Timestamps are in GMT/BST.
[17:42] <ShaunMacPherson> Is Freenet both a key based routing method and a DHT?
[17:43] * TLF (~francisco@91.Red-81-40-116.pooles.rima-tde.net) has joined #freenet
[17:46] * greycat (~wooledg@192.35.79.70) Quit ("Lick Bush in '04")
[17:47] <toad_> hi
[17:47] <toad_> ShaunMacPherson: yes
[17:47] <ShaunMacPherson> koool
[17:47] <ShaunMacPherson> I am editing up the wikipedia articles on network routing methods and noticed Freenet had 2 :)
[17:50] <ShaunMacPherson> just making sure thanks
[17:50] <toad_> ahhh
[17:50] <toad_> :)(
[17:50] <toad_> :)
[17:56] * JustMe_ (JustMe_@cs68206255-46.houston.rr.com) has joined #freenet
[18:00] * Toast (~Toast@82-34-170-133.cable.ubr03.maid.blueyonder.co.uk) has joined #freenet
[18:01] <Toast> I'm much liking what I'm hearing on Devl lately.
[18:03] <toad_> Toast: specifically?
[18:04] <mikeDOTd> has anyone ever run freenet's code through this? --> http://pmd.sourceforge.net/
[18:05] <ShaunMacPherson> Pretty interesting mikeDOTd
[18:05] <toad_> argh
[18:05] <ShaunMacPherson> That is very interesting, do other languages have auto bug checkers?
[18:05] <toad_> mutt's encryption support is broken...
[18:07] <cbreak> I worked with a language where you had to specify pre and post conditions...
[18:07] * FallingBuzzard (~srademach@207.152.112.129) Quit (Read error: 104 (Connection reset by peer))
[18:09] <cbreak> also, most compilers should mention syntax errors and minor problems like unused local variables. And auto Inlining gets rid of empty functions.
[18:09] <toad_> even Eclipse does
[18:09] <toad_> in real time
[18:15] <ShaunMacPherson> Is PMD of mikeDOTd showed, a java autochecker, that might help
[18:15] <toad_> does it test anything eclipse doesn't?
[18:15] <ShaunMacPherson> Good point...
[18:15] <ShaunMacPherson> probably not, and if it doesnt it will be put into eclipse i'd guess :)
[18:16] <mikeDOTd> it's a plugin for eclipse
[18:16] <mikeDOTd> so it's likely it adds functionality :-)
[18:17] <ShaunMacPherson> heh
[18:17] <ShaunMacPherson> One would hope so :)
[18:17] <toad_> well we haven't fixed all the warnings eclipse finds
[18:17] * Zorix (Brandon@fl-65-40-46-10.dyn.sprint-hsd.net) Quit ()
[18:18] <toad_> most of the things it talks about are in the build of eclipse that i am using
[18:18] <toad_> Empty try/catch/finally/switch blocks - yes, some of them
[18:18] <ShaunMacPherson> I thought it wasnt wise to use eclipse since it might lack features in the '
[18:18] <toad_> Unused local variables, parameters and private methods - yes
[18:18] <ShaunMacPherson> offical' java version, or should we all use eclipse?
[18:19] <toad_> # Empty if/while statements - yes, think so
[18:19] <toad_> Overcomplicated expressions - unnecessary if statements, for loops that could be while loops - not sure
[18:19] <toad_> Classes with high Cyclomatic Complexity measurements - don't think so
[18:19] <cbreak> what is cyclomatic complexity?
[18:19] <ShaunMacPherson> i'd also like to know :)
[18:20] <ShaunMacPherson> i'll add it to wikipedia heh
[18:20] <ShaunMacPherson> i think its a measure of how complex a program is and the riskiness based on it
[18:23] <toad_> hmmm
[18:23] <toad_> what is the next priority?
[18:23] <toad_> compile 60243 on dodo, check it, send the announce mail
[18:23] <Toast> toad: The progress that the simulations are making, the drive to simplify and the steady progress on the stable and unstable nets
[18:23] <toad_> then probably get back to simulations
[18:23] <Toast> Everything seems to be going in the right direction.
[18:23] <toad_> Toast: I'm skeptical on the drive to simplify :)
[18:24] <cbreak> fixed size keys, htl elimination, pre mix routing :)
[18:24] * pupok_ (~r00t@81-178-94-146.dsl.pipex.com) Quit (Read error: 104 (Connection reset by peer))
[18:24] <Toast> toad: I can understand that and I don't think anything should be absolutely thrown away, but it would be best to strip things down a little then build back up using a little more riggor
[18:24] <toad_> fixed size keys and htl elimination are definitely simplifying steps
[18:24] * pupok_ (~r00t@81-178-94-146.dsl.pipex.com) has joined #freenet
[18:24] <Toast> and the simulations provide that riggor
[18:25] <lolo-laptop> toad_: I wish I had home net connection to test out your latest work...
[18:25] <toad_> lolo-laptop: you could play with the simulations?
[18:25] <toad_> okay
[18:25] <toad_> options:
[18:26] <lolo-laptop> hmm... oh, and i blew up one of my machines, so I can't really do that either :-(
[18:26] <Toast> Another thing that would be nice is that before making a routing change to the network, indicate what difference you expect it to make and name the stats/indicators in the node you would like people to record before and after.
[18:26] <toad_> 1. more debugging.
[18:26] <toad_> 2. more minor features: transfer coalescing would be the obvious thing.
[18:26] <toad_> 3. more work on simulations: specifically, non-fully-connected-network support.
[18:26] <toad_> pick one of the above 3 options...
[18:26] <Toast> 3
[18:27] <Toast> let others find the bugs (1), but you can fix them when they have been tracked down.
[18:27] <toad_> heh
[18:27] <lolo-laptop> i'll rank the tree: 2-3-1
[18:28] <lolo-laptop> s/tree/three/
[18:28] <cbreak> Hmm... I don't like bugs. I like new features... hmm... :)
[18:28] <KenMan> more sim work , yeah !!
[18:28] <Toast> I say 3 before 2, because new features should be simulated first.
[18:28] <lolo-laptop> transfer coalescing is an _obvious_ new feature...
[18:28] <toad_> Toast: only if they have a serious routing impact
[18:29] <lolo-laptop> obvious new features, then more simulation, and fix bugs as they show up
[18:29] <KenMan> oh, how about the usual 'take it as it comes' and do what YOU most want to do, on a minute-by-minute basis :)
[18:30] <lolo-laptop> lol
[18:30] <lolo-laptop> heya KenMan
[18:30] <lolo-laptop> :)
[18:30] <KenMan> see, by asking for input, you can do this better.
[18:30] <KenMan> hi lostlo
[18:30] <toad_> lol KenMan
[18:31] <toad_> rofl
[18:31] <KenMan> why are you guys laughing at me ?? ?
[18:31] <toad_> what I mean is what should I do now ?
[18:32] <KenMan> how about more thinking out loud ? I'm building up my new fedora install to do 'stuff' like java!
[18:32] * toad_ thinks he should work on the simulator
[18:32] <KenMan> whippee
[18:32] <toad_> not whoopee, whippeee... hmmm
[18:32] <cbreak> that will give us a lot of nice pictures. good :)
[18:33] * toad_ updates the current lot of pretty pictures...
[18:33] * xolo (1000@xoder.rh.rit.edu) has joined #freenet
[18:34] * pittaman (~mistery_b@d51A410CA.kabel.telenet.be) Quit ("Gesmolten. Weerstand nutteloos. Kom me nu toch eindelijk eens ophalen ;)")
[18:36] <toad_> http://amphibian.dyndns.org/simulations2-new.png
[18:36] <toad_> that's the current pretty picture
[18:36] <KenMan> whippee vs whoopee is like the difference between a puppy saying 'birk birk' instead of a mature dog saying bark bark
[18:36] <toad_> if brown doesn't turn down like the newbie-announce curves do, then we're in business
[18:37] <toad_> green and purple are promising but a little suspicious
[18:37] <toad_> unfortunately it's not really practical to simulate 100M requests...
[18:37] <KenMan> i'll say this , even though someone else pointed it out already. Is this meaningful to us, considering the values on the x axis ??
[18:38] <toad_> KenMan: well, IF brown works, then brown is very meaningful
[18:38] <toad_> KenMan: but I need to know that it doesn't collapse with time#
[18:38] <toad_> and it is 35M requests over 100 nodes
[18:38] <toad_> which is only 350K requests per node
[18:38] <KenMan> whether this represents 1M queries against all routes, or 1M to one route, it's not practical, is it ??
[18:39] <KenMan> so it would take the average node a month or longer to achieve this level of data ?
[18:40] <cbreak> my node has about 76513.608 queries per day.
[18:40] <cbreak> (Well, it had, if it would run 24 Hours...)
[18:40] * KenMan will have to spend some time learning & understanding what toad be doing :(
[18:41] <toad_> cbreak: most of those fail...
[18:42] <cbreak> yes.
[18:43] <KenMan> i would think that 100 successes (5000 attempts) would HAVE to be good enough to form a reasonable basis for statistics (estimates) to do any good.
[18:44] <KenMan> We are constantly refining them over time, but it has to be meaningful from early on, to have non-negative effects.
[18:44] <KenMan> and then there is the 'shifting spec' aspect to be considered :(
[18:44] <ShaunMacPherson> 100 successes? you could do meaningful stats with a sample size of 5 :)
[18:45] <ShaunMacPherson> Need to do the t-test though
[18:45] <cbreak> is one request suposed to help all nodes allong the path to specialize?
[18:45] <toad_> KenMan: well, 100 nodes 100 node RTs ~ 100 requests
[18:45] <toad_> = ~ 1M requests for reasonable performance, you would hope...
[18:46] <KenMan> How important is it that most/all requestors develop 'similar' estimates of a single node ??
[18:47] <toad_> KenMan: pretty important!
[18:47] <KenMan> could we analyze from that perspective then ?
[18:48] <KenMan> Maybe as a simplification, the 'target' node could have a fixed DS within some sort of sim ? and we could compare the resulting estimates of various peers ?
[18:48] * KenMan will stop thinking out loud now...
[18:49] <toad_> possibly
[18:49] <toad_> hack it up
[18:49] <toad_> okay, how do i bootstrap the network?
[18:49] <KenMan> the only important measurements would be 'how similar' and 'how quickly can they converge' ?
[18:50] <toad_> make 25 fully connected nodes, then do lots of requests, slowly adding new nodes?
[18:52] <toad_> or what?
[18:52] <KenMan> are you talking to me ? or just out loud ? what are you trying to do ?
[18:53] <cbreak> if the seednodes contain estimators, you could collect a few different ones and compare the real life equalness...
[18:53] <toad_> I'm trying to make the simulator do non-fully-connected topologies
[18:53] <seberino> help! friend of mine says freenet hasn't really developed/evolved in the last 2 years and current implementations don't really do much. is this true? is freenet is vibrant project that will surely see version 1.0 someday for sure or not?
[18:54] <cbreak> Freenet has changed a lot in the last two years.
[18:55] <seberino> cbreak: thanks. i don't know what chatroom/mailing lists/etc. my friend is basing his opinions on
[18:55] <cbreak> I know there is a continuing flow of new builds comming from toad, containing lots of bug fixes and new code.
[18:56] <KenMan> toad - depends on what you are doing. For some exams, just constructing the links manually would be sufficient. For others, yeah, build it up progressively. RT links are effectively persisted to disk, i suppose.
[18:56] <toad_> seberino: I think so, on the basis of recent progress
[18:56] <seberino> cbreak, toad_: are the design problems nearly insurmountable? why hasn't it reached version 1.0 in all these years?
[18:56] <cbreak> More recent changes include NGR (Routing), NIO (Reduced Load) and Queueing (Better Balancing)
[18:56] <toad_> seberino: in a couple weeks we're going to launch an appeal for funds, backed by some very interesting simulation results and a network working a lot better than it did a year ago
[18:56] <toad_> it's not perfect, it's not even enough for 0.6 yet
[18:57] <toad_> but it's better than it was
[18:57] <KenMan> the problem has also grown more complex over those years !!
[18:58] <toad_> :)
[18:58] <toad_> some of that is inevitable and necessary
[18:58] <toad_> some of it isn't
[18:58] <seberino> KenMan: but the problem hasn't changed!? it is still to make an anonymous attack resistant p2p system
[18:58] <toad_> there are various things we'd like to do that shift complexity away from the core
[18:58] <toad_> seberino: indeed, and the thing about that is that it's bloody hard
[18:58] <KenMan> as you lay a few bricks down, you learn about rearranging them for tornado resistance.
[18:59] <toad_> if you are doing something new and innovative that nobody has done before, sometimes it takes a while
[18:59] <seberino> toad_: ahhh, see! it *is* a design problem
[18:59] <toad_> seberino: indeed
[18:59] <cbreak> If you count "Anonymous P2P" as Design...
[18:59] <toad_> seberino: we don't really know what the perfect design is
[19:00] <KenMan> the problem has been better defined over time. But it is still being specified, let alone building a solution.
[19:00] <toad_> it's research led to some degree
[19:00] <toad_> that's part of the problem
[19:00] <seberino> toad_: must be really fascinating work to design something so original and cool =)
[19:00] <seberino> toad_: how are you paid to work on freenet? ru a grad student? employed by freenet funds? , etc.
[19:01] <toad_> seberino: I'm paid directly out of freenet funds as a long term contractor
[19:01] <seberino> toad_: do they need more independent contractors? =) i'm interested seriously.
[19:02] <KenMan> we could use some more "independent" contractors, as in "independent of needing $$'s"
[19:02] <toad_> seberino: hmmm. will you work for $2100/mo?
[19:02] <seberino> toad_: part time? =)
[19:02] <toad_> seberino: for more or less full time (~35h/week) ?
[19:02] <toad_> well, part time is an option, but I'm not sure how practical it is...
[19:03] <toad_> depending on how your life works, maybe 35h/wk IS part time :)
[19:03] <toad_> where do you live? the US?
[19:03] <KenMan> hey, i want to get in line ahead of seberino !! I am about to go bankrupt...
[19:04] <toad_> I'll probably ask for another pay rise after we get 0.6 out, as our financial situation is looking rather dubious ATM
[19:04] <toad_> KenMan: for what we pay?
[19:04] <toad_> I don't think taking on extras is entirely out of the question, but right now we only just cover my salary, so we'd need to get value for money
[19:04] <seberino> toad_: couple hundred for 10 hours/week for starters would be good for you perhaps... you could test drive me on small $ now... i have a phd and work in cryptography for the nsa among other things
[19:04] <KenMan> i could almost scrape by at your level of pay !! but I also know Hops can't support 2 full time dudes...
[19:05] <toad_> in any case, anyone who wanted to should establish some credibility first - you got CVS access?
[19:05] <KenMan> now you know you can't trust him. They don't allow their employees to walk around proclaiming what they do.
[19:05] <toad_> LOL
[19:05] <toad_> well
[19:05] <seberino> KenMan: huh?
[19:05] <toad_> I recommend you contact ian
[19:05] <toad_> I certainly can't make this sort of decision on my own
[19:06] <seberino> KenMan: i didn't tell you the details of my project. *that* would be in violation
[19:06] <toad_> I also recommend you make some sort of contribution in advance to gain credibility - that's what I did anyway
[19:06] <cbreak> maybe he is a spy...
[19:07] <KenMan> seberino: uh, yeah. You are exactly right.
[19:07] <cbreak> (OK, since freenet is OSS...)
[19:07] <toad_> heh, yeah..
[19:07] <seberino> cbreak: yea, it is so hard to get a look at freenet code without spies
[19:07] <toad_> LOL
[19:08] <seberino> cbreak: plus, the first thing a spy would do is announce to toad_ where he works
[19:08] * toad_ nods seberino, good point
[19:08] <seberino> cbreak: and give toad_ a url to a resume at pythonsoft.com/resume.pdf =)
[19:09] <KenMan> don't fall for it! His is clearly a triple-agent...
[19:09] <toad_> okay, my recommendations: 1. Make some direct contribution to freenet gratis. 2. Mail ian. I'd certainly be involved in the decision making process, but it's not my call.
[19:09] <KenMan> I know you know that I know what you know, but now do you know what I am thinking ?
[19:10] <KenMan> it's okay if you don't ;)
[19:10] <cbreak> I know I don't know enough... But more than you :)
[19:10] <toad_> there are issues of incentives for volunteer developers - it's straightforward to support me, as everyone trusts me, and I'm full time. But if we start hiring people part time, then there is the risk that people won't contribute at all unless we pay them - right now maybe 1/3rd of dev time spent on freenet is from volunteer devs
[19:11] <KenMan> i work in the counter-counter-counter-espionage office of the CIA, or is it the Russian gov ? I'm always getting those mixed up :)
[19:11] <toad_> I'm not saying it's necessarily a credible risk
[19:11] <toad_> but there are some serious issues to talk through
[19:11] <toad_> and I don't have the authority to say yes anyhow
[19:11] * toad_ reads your resume...
[19:12] <toad_> "ill-defined complex problems requiring independant creative thought"
[19:12] <toad_> nice
[19:12] <toad_> perfect description of freenet :)
[19:12] <seberino> toad_: yea! i was born to help you!
[19:13] <toad_> that's an impressive resume
[19:13] <seberino> toad_: go ahead and give me CVS access then and i'll try to impress you and ian for free first
[19:13] <toad_> seberino: hehe
[19:13] <toad_> well the normal rule is you send me a patch
[19:13] * cbreak is paranoid...
[19:14] <toad_> THEN I give you CVS access, if I like it
[19:14] <seberino> toad_: fine
[19:14] <toad_> that's my rule; we never did it before that, but I am paranoid :)
[19:14] <toad_> s/that/we had me/
[19:14] <seberino> toad_: how about sending me your email and any other info in private then ?
[19:23] * Ash-Fox (Hal-9000@aba101.neoplus.adsl.tpnet.pl) Quit (Read error: 104 (Connection reset by peer))
[19:23] <toad_> okay, first off, set up 25 nodes, fully connected
[19:23] * pupok_ (~r00t@81-178-94-146.dsl.pipex.com) Quit (Read error: 104 (Connection reset by peer))
[19:24] <toad_> then... should we run the requests in parallel to adding more nodes? how many requests per added node?
[19:24] <toad_> probably we should
[19:28] * Ash-Fox (Hal-9000@aba101.neoplus.adsl.tpnet.pl) has joined #FreeNET
[19:32] * pupok_ (~r00t@81-178-94-146.dsl.pipex.com) has joined #freenet
[19:35] <Toast> Toad, howlong does it take to run a line on your graphs?
[19:37] * Ash-Fox (Hal-9000@aba101.neoplus.adsl.tpnet.pl) Quit (Read error: 104 (Connection reset by peer))
[19:43] * TLF (~francisco@91.Red-81-40-116.pooles.rima-tde.net) Quit ("Adi?s a todos y a todas, y que les vaya bien. || http://www.vitaliano.esp.cc")
[19:43] * sanity (~ian@81-178-94-146.dsl.pipex.com) Quit (Read error: 104 (Connection reset by peer))
[19:43] * greycat (rfc1413@wooledge.org) has joined #freenet
[19:46] * Willard (~Willard@195.97.73.194) has joined #freenet
[19:46] * Willard (~Willard@195.97.73.194) has left #freenet
[19:46] * lolo-laptop (~lostlogic@68.251.84.226) Quit ("Client exiting")
[19:47] <toad_> Toast: far too long
[19:47] <toad_> even with the new estimators
[19:47] <toad_> it may be that the DP opteron i'm running the brown and yellow lines on is REALLY busy and that's slowing it down...
[19:48] * sanity (~ian@81-178-94-146.dsl.pipex.com) has joined #freenet
[19:49] * Ash-Fox (Hal-9000@aba101.neoplus.adsl.tpnet.pl) has joined #FreeNET
[19:59] <d-ArkAngel> toad, how do I get the simulator compiled to give it a test run and see how it proccesses on my box? (my java skills are rusty :-) and comming into something in a lang I've not used for so long in the middle is taking some getting used to :-) )
[20:04] <toad_> d-ArkAngel: ant clean jar (you have ant installed?)
[20:04] <toad_> then have a look at the scripts in scripts/simulator
[20:05] <greycat>
[20:05] <greycat> grr, damned fingers.
[20:06] <d-ArkAngel> not sure about ant.... me tries the command :-)
[20:07] <d-ArkAngel> aparently I do :-)
[20:07] <d-ArkAngel> lol 191 errors...
[20:07] * d-ArkAngel looks :-)
[20:08] <toad_> oops :)
[20:08] <toad_> where did you run it?
[20:10] <d-ArkAngel> in the freenet dir
[20:10] <d-ArkAngel> I guess I need to be a level up :-)
[20:11] <d-ArkAngel> hm, that's odd...
[20:12] <d-ArkAngel> oh it's cos I've not got classpath set right... Duh
[20:13] <toad_> :)
[20:14] <toad_> just run ant
[20:14] <toad_> not clean jar
[20:14] <toad_> it'll automatically fetch the prerequisites
[20:14] <toad_> bbl zzz
[20:14] <d-ArkAngel> give me a hint, what DIR is classpath meant to point at on a linux box? jdk/lib?
[20:14] <toad_> (got to get up tomorrow for KK)
[20:14] <toad_> d-ArkAngel: I have no idea :(
[20:14] <d-ArkAngel> ahh, ok I'll guess :-)
[20:14] <toad_> i think you just set JAVA_HOME
[20:14] <toad_> and it'll figure out the classpath
[20:15] <toad_> (you are using Sun, right?)
[20:15] <d-ArkAngel> yeah
[20:16] <toad_> ok got to go
[20:16] <toad_> bbl
[20:16] <d-ArkAngel> c ya. I'll start this box back up in about 8 hrs
[20:21] * ShaunMacPherson (~SMP-Bot@Sudbury-HSE-ppp3976677.sympatico.ca) has left #Freenet
[20:24] * Toast (~Toast@82-34-170-133.cable.ubr03.maid.blueyonder.co.uk) Quit ()
[20:33] * Ash-Fox (Hal-9000@aba101.neoplus.adsl.tpnet.pl) Quit ()
[20:35] * cbreak (~cbreak@dhcp-32-094.via-eth.ch) Quit ("leaving")
[20:40] * Muixirt (~trixium@pD9E615F9.dip.t-dialin.net) Quit ("Gute Nacht !")
[20:51] * Ash-Fox (Hal-9000@abi80.neoplus.adsl.tpnet.pl) has joined #FreeNET
[21:09] * guido^pe (~unknown@dsl-082-082-159-069.arcor-ip.net) Quit (Remote closed the connection)
[21:10] <seberino> toad_: tell me what he thinks about our thoughts if you are going to cc him please by email
[21:22] <greycat> Connections open (Inbound/Outbound/Limit)17 (4/13/32)
[21:22] <greycat> Transfers active (Transmitting/Receiving) 77 (65/12)
[21:22] <greycat> Hmm... interesting.
[21:23] <toad_> 17 conns seems a bit low
[21:23] <toad_> especially for 77 xfers!
[21:23] <toad_> any client activity?
[21:23] <greycat> I'm more amazed that I don't have my "lag symptoms" at the moment.
[21:23] <greycat> yes, I'm inserting and fetching
[21:24] <greycat> with both 5095 and 5096, I've been very slow to acquire new connections. It takes many hours to get up to 32.
[21:25] <greycat> (and often my watchdog will kill the node before it gets that far)
[21:34] <toad_> cool... http://www.linuxjournal.com/article.php?sid=3882 - /me should really learn python some day
[21:35] * seberino (~seberino@seberino3.spawar.navy.mil) Quit (Remote closed the connection)
[21:48] * kers (~kers@5.ppp132.rsd.worldonline.se) has joined #freenet
[21:52] * moskau23 (~Miranda@dsl-082-082-232-109.arcor-ip.net) Quit (Read error: 110 (Connection timed out))
[21:54] <kers> http://www.evilscheme.org/defcon/
[22:52] * hobx_ (~chatzilla@ankh.math.chalmers.se) Quit (Remote closed the connection)
[22:52] * kers (~kers@5.ppp132.rsd.worldonline.se) Quit (Read error: 104 (Connection reset by peer))
[23:38] * goatee_ (~goatee@ip216-239-84-104.vif.net) has joined #freenet
[23:47] * goatee (~goatee@ip216-239-88-178.vif.net) Quit (Read error: 60 (Operation timed out))
[23:56] * goatee_ is now known as goatee
These logs were automatically created by Jay Oliveri with his gimp hapi on irc.freenode.net.
