mikeeusa> Freenet rocks
[16:08] <mikeeusa> ... damn women, hating freenet...
[16:08] <i2p_iip> <Sonax> Michelle is against the natural order of things, for shame.
[16:10] <mikeeusa> most women today are against the natural order of things
[16:13] <TheSeeker> Is being malnourished while pregnant, thus causing a miscarriage, a crime in the USA?
[16:14] <mikeeusa> no
[16:14] <mikeeusa> in the USA it is legal for a woman to murder her children
[16:14] <mikeeusa> also it is defacto legal for women to cut off their husband/boyfriend's penis
[16:15] <mikeeusa> (they get no jail time)
[16:15] <TheSeeker> heh
[16:15] <mikeeusa> and it is almost legal for them to murder their husbands (1 year jail time).
[16:15] <mikeeusa> now if a man says something mean to a woman in the USA
[16:15] <mikeeusa> 5 years for sexual harrasment
[16:20] <mikeeusa> (round there)
[16:23] <TheSeeker> is ending a life before it starts better than bringing life into conditions which almost certainly cause severe depresion and suicide? (welfare+drugs+abusive parent[s]+inner city public school system)
[16:29] * kevloral is now known as kevlAway
[16:32] <i2p_iip> <Michelle> what the hell are you talking about, mikeeusa ?
[16:48] * m[a]zzanet (~mazzanet@mazzanet.user) Quit (Client Quit)
[16:53] * mazzanet (~mazzanet@mazzanet.user) has joined #freenet
[16:53] <mikeeusa> <TheSeeker> is ending a life before it starts better than bringing life into conditions which almost certainly cause severe depresion and suicide? (welfare+drugs+abusive parent[s]+inner city public school system)
[16:53] <mikeeusa> it's called murder
[16:53] <mikeeusa> <mikeeusa> also it is defacto legal for women to cut off their husband/boyfriend's penis
[16:53] <mikeeusa> <mikeeusa> (they get no jail time)
[16:53] <mikeeusa> <TheSeeker> heh
[16:53] <mikeeusa> seems you don't care that women can mutilate freely either
[16:54] <mikeeusa> pathetic girly man
[16:54] <mikeeusa> worshipping your goddesses so you can "get some"
[17:04] * pittaman (~mistery_b@d51A410CA.kabel.telenet.be) Quit ("http://www.ic-musicmedia.com/pittaman subtle spam")
[17:15] * Elly|afk is now known as Elly
[17:18] * mikeeusa (~r00t@ool-18be3a49.dyn.optonline.net) Quit (Read error: 110 (Connection timed out))
[17:21] * mikeeusa (~r00t@ool-18be3a49.dyn.optonline.net) has joined #freenet
[17:26] * MANCOM (MANCOM@p548101E5.dip.t-dialin.net) Quit ()
[17:27] * Elly is now known as Elly|afk
[17:41] * Elly|afk is now known as Elly
[17:48] <KenMan> mikeeusa: are we to take it that you do not believe women are the fairer sex ?
[17:50] <TheSeeker> KenMan: still wondering if you've tested your sim with netwrks that have only a few RT conenctions vs those that have many... and if you have link speed at all accounted for with regards to transfers
[17:50] <KenMan> i've used small (3) and large (100) sized RT. I am not simulating link speed.
[17:51] <KenMan> I'm just trying to get the basics nailed down. Very basic basics...
[17:52] <TheSeeker> how about in the middle, around 30?
[17:52] <TheSeeker> though taking link speed and having variable speeds per node would be required to answer my original qestion...
[17:53] <TheSeeker> which is: Would freenet work better if everyone limited their number of RT connections to allow for a much improved minimum average upload speed per node?
[17:54] <KenMan> Right now i'm running 2457nodes x 12routes , it has done 225M unique key searches.
[17:54] <TheSeeker> an automatic RT node conenction number based off the user's upload limit seems reasonable... presuming that the limit is accurate.
[17:54] <TheSeeker> (meaning not above actual physical limit)
[17:54] <KenMan> I like that philosophy. The basic answer is , the more routes you have, the longer it takes to learn about each one.
[17:55] <TheSeeker> so, with fewer routes, not only are they faster, but you learn about them more quickly?
[17:55] <KenMan> that is generally agreed upon by most participants.
[17:56] <TheSeeker> but the general populace also seems to think that increasing the number of maxconnections to 100 or 200 will improve their freenet performance...
[17:56] <KenMan> Unfortunately, the current implementation of mRI (request rate limiting) throws a wrench into the works.
[17:56] <TheSeeker> ah
[17:56] <KenMan> increasing it will improve performance, thanks to mRI :(
[17:56] <TheSeeker> doesn't the removal of the failure table help eleviate some of that ?
[17:56] <TheSeeker> along with the addition of request queueing?
[17:57] <KenMan> to date , that is not determined
[17:57] <KenMan> request queueing did not do much at all to alleviate the negative impact of the current mRI system. It is more of a 'workaround' than a solution.
[17:58] <KenMan> I mean, it was a positive effect, but netwide, probably not a significant improvement.
[17:59] <KenMan> And, if anything, it increased search times. But that is irrelevant when fetches take as long as they do.
[18:01] <TheSeeker> what I'd really like to see is the number of connections transferring limited to "as many as possible before lowering the average upload speed below X"
[18:01] <KenMan> i think the failure table was implemented prior to mRI, and the two are vaguely similar
[18:02] <KenMan> we would all like to see a solution that does what you say
[18:02] <TheSeeker> if a search only takes N seconds to find the data on the network, and the data can be sent back at least faster than 1 K/s, most freesites (html) will probably load in under N+30 seconds ...
[18:02] <KenMan> sounds good, until you throw a bunch of nodes at each other, and they all start to blow smoke.
[18:03] <KenMan> and leak oil...
[18:03] <TheSeeker> sounds like I need to write my own sim x.x
[18:03] <KenMan> go fer it ;) anything you can contribute helps !
[18:03] <TheSeeker> but I have no clue how to implement the current estimators *cries*
[18:03] <KenMan> don't bother. Do something simpler. I am in the same boat.
[18:03] <TheSeeker> so I'd just have to make up my own code that does freenet-ish things :P
[18:04] <KenMan> or, work from the existing java simulation.
[18:04] <TheSeeker> I know 0 java, mine would be written in XBasic or Euphoria.
[18:04] <Elly> Java is not a terribly difficult language to pick up
[18:04] <TheSeeker> probably Euphoria because of it's unique array based data structures...
[18:04] <KenMan> oh, well, don't let that stop you. Just use the simplest form of an estimator you can come up with.
[18:05] <TheSeeker> KISS?
[18:05] <KenMan> yes, KISS :)
[18:06] <KenMan> for example, my nodes remember every successful key retrieved, as well as each insert, (up to 45 of them) for every connected peer
[18:07] <TheSeeker> once upon a time I started writing a sim for Gnutella back before download queues were accepted as being useful... I wonder if I still have that code around...
[18:07] <KenMan> then, for an estimate, I simply add the distances to the 5 closest known keys for that peer.
[18:07] <KenMan> It clearly makes them specialize, according to the pictures I draw.
[18:07] <TheSeeker> hmm
[18:08] <KenMan> that's all it needs to do. Eventually, I could compare the existing estimators' performance to my simplistic version, but I'm not there yet.
[18:09] <TheSeeker> You know.... I wonder what would happen if you told probalistic caching to start at 0% full unstead of like 95% it's at now?
[18:09] <KenMan> it doesn't make any difference after enough time has passed.
[18:09] <TheSeeker> I'm tempted to point my datastore at a new directory just to find out...
[18:09] <KenMan> but, if it improves the initial window of a connection's lifetime, that would be very valuable.
[18:10] <TheSeeker> I think we lost eachother in our own thoughts.
[18:10] <KenMan> faster learning is good. But I don't think it will make much difference on the learn-rate of your peers. give it a whirl anyway.
[18:11] <TheSeeker> first I have to write some code that simulates a network :P
[18:11] <TheSeeker> or did you mean the resetting the DS thing?
[18:11] <KenMan> resetting the DS. It is easy, temporary, and might be educational.
[18:12] <KenMan> The primary idea is to get a full store as fast as possible, then you can refine it later.
[18:13] <TheSeeker> a "full store as fast as possible" is useless if it's equally distributed among the keyspace... there's plenty of data in any range of keyspace to fill anyone's store...
[18:13] <KenMan> pcaching was (i think) designed to improve anonymity, not performance. It does have a side-effect of increasing the global storage/reducing redundancy net-wide.
[18:14] <KenMan> which in general will decrease odds of success for a single chosen key
[18:14] <TheSeeker> "minStoreFullPCache=0.9"
[18:14] <TheSeeker> I thought Pchaching was all about only storing keys close to your specialization?
[18:15] <KenMan> full store ASAP is NOT useless, even if it is perfectly uniform. It will generate hits, which can then grow into specialization.
[18:15] <TheSeeker> or is it supposed to be that since other nodes all agree that you are specialized at X, your node will naturally only store data near X because that's all that gets routed to you?
[18:15] <KenMan> no, pcaching has nothing to do with your spec.
[18:15] <KenMan> that is the correct theory re:spec
[18:16] <TheSeeker> that's stupid. how are other nodes supposed to agree what you're specialized in if there's no rational way of defining it? :P
[18:16] <KenMan> exactly. And what are the odds of them agreeing on the same/similar spec, if they don't talk to each other ?
[18:17] <KenMan> well, the rational way is that if you have a specialized set of routes, the DS will converge on that spec over time, and they will sense it.
[18:17] <TheSeeker> "I know I'm good at routing X, but I'm not going to tell you" followed by "hey, can you route X? -- No! I'm busy, go away"
[18:17] <KenMan> having a specialized set of routes requires that you can smell other nodes spec sufficiently, and that you mingle with those most like yourself
[18:17] <KenMan> heh, you have a deeper insight than I realized ;)
[18:18] <TheSeeker> I should probably have asked this first...
[18:18] <TheSeeker> does your node prioritize which requests it will take based on self-observed routing effectiveness?
[18:19] <KenMan> in general, no. The queueing that toad built allows reordering, i think.
[18:19] <Elly> Out of curiosity, how old are you two?
[18:19] <KenMan> too old, and just young enough
[18:20] <KenMan> (I'm too old at 32). The Seeker is in his 20's (or less?)
[18:20] <TheSeeker> I think that an inverse proportion to overload status * '1/distance to your specialization point' (I think?) chance of accepting a request
[18:20] <Elly> Aha, that makes me feel better.
[18:20] <TheSeeker> in my 20s, often thought of as less 'cause I'm short >.<
[18:21] <KenMan> it has to be statistical. we want nodes to accept some random keys, but mostly spec'ed keys
[18:22] <KenMan> Nodes can have more than one spec point, like galaxies in the universe. But they tend to have one that is largest.
[18:23] <KenMan> No one has ever spilled the beans on precisely what specialization is, because no one knows.
[18:23] <TheSeeker> I am one that says that you should only have one spec point, and who cares if anyone knows what it is?
[18:24] <TheSeeker> I say specialization is a point in keyspace that you 1) know how to route, and 2) that you have filling your datastore, because why would you have any other areas of the keyspace covered if you're only wanting to route those keys close to your spec?
[18:24] <KenMan> well, suppose you have 20 routes, thus 20 spec points. Each one had better be a wide-enough mountain (2^160/20) to cover your needs.
[18:25] <KenMan> both your local needs, and those statistically rare requests that are not on your mountain.
[18:25] <TheSeeker> why? you only need to get closer on the first hop, not right to the key in question
[18:26] <KenMan> how can you guarantee that you are always getting closer ? I haven't found any way other than DHT, like Chord's finger tables.
[18:26] <TheSeeker> for good routing it seems to me that ideally you would want a large number of nodes near your own spec, and a smaller number of nodes equally distributed among the keyspace.
[18:27] <KenMan> yes, that seems intuitive
[18:27] <TheSeeker> well, if each node *knows* the spec of the nodes it's connected to, you send it to the node with the spec closest to your goal.
[18:27] <KenMan> how are these nodes' single spec points to be chosen ?
[18:28] <TheSeeker> though, thinking about it a step further I can see how that might lead to attacks on anonymity... which I guess is a higher concern than performance for freenet...
[18:28] <KenMan> yes
[18:28] <Elly> hey, wow
[18:28] <Elly> freesites seem to load almost as fast as normal web pages now
[18:29] <TheSeeker> if you have a disproportionatly large number of nodes connected at your spec point, then most requests that come from your node to other non-spec nodes are more than likely going to be coming from you ... though not neccesarily true, overwhelmingly statistically true...
[18:30] <TheSeeker> so if you are FBI or whateer, you hack a node to connect to non-spec poitn nodes only (being on everyone else's "other" list) then watch requests come in and match them to who sent the request...
[18:30] <TheSeeker> not 100%, but convincing...
[18:31] <TheSeeker> but if you don't connect to a disproportionately large number of nodes near your spec point, then very few routes will get to you because of HTL and everyone with similar specs being spread out so much...
[18:31] <TheSeeker> freenet is doomed! :P
[18:31] <KenMan> sort of. That's the kind of problems to be solved ;)
[18:32] <TheSeeker> absolute anonymity cannot coexist with any sort of usable performance :P
[18:32] <KenMan> ah hah! Science in action :)
[18:32] <TheSeeker> though, even as it stands, all freenet seems to be looking for is plausible deniability.
[18:32] <TheSeeker> so the previous large % near and rest spread idea is still valid.
[18:34] <TheSeeker> while it may be possible to hack a node to sniff out what a person is requesting by knowing their spec and seeing what non-spec stuff they request, there's no way of proving that you're not just a guy closer to the key soemone was looking for...
[18:37] <TheSeeker> very simple 2D format of my thinking .... I see the keyspace as a disk, on the inside of the disk you have fast nodes, on the outside edge of the disk you have slow nodes... now create a concentric ring of dots to fill the disk, and conenct the dots of each node to the ones within a radius of N to itself, as well as half as many other nodes evenly spread at the same radius on the disk ...
[18:38] <TheSeeker> now how do you manage the insertion of new nodes into such a network?
[18:38] * Ash-Fox (HAL-9000@adk220.neoplus.adsl.tpnet.pl) Quit (Client Quit)
[18:39] <TheSeeker> immediately after that thought came the old gnet debates about topology, and how a distributed community could assign a new node a place in the tolpology, as well as promote or demote/ban a node from the grid ... there were discussion of N-dimensional planes to describe interconnections and it made my brain hurt.
[18:40] <TheSeeker> I probably see single spec as the answer because I have trouble thinking in 3 dimensions, let alone 24 :P
[18:41] <TheSeeker> I should get started making notes for my sim, thinking this much is making me uncomfortable
[18:43] * Ash-Fox (HAL-9000@adm245.neoplus.adsl.tpnet.pl) has joined #FreeNET
[19:02] <KenMan> good luck Seeker. I find that people who can express such concepts in 2 or 3 dimensions (physically), as you have done, have a leg up on the folks who cannot.
[19:03] <Elly> sigh
[19:03] <Elly> I need 512MB more RAM if I'm gonna leave freenet running 24/7
[19:03] <TheSeeker> oh?
[19:04] <TheSeeker> Freenet WILL eat as much RAM as you let it have, giving it more just means it will use more :P
[19:05] <TheSeeker> ooh, ooh, I'm about to hit 20k keys!
[19:05] <TheSeeker> even though I've only got 17.3k files in my ds ...
[19:06] <TheSeeker> http://theseeker.bouncme.net:8888/servlet/nodestatus/ds_size_histogram_data.txt
[19:06] <TheSeeker> oops
[19:06] <TheSeeker> http://theseeker.bounceme.net:8888/servlet/nodestatus/ds_size_histogram_data.txt
[19:11] * d-ArkAngel (~Rob@cpc2-midd2-5-0-cust16.midd.cable.ntl.com) has joined #freenet
[19:12] <KenMan> howdy rob :)
[19:12] <KenMan> how's life treating you these days ?
[19:26] * hobx (~hobx@83.248.111.142) Quit ("Client exiting")
[19:26] * hobx (~hobx@83.248.111.142) has joined #freenet
[19:30] * moskau23 (~Miranda@dsl-213-023-248-084.arcor-ip.net) has joined #freenet
[19:30] * d-ArkAngel (~Rob@cpc2-midd2-5-0-cust16.midd.cable.ntl.com) Quit (Read error: 110 (Connection timed out))
[19:37] <KenMan> he was a monster, black, dressed in leather...
[19:37] <KenMan> she was a princess, queen of the highway...
[20:03] * guido^pe (~unknown@dsl-213-023-244-084.arcor-ip.net) has joined #freenet
[20:14] * guido^pe (~unknown@dsl-213-023-244-084.arcor-ip.net) Quit (Remote closed the connection)
[20:15] * sanity (~ian@81-178-106-165.dsl.pipex.com) Quit ()
[20:42] * moskau23 (~Miranda@dsl-213-023-248-084.arcor-ip.net) Quit ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
[20:51] * Hory` (~Miranda@82.76.81.56) Quit ("CyberLore.net - Recommendations on the best games, freeware and websites.")
[21:02] <TheSeeker> http://www.rapideuphoria.com/ <- I'm tempted to buy the full version of this so I can compile the interpeter and the sources I make into a standalone exe ...
[21:04] <TheSeeker> !! woot!
[21:04] <TheSeeker> The free Public Domain Interpreter (ex.exe, exw.exe, exu), now includes full support for trace() beyond 300 statements, and for profiling. Thus there is no longer a registered "Complete Edition" interpreter. However, registration is still required for the Binder / shrouder.
[21:04] <TheSeeker> yey, I was wanting to profile the last application I wrote but it quikly went WAY beyond 300 lines...
[22:34] * kmag (~kmag@pcp04402498pcs.nrockv01.md.comcast.net) has joined #freenet
[22:37] <kmag> How do I control which JVM my freenet node starts up in a sub-process in the startup sequence?
[22:37] <kmag> I start freenet in a 1.5.0 JVM
[22:37] <kmag> and it tells me that it has found a 1.4.2 JVM and is starting that
[22:39] <kmag> however, my tests show that TCP accept latency in the 1.5.0 JVM is 10% that of the 1.4.2 JVM
[22:39] <kmag> at least for NIO
[22:40] * Elly is now known as Elly|afk
[22:55] <kmag> Is anyone out there running freenet on Sun's 1.5.0 JVM?
[22:55] <kmag> Jikes RVM would also be cool
[23:00] <mikeeusa> <KenMan> mikeeusa: are we to take it that you do not believe women are the fairer sex ?
[23:00] <mikeeusa> correct
[23:00] <mikeeusa> women are manipulative etc
[23:03] <kmag> fairer is usually taken to mean prettier in that context
[23:04] <kmag> though, I jumped into the middle of the conversation, so maybe that has also been covered
[23:04] <mikeeusa> most are fat and ugly now adays
[23:04] <mikeeusa> and even if they are pretty, facades are worthless if they are only facades
[23:05] <kmag> I missed the beginning of the conversation. Are you asserting that men are prettier than women?
[23:06] <mikeeusa> no
[23:07] <mikeeusa> i am asserting that women are crap bitches
[23:07] <mikeeusa> and thus it dosn't matter if they are pretty
[23:07] <mikeeusa> because they have nothing to back that 'virtue' up with
[23:09] <kmag> recent break-up?
[23:15] <kmag> trying to talk yourself into joining a monastary or becoming gay?
[23:21] <kmag> perhaps trying to justify a career in software development?
[23:25] <mikeeusa> none
[23:26] <mikeeusa> You see I want a submissive pretty slave wife
[23:26] <mikeeusa> but to no avail
[23:26] <mikeeusa> because they don't make them like they used to
[23:26] <mikeeusa> (1k years ago)
[23:27] <kmag> just move to a country that still lives in the stone age
[23:28] <kmag> in some countries you are legally allowed several slave-wives at the same time!
[23:33] <mikeeusa> according to the Bible you are
[23:33] <TheSeeker> which one?
[23:33] <mikeeusa> so *logically* Christian nations should be similar
[23:33] <mikeeusa> TheSeeker, the regular one
[23:34] <mikeeusa> read it sometime, it has nice stuff for men (tm)
[23:35] <TheSeeker> I tend to avoid obvious fear-mongering propaganda if I can help it...
[23:36] <mikeeusa> ?
[23:36] <mikeeusa> ok
[23:37] <mikeeusa> hopefully there will come a time when all the world is socially in the stone age
[23:37] <mikeeusa> where women have no rights
[23:37] <mikeeusa> where the curtain of opression falls anew
[23:37] <TheSeeker> most religions are founded on peoples fear of the unknown, and primarly death ... smarter than average people see this as an opportunity to politically control the masses, and do a pretty good job of it.
[23:37] <mikeeusa> and crushes those who once sought to turn the tables of men
[23:38] <mikeeusa> TheSeeker: as long as I get my slave wives
[23:38] <mikeeusa> and make it a world wide order
[23:38] <mikeeusa> of female servitude
[23:38] <mikeeusa> I will be happy
[23:39] <TheSeeker> trying to re-learn euphoria for the purpose of making a simulator is making my brain hurt, I think I'll get some sleep
[23:41] <i2p_iip> <stefanboy> ew, euphoria
[23:44] <TheSeeker> you don't like fast, interpeted, array-based languages? :P
[23:45] <TheSeeker> ah well, it ensures that nobody will ever want to look at the source of my sim, if it's ever working...
[23:46] <i2p_iip> <stefanboy> euphoria used to be hideous when it came out years and years ago, dunno what it's like now
[23:46] <i2p_iip> <stefanboy> euphoria used to be hideous when it came out years and years ago, dunno what it's like now
[23:46] <i2p_iip> <stefanboy> what are you simulating?
[23:46] <TheSeeker> I didn't start using it until 2.4 2.5 came out recently so I'm taking another look at it, along with the win32 libraries and IDE
[23:47] <TheSeeker> freenet hopefully, or something similar to it
[23:47] <i2p_iip> <stefanboy> hmm. to what end?\
[23:48] <TheSeeker> it'll have 'nodes' that 'connect' to other nodes and do searches for keys, insert new keys, and generally try to be a freenet without any real data... just keys and attributes of keys (file sizes)
[23:48] <TheSeeker> I intend to implement routing ideas and offer simulation results as some sort of proof that they are valid
[23:49] <i2p_iip> <stefanboy> sounds like OOP would be better suited for that sort of application
[23:50] <TheSeeker> why's that?
[23:50] <i2p_iip> <stefanboy> design a node object and have them interact
[23:50] <TheSeeker> it's just a loop that increments a timeslice and moes events around in a simulation.
[23:50] <TheSeeker> *moves
[23:52] * Elly|afk (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Read error: 60 (Operation timed out))
[23:53] <TheSeeker> since I know little about OOP, and don't feel like attempting to learn about it, AND another languave I've never used before to implement it, I think I'll stick to my original plan, which is rapid development of code in a simple, yet powerful, language.
[23:54] <i2p_iip> <stefanboy> if you take the time to learn say, java, you will kick yourself for using e
[23:55] <TheSeeker> yeah, but "taking the time to learn java" will add a couple years to my sim project, and I'd rather be mostly done by the end of the week.
[23:56] <i2p_iip> <stefanboy> OOP is pathetically simple, it will take you to the end of the week to learn java =)
[23:56] <TheSeeker> I doubt it. I've tried understanding concepts of OOP for years, but it's never clicked.
[23:57] <TheSeeker> I think in terms of monolithic programs that have a main loop, and maybe some subs and functions if I have a block of code I use more than a couple times.
[23:57] <mikeeusa> OOPS!
[23:57] <mikeeusa> OOPS!
[23:57] <mikeeusa> LOLZ
[23:57] <i2p_iip> <stefanboy> what concepts do you not understand?
[23:57] <mikeeusa> oop aint very good
[23:57] <TheSeeker> most of them.
[23:58] <mikeeusa> but java does have a main function
[23:58] <TheSeeker> what's the difference between an "object" and a TYPEd array?
These logs were automatically created by Jay Oliveri with his gimp hapi on irc.freenode.net.