500 lines
42 KiB
HTML
500 lines
42 KiB
HTML
{% extends "_layout.html" %}
|
|
{% block title %}I2P Development Meeting 107{% endblock %}
|
|
{% block content %}<h3>I2P dev meeting, September 14, 2004</h3>
|
|
<div class="irclog">
|
|
<p>14:06 < jrandom> 0) hi</p>
|
|
<p>14:06 < jrandom> 1) 0.4.0.1</p>
|
|
<p>14:06 < jrandom> 2) Threat model updates</p>
|
|
<p>14:06 < jrandom> 3) Website updates</p>
|
|
<p>14:06 < jrandom> 4) Roadmap</p>
|
|
<p>14:06 < jrandom> 5) Client apps</p>
|
|
<p>14:06 < jrandom> 6) ???</p>
|
|
<p>14:06 < jrandom> 0) hi</p>
|
|
<p>14:06 * jrandom waves</p>
|
|
<p>14:06 < cervantes> evening</p>
|
|
<p>14:06 < jrandom> weekly status notes posted to http://dev.i2p.net/pipermail/i2p/2004-September/000444.html</p>
|
|
<p>14:07 < jrandom> (before the meeting this time too ;)</p>
|
|
<p>14:07 < deer> <jrand0m> woah, 30 people over here</p>
|
|
<p>14:07 -!- Irssi: #i2p: Total of 21 nicks [0 ops, 0 halfops, 0 voices, 21 normal]</p>
|
|
<p>14:07 < jrandom> ok, anyway, lets jump right in to 1) 0.4.0.1</p>
|
|
<p>14:08 < jrandom> the release is out and things seem to be working more or less</p>
|
|
<p>14:09 < jrandom> i see a variety of connection times on irc, though in discussions with people, it seems there are congestion issues when e.g. downloading large files and using irc at the same time</p>
|
|
<p>14:09 < jrandom> are many people running into that?</p>
|
|
<p>14:10 < jrandom> i guess not</p>
|
|
<p>14:11 < cervantes> I've been doing various bandwidth tests recently and haven't encounter problems in that area yet...although I'm not using the bandwidth limiter</p>
|
|
<p>14:11 * nicktastic hasn't downloaded much since raiding alexandria weeks ago</p>
|
|
<p>14:11 < dm> I remember getting disconnect more often on IRC when I was using eepsites, but that was 2 months ago</p>
|
|
<p>14:11 < dm> disconnected</p>
|
|
<p>14:11 < dm> not sure if it still happens</p>
|
|
<p>14:11 < jrandom> ah, yeah, we need to harass the alexandria folks to give us more books :)</p>
|
|
<p>14:12 < Nightblade> thanks for keeping us up to date dm</p>
|
|
<p>14:12 < jrandom> i've had good luck w/ irc while downloading some large files from thetower, but, like cervantes, i dont have bandwidth limiting set</p>
|
|
<p>14:13 < jrandom> (though that router's bw average was a steady 11KBps at the time, while downloading 8KBps of music)</p>
|
|
<p>14:13 * nicktastic finds something to download</p>
|
|
<p>14:13 * jrandom watches your irc.duck.i2p connection quickly get dropped ;)</p>
|
|
<p>14:13 < jrandom> ok, anyway, does anyone have anything else they want to bring up wrt 0.4.0.1?</p>
|
|
<p>14:14 < dm> Nightblade: hehe, no problem :)</p>
|
|
<p>14:14 < dm> jrandom: good work, ever onwards</p>
|
|
<p>14:14 < fvw> the installer is pretty? (not sure if that's new in .1?)</p>
|
|
<p>14:14 < jrandom> gracias dm</p>
|
|
<p>14:15 < jrandom> fvw: same as 0.4, but i agree, hyper did some great work there (as did our anonymous designer!)</p>
|
|
<p>14:15 < fvw> also, I'm not going to commit myself as to pretty _what_ it is :)</p>
|
|
<p>14:15 < jrandom> sonofabi...</p>
|
|
<p>14:16 < jrandom> ok, moving on to 2) Threat model updates</p>
|
|
<p>14:16 < cervantes> yes well done.. :) writing documentation always sucks </p>
|
|
<p>14:17 < jrandom> yeah, it was a painful 2-3 days</p>
|
|
<p>14:17 < jrandom> i'm not sure if any of y'all have read http://www.i2p.net/how_threatmodel but if you ever want to know wtf we're talking about when we say "anonymous", thats what we mean</p>
|
|
<p>14:18 < jrandom> most of the categories there were just ripped from http://citeseer.ist.psu.edu/454354.html (linked to on the page)</p>
|
|
<p>14:18 < jrandom> there's a lot more i'd like to do in the threat model, but i just dont have the time.</p>
|
|
<p>14:18 < jrandom> i'd love to see a matrix of those threats vs cost of mounting them vs the type of user who cares about them</p>
|
|
<p>14:19 < jrandom> (e.g. joe sixpack does't care about global active adversaries)</p>
|
|
<p>14:19 < jrandom> so if anyone is bored... ;)</p>
|
|
<p>14:19 < cervantes> something that occurred to me whilst reading your doc... we need a decent glossary...</p>
|
|
<p>14:20 < fvw> doesn't he? joe sixpack likes to download mp3s...</p>
|
|
<p>14:20 < jrandom> someone just published one iirc...</p>
|
|
<p>14:20 < cervantes> really?</p>
|
|
<p>14:20 < cervantes> on an eep?</p>
|
|
<p>14:20 < jrandom> no, some research paper</p>
|
|
<p>14:20 < jrandom> its not on freehaven yet, lemmie dig it up</p>
|
|
<p>14:21 < jrandom> bugger, i dont seem to have my copy anymore. </p>
|
|
<p>14:21 < jrandom> i'll try to track it down after the meeting</p>
|
|
<p>14:22 < cervantes> does it tackle i2p specific concepts to?</p>
|
|
<p>14:22 < jrandom> oh, no</p>
|
|
<p>14:22 < jrandom> its just a general glossary for anonymous networks, dealing with mixes, cascades, attackers, etc</p>
|
|
<p>14:22 < jrandom> no garlic routing or tunnels ;)</p>
|
|
<p>14:23 < cervantes> a nice single paragraph summary of all "in" buzzwords so people can quickly see the difference between onion and garlic routing (for example) withou having to read the whole "how" document</p>
|
|
<p>14:23 < jrandom> you realize a glossary would be larger than the how_* pages combined, right? </p>
|
|
<p>14:23 < jrandom> but yeah, i agree, we should do that</p>
|
|
<p>14:23 < cervantes> sure... but.. ;)</p>
|
|
<p>14:23 * jrandom volunteers cervantes to work on it ;)</p>
|
|
<p>14:23 * dm concurs</p>
|
|
<p>14:23 < cervantes> hehe I don't know what half that shit means :)</p>
|
|
<p>14:24 < jrandom> write up what you do know and ask me questions</p>
|
|
<p>14:24 < cervantes> I'll have a crack at it</p>
|
|
<p>14:24 < jrandom> w00t! cervantes++</p>
|
|
<p>14:24 < cervantes> if I put it on the forum then others can contribute...</p>
|
|
<p>14:24 < jrandom> good idea</p>
|
|
<p>14:24 < deer> * Pseudonym cheers</p>
|
|
<p>14:25 < cervantes> _but_ that doc you mentioned would be handy :o)</p>
|
|
<p>14:25 < dm> tunnel: artificial underground passage</p>
|
|
<p>14:25 < jrandom> agreed, i'll try to find it again</p>
|
|
<p>14:25 < cervantes> I'll do a special version for you dm</p>
|
|
<p>14:25 < dm> yay!</p>
|
|
<p>14:26 < jrandom> ok, anything else on the threat model, or shall we move on to 3) Website updates ?</p>
|
|
<p>14:27 < jrandom> ok, as anyone who has been to the site today has seen, Curiosity has come up with some nice usability updates</p>
|
|
<p>14:27 < dm> I think cervantes and I are the only ones still awake.</p>
|
|
<p>14:27 < korkakak> I think that in threat models</p>
|
|
<p>14:28 < korkakak> you should add some mixnetwork attacks</p>
|
|
<p>14:28 < jrandom> what sort of mix attacks?</p>
|
|
<p>14:28 * dm loads up www.i2p.net</p>
|
|
<p>14:28 < korkakak> like collusion attacks</p>
|
|
<p>14:28 < jrandom> thats the thing that sucks about the taxonomies i used. they're all pretty much collusion attacks.</p>
|
|
<p>14:29 < korkakak> With mix attacks i mean attacks that may happen in a mix network</p>
|
|
<p>14:29 < korkakak> ah ok sorry</p>
|
|
<p>14:29 < jrandom> (and most can be used for either probabalistic or confirmation attacks, etc)</p>
|
|
<p>14:29 < dm> I like the increasing-in-size paragraphs. Helps drag people in. Far too technical for a front page though.</p>
|
|
<p>14:29 < korkakak> Another 5 cents from me: Can i2p detect a collusion automatically?</p>
|
|
<p>14:30 < jrandom> but if you have some suggestions for things we need to add, please, let me know</p>
|
|
<p>14:30 < jrandom> oh, definitely not. we haven't imported morphmix's algorithms</p>
|
|
<p>14:30 < korkakak> I c</p>
|
|
<p>14:30 < korkakak> ok keep on</p>
|
|
<p>14:30 < jrandom> though theirs wouldn't really fly with us though, since we're a free route mixnet</p>
|
|
<p>14:31 < korkakak> Well yes and no</p>
|
|
<p>14:31 < korkakak> but it is ok. SOrry for the interrutp</p>
|
|
<p>14:32 < cat-a-puss> It might also be a good idea to mention up frount some of the obvious attacks that I2p is NOT vunerable to</p>
|
|
<p>14:32 < jrandom> hmm? their algorithms are based off detecting the influence of colluding peers in the peer selection - within i2p, the local router explicitly defines the entire peer selection algorithm</p>
|
|
<p>14:33 < korkakak> I guess that this is true due to the size of todays network</p>
|
|
<p>14:33 < jrandom> ah, thats a good idea cat-a-puss, w/ MITM/etc. would you be interested in posting up some ideas for that?</p>
|
|
<p>14:33 < cat-a-puss> sure</p>
|
|
<p>14:33 < dm> MITM?</p>
|
|
<p>14:33 < dm> Ah, man in the middle.</p>
|
|
<p>14:33 < jrandom> muchas gracias cat-a-puss!</p>
|
|
<p>14:34 * cervantes jots down MITM for the glossary</p>
|
|
<p>14:34 < jrandom> korkakak: hmm. i'm not sure how that aspect is affected by the size of the net, but there may be things we can learn from morphmix's collusion detection, certainly</p>
|
|
<p>14:34 < jrandom> (perhaps wrt the netDb responses, for instance)</p>
|
|
<p>14:34 < korkakak> wrt = ?</p>
|
|
<p>14:35 < dm> hehee</p>
|
|
<p>14:35 < jrandom> sorry, with regards to</p>
|
|
<p>14:35 < dm> I know that one!</p>
|
|
<p>14:36 < jrandom> we would certainly benefit from more discussion on the threat model. perhaps we can start up a thread on the list or in the forum?</p>
|
|
<p>14:36 < dm> "The result is that the number of peers relaying each end to end message is the absolute minimum necessary to meet both the sender's and the receiver's threat model."</p>
|
|
<p>14:36 < dm> I like this way of looking at it.</p>
|
|
<p>14:37 < dm> Although it's not true.</p>
|
|
<p>14:37 < jrandom> hmm? </p>
|
|
<p>14:37 < jrandom> if both sender and receiver want only plausible deniability, they can talk directly</p>
|
|
<p>14:37 < jrandom> (etc)</p>
|
|
<p>14:37 < dm> The absolute minimum number of peers required to meet the threat model of A and B is the number of peers required by A or B, whichever has more stringent requirements :)</p>
|
|
<p>14:38 < jrandom> not true dm</p>
|
|
<p>14:38 < jrandom> if they both require 2 hop tunnels to defend against colluding attackers in their tunnels, they can't both have 1 hop tunnels</p>
|
|
<p>14:39 < dm> If A is willing to talk to someone with 10 indirections, and B is willing with 5, the minimum needed is 10, not 15!?</p>
|
|
<p>14:39 < jrandom> no, 15. B shouldn't trust A's tunnels.</p>
|
|
<p>14:39 < dm> Yeah, he shouldn't.</p>
|
|
<p>14:39 < dm> But theoratically.. Anyway, stupid discussion. I like that sentence though.</p>
|
|
<p>14:40 < jrandom> its one of the more important design decisions in i2p ;)</p>
|
|
<p>14:40 < jrandom> anyway, back to 3) Website updates</p>
|
|
<p>14:41 < deer> <nicktastic> (fyi - irc dropped while downloading two large files, but latency to the server is as it was before the downloads started, so could've been a fluke (ungraceful shutdown somewhere?))</p>
|
|
<p>14:41 < jrandom> Curiosity and I discussed the length of the new homepage, and while we all agree that its a little long, its better than the old 1 liner</p>
|
|
<p>14:41 < cervantes> agreed</p>
|
|
<p>14:42 < jrandom> ah ok. perhaps even network congestion while downloading, since the eepproxy and the irc client use the same I2P destination (by default)</p>
|
|
<p>14:42 < nicktastic> Aaah....</p>
|
|
<p>14:42 < jrandom> (so both would be trying to use the same pair of inbound tunnels)</p>
|
|
<p>14:42 < nicktastic> I was wondering why only one showed up</p>
|
|
<p>14:43 < jrandom> yeah, thats the default within I2PTunnel and the ministreaming lib. perhaps if someone cares we can expose a way to configure that ;)</p>
|
|
<p>14:43 < nicktastic> sorry to interrupt</p>
|
|
<p>14:43 < deer> * Pseudonym cares</p>
|
|
<p>14:43 < dm> such polite lads we have in this room</p>
|
|
<p>14:43 < interrupt> you are forgiven</p>
|
|
<p>14:44 < interrupt> ;)</p>
|
|
<p>14:44 * nicktastic rolls eyes</p>
|
|
<p>14:44 < jrandom> patches welcome Pseudonym ;) (naw, i'll see if i can find an easy way.. shouldnt be too hard)</p>
|
|
<p>14:44 < jrandom> ok, anyway</p>
|
|
<p>14:44 < deer> <Pseudonym> good, 'cause I don't know crap about how to code in java</p>
|
|
<p>14:45 < jrandom> there may be further website updates, but if anyone has any suggestions, please post 'em to the forum or the list, or point 'em out to Curiosity on irc and we'll get things rolling</p>
|
|
<p>14:45 < jrandom> anyone have anything they want to bring up wrt the website?</p>
|
|
<p>14:45 < cervantes> umm bouties perhaps</p>
|
|
<p>14:46 < cervantes> although maybe that's best saved for 5</p>
|
|
<p>14:46 < jrandom> prolly so</p>
|
|
<p>14:46 < jrandom> ok, moving on to 4) Roadmap</p>
|
|
<p>14:46 < jrandom> lots of updates. see the email for info</p>
|
|
<p>14:47 < jrandom> (or look at the pretty gantt chart ;)</p>
|
|
<p>14:47 < dm> Was that done in MS Project?</p>
|
|
<p>14:47 < jrandom> http://ganttproject.sourceforge.net/</p>
|
|
<p>14:47 < cervantes> eerm gantt :)</p>
|
|
<p>14:47 < dm> oh.. gantt is a product. My bad.</p>
|
|
<p>14:48 < dm> Nice to see there are no dependencies in the roadmap.</p>
|
|
<p>14:48 < jrandom> i've posted a few different revs of the roadmap over the last few days, but this one seems to be solid</p>
|
|
<p>14:48 < cervantes> it's all dependant on jrandom ;-)</p>
|
|
<p>14:48 * jrandom whimpers</p>
|
|
<p>14:48 < fvw> 3.0 in febuary? Wow.</p>
|
|
<p>14:48 < jrandom> the 2.0 and 3.0 releases are really just 1 (big) feature each</p>
|
|
<p>14:48 < dm> Don't forget: exponential versioning</p>
|
|
<p>14:49 < jrandom> heh</p>
|
|
<p>14:49 < jrandom> yeah, we'll be 1183 by next july</p>
|
|
<p>14:50 < dm> Well, it's more interesting than the abritrary +0.1 per build of most projects, so I'm not complaining.</p>
|
|
<p>14:50 < jrandom> the 2.0 and 3.0 releases may be delayed to stay in line with other apps though. e.g. 3.0 would work great with an email app</p>
|
|
<p>14:51 < jrandom> the release criteria for 1.0 has been the usual - functional, secure, scalable, and anonymous</p>
|
|
<p>14:51 < jrandom> thats why i moved the udp transport in, as our current tcp transport would shit bricks if we had a few thousand peers</p>
|
|
<p>14:51 < dm> so we should have a 0.9 - The Slashdot</p>
|
|
<p>14:51 < dm> if it survives we can check off scalable and move to 1.0</p>
|
|
<p>14:51 < jrandom> heh</p>
|
|
<p>14:52 * jrandom would rather grow organically, thankyouverymuch</p>
|
|
<p>14:52 < cervantes> we don't to tell _them_ about it</p>
|
|
<p>14:52 < cervantes> *don't want</p>
|
|
<p>14:52 < korkakak> btw may i say something about the global timing?</p>
|
|
<p>14:52 < cervantes> let them all stay on the internet while we move here</p>
|
|
<p>14:52 < jrandom> sure korkakak</p>
|
|
<p>14:53 < korkakak> As far as i am concerned you cannot simulate a synchronus network over an asynchronous</p>
|
|
<p>14:53 < korkakak> it is just bad design and should lead to network splits [i think] in the way it is used</p>
|
|
<p>14:54 < korkakak> as a timestamp for UDP packets</p>
|
|
<p>14:54 < jrandom> the timing is not synchronized for messaging, merely to help us know the freshness of data</p>
|
|
<p>14:54 < korkakak> yes that's the point</p>
|
|
<p>14:54 < jrandom> without knowing the freshness of the netDb entries, you're vulnerable to a whole slew of attacks</p>
|
|
<p>14:55 < korkakak> Yes</p>
|
|
<p>14:55 < korkakak> but imagine a growing network</p>
|
|
<p>14:55 < korkakak> a huge network</p>
|
|
<p>14:55 < jrandom> like the internet</p>
|
|
<p>14:55 < dm> bigger!</p>
|
|
<p>14:55 < fvw> two internets tied together with bits of string!</p>
|
|
<p>14:55 < jrandom> that has a network time protocol for scaling to such sizes... ;)</p>
|
|
<p>14:56 < korkakak> I don't think I understand your point but</p>
|
|
<p>14:56 < dm> korkakak: what are you trying to say?</p>
|
|
<p>14:57 < korkakak> that net splits may happend due to invalid timestamps</p>
|
|
<p>14:58 * dm is not sure how syncing works currently</p>
|
|
<p>14:58 < korkakak> the case is called localization effect [english translation from greek]</p>
|
|
<p>14:58 < deer> <soros> i hear i2p's anonymity has been cracked</p>
|
|
<p>14:59 < deer> <soros> true ?</p>
|
|
<p>14:59 < jrandom> i believe we can address the time sync issue the same way the NTP networks do. there are a massive number of tier 2 and 3 NTP hosts, and while our current SNTP implementation is of course unsuitable for congested environments, there is no reason to believe time synchronization isn't possible</p>
|
|
<p>14:59 < jrandom> heh soros</p>
|
|
<p>14:59 < jrandom> soros: the thread you're referring to (someone else mentioned it to me) on devl was talking about JAP being compromised, not I2P.</p>
|
|
<p>15:00 < dm> so all I2P nodes must stay synced at all times for it to work?</p>
|
|
<p>15:00 < korkakak> NTP nets are synhcronus networks over synchronus networks ;-)</p>
|
|
<p>15:00 < jrandom> but if someone has a compromise for I2P, I would certainly love to hear about it</p>
|
|
<p>15:00 < deer> <soros> i have one but i'm keeping it a secret</p>
|
|
<p>15:00 < jrandom> at various levels of abstraction korkakak, sure. my ethernet cable is synchronized too</p>
|
|
<p>15:01 < deer> <soros> :)</p>
|
|
<p>15:01 < jrandom> yes dm, synchronized to the network time</p>
|
|
<p>15:01 < korkakak> jrandom it is nick or korki :-)</p>
|
|
<p>15:01 < jrandom> (the point is that we dont use synchronous messaging)</p>
|
|
<p>15:01 < jrandom> :) 'k</p>
|
|
<p>15:01 < jrandom> (please dont be offended if i dont tell you my name ;)</p>
|
|
<p>15:02 < korkakak> No I am not</p>
|
|
<p>15:02 < dm> His name is Abdul</p>
|
|
<p>15:02 < jrandom> ok where were we</p>
|
|
<p>15:02 < nicktastic> 4)</p>
|
|
<p>15:03 < jrandom> ok right, thanks. the roadmap</p>
|
|
<p>15:03 < jrandom> anyone have any concerns / ideas / suggestions?</p>
|
|
<p>15:03 < dm> so when you say some work is going to be done on the transport, do you mean reworking TCP, or moving to UDP?</p>
|
|
<p>15:04 < jrandom> UDP is 0.4.4</p>
|
|
<p>15:05 < dm> I thought I saw something about work on the transport layer</p>
|
|
<p>15:05 < dm> in the near future</p>
|
|
<p>15:05 < jrandom> yes, 0.4.1 will be a revamp of the TCP transport</p>
|
|
<p>15:05 < dm> why revamp TCP in 0.4.1 if going UDP in 0.4.4?</p>
|
|
<p>15:05 < dm> We'll need both?</p>
|
|
<p>15:05 < cervantes> only to point out that your are still the only resource in the project plan... ...are we suffering from a lack of contributors or just project fragmentation?</p>
|
|
<p>15:06 < jrandom> dm: some people cannot use UDP</p>
|
|
<p>15:06 < dm> firewalls?</p>
|
|
<p>15:06 < jrandom> cervantes: we certainly could parallelize many of those tasks with more contributors</p>
|
|
<p>15:07 < jrandom> (but the roadmap does not assume more)</p>
|
|
<p>15:07 < cervantes> so hopefully it represents the worse case scenario</p>
|
|
<p>15:07 < jrandom> there is however other important work going on not reflected on the roadmap, such as client mods, services on top of i2p, etc</p>
|
|
<p>15:08 < cervantes> asside from you being assassinated</p>
|
|
<p>15:08 < dm> I wish we could afford toad</p>
|
|
<p>15:08 < deer> <Pseudonym> now that 0.4 is out and pretty much working, should we announce somewhere (not necessarily /.) to try to increase the number of developers/testers/donors?</p>
|
|
<p>15:08 < jrandom> more contributors would certainly be welcome</p>
|
|
<p>15:08 * korkakak farewells all. REady to go to his bed. It is late at korkakak lands...</p>
|
|
<p>15:08 < korkakak> bye gayz</p>
|
|
<p>15:08 < cervantes> g'night</p>
|
|
<p>15:08 < jrandom> thanks for swinging by nick, ttyl</p>
|
|
<p>15:10 < dm> nite</p>
|
|
<p>15:10 < jrandom> a /. would probably be premature, but it would be good to bring new folks on board through other means</p>
|
|
<p>15:10 < dm> You're very open to Pseudonym's suggestion. I thought you were going to freak out.</p>
|
|
<p>15:10 < jrandom> but i think through word of mouth we're growing steadily</p>
|
|
<p>15:11 < deer> <Pseudonym> and if we do want to announce, where should we do it?</p>
|
|
<p>15:11 < jrandom> i dont think we should have any announcements yet, not till 1.0</p>
|
|
<p>15:11 < deer> <Pseudonym> seems like we could use an influx of cash/talent</p>
|
|
<p>15:11 < jrandom> but if you hear someone talking about how they wish there were some way to help do stuff anonymously, point 'em at i2p ;)</p>
|
|
<p>15:12 < deer> * DrWoo suggests a whisper campaign</p>
|
|
<p>15:12 < cervantes> we have a fair amount of unnallocated cash already...</p>
|
|
<p>15:12 < jrandom> we're an open team, but you only have one chance to make a first impression.</p>
|
|
<p>15:13 < cat-a-puss> I would not recomend going from no publicity to /. there needs to be an intermediate step to make sure we can handel the load</p>
|
|
<p>15:13 < deer> <Pseudonym> then we should allocate it to bounties we think are important</p>
|
|
<p>15:13 < dm> We need to hire a full-time dev or find someone REALLY REALLY bored</p>
|
|
<p>15:13 < jrandom> agreed. i'd like to see at least 500 routers online prior</p>
|
|
<p>15:13 < jrandom> actually, y'all are moving us right along to 5) Client apps :)</p>
|
|
<p>15:14 < jrandom> we do have ~300 in the pot at the moment (well, almost, but thats another story)</p>
|
|
<p>15:14 < deer> <Pseudonym> any suggestions on what the intermediate step could be?</p>
|
|
<p>15:14 < jrandom> pseudonym: we can't have 1000s of nodes until 0.4.4</p>
|
|
<p>15:15 < jrandom> (and we'd want to stress test the net out first)</p>
|
|
<p>15:15 < fvw> Actually, we probably can on most unices. Needs adjusting the rlimits though.</p>
|
|
<p>15:15 < jrandom> right right</p>
|
|
<p>15:15 < jrandom> it'd be painful, anyway ;)</p>
|
|
<p>15:16 < deer> <Pseudonym> right. so no /. but it seems like there should be somewhere we can get a couple hundred</p>
|
|
<p>15:16 < jrandom> we can do larger sims though</p>
|
|
<p>15:16 < deer> <Pseudonym> does anyone know someone at the EFF? maybe they have a mailing list</p>
|
|
<p>15:17 < jrandom> yeah, i've spoken with some eff folks about some things</p>
|
|
<p>15:17 < fvw> I think any announcement will cause it to filter through to slashdot. I agree with jrandom, a little waiting isn't bad at this time.</p>
|
|
<p>15:18 < dm> you have to be aware that if you hit 200-300 nodes, you're most likely to get an automatic /. mention ;)</p>
|
|
<p>15:18 < jrandom> (especially since we've been going for ~ 15 months already)</p>
|
|
<p>15:18 < dm> critical mass/hype and all that</p>
|
|
<p>15:18 < jrandom> well, thats also another thing that leads in to 5) Client apps</p>
|
|
<p>15:19 < jrandom> i'm watching some stats and it seems probably 1/4th of the routers out there aren't even really doing any client activity</p>
|
|
<p>15:19 < jrandom> which is great and wonderful that people are willing to donate their resources to act as I2P routers, but its too bad that we dont have something to suck them in :)</p>
|
|
<p>15:19 < fvw> Yeah, I'd like to do a proper chat app (as in irc, but in a way that makes sense for i2p), but this is very much a long term thing, no time the next few months...</p>
|
|
<p>15:20 < jrandom> we have had an influx of kickass eepsites recently though</p>
|
|
<p>15:20 < jrandom> ah cool fvw</p>
|
|
<p>15:20 < cervantes> many people run more than 1 router though</p>
|
|
<p>15:20 < jrandom> a solid IM/group chat for I2P would certainly rule</p>
|
|
<p>15:20 < nicktastic> fvw: Instant messenger with multi-user chat, perhaps?</p>
|
|
<p>15:20 < deer> <mrflibble> dudes, in 0.4.0.1, how do i allow the router to listen on more than just localhost?</p>
|
|
<p>15:20 < cat-a-puss> hey, could someone write a gaim plugin? that would be a good way to do it</p>
|
|
<p>15:20 < jrandom> right cervantes </p>
|
|
<p>15:20 < cervantes> they maybe use 1 for apps...and donate the others</p>
|
|
<p>15:21 < jrandom> mrflibble: http://localhost:7657/i2ptunnel/ to configure the http and irc proxies to listen on "any interface"</p>
|
|
<p>15:21 < fvw> which reminds me: could we do something multicastish for outbound tunnels? ie have one message delivered to multiple inbouds?</p>
|
|
<p>15:21 < nicktastic> cat-a-puss: Certainly possible</p>
|
|
<p>15:21 < fvw> yeah, in essence there's not much difference between irc and im, apart from the user interface.</p>
|
|
<p>15:22 < jrandom> fvw: yes and no. it wouldn't offer much savings (as messages are end to end encrypted, so you'd have to garlic wrap the message to the outbound tunnel's endpoint and direct the cloves seperately from there)</p>
|
|
<p>15:22 < jrandom> imho multicast would want to use an application layer overlay</p>
|
|
<p>15:22 < deer> <mrflibble> oh, thanks jrandom!</p>
|
|
<p>15:22 < fvw> what do you mean by application layer overlay?</p>
|
|
<p>15:22 < jrandom> ala shoutcast/etc</p>
|
|
<p>15:23 < hypercubus> he means do the multicasting in the applcation layer</p>
|
|
<p>15:23 < hypercubus> not in the i2p layer</p>
|
|
<p>15:23 < cervantes> 'lo hyper</p>
|
|
<p>15:23 < fvw> yes ok. Fair enough.</p>
|
|
<p>15:24 < jrandom> ok, I ranted enough in the email about the client apps, so I'm not going to repeat myself here.</p>
|
|
<p>15:25 * fvw pouts and puts away the popcorn.</p>
|
|
<p>15:25 * jrandom !thwaps the wiseass</p>
|
|
<p>15:26 < jrandom> but, basically I think before we go "live", we need something engaging to go live *with*</p>
|
|
<p>15:26 < dm> If you build it, they will come...</p>
|
|
<p>15:26 < dm> hahaha, or not!!!</p>
|
|
<p>15:26 < fvw> yes. Though we could probably pull quite some crowd from freenet just by having dynamic (not to mention _working_) freesites.</p>
|
|
<p>15:27 < deer> <Pseudonym> what about using some of the money in the general fund to create/increase bounties for the engaging stuff</p>
|
|
<p>15:27 < nicktastic> ...and dht</p>
|
|
<p>15:27 < cervantes> I have no knowledge of freenet... how do freesites differ to eepsites?</p>
|
|
<p>15:27 < cervantes> if they are in any way the same</p>
|
|
<p>15:27 < deer> <Pseudonym> eepsites work</p>
|
|
<p>15:28 < deer> <soros> heh</p>
|
|
<p>15:28 < hypercubus> imo you guys are impatient</p>
|
|
<p>15:28 < cervantes> apart from that</p>
|
|
<p>15:28 < nicktastic> hypercubus: How's that?</p>
|
|
<p>15:28 < hypercubus> contribute to the project, or shut up</p>
|
|
<p>15:28 < deer> <soros> freesites are static.</p>
|
|
<p>15:28 < jrandom> bounties/voting some of the general fund to give $$$ to service providers / eepsites that do kickass things does sound like a good idea</p>
|
|
<p>15:28 * jrandom is the impatient one hypercubus ;)</p>
|
|
<p>15:28 < jrandom> Pseudonym: is that what you mean?</p>
|
|
<p>15:28 < cervantes> these applications are certainly not going to materialise overnight</p>
|
|
<p>15:29 < jrandom> right, thats why we need to talk about it now</p>
|
|
<p>15:29 < jrandom> duck: you 'round?</p>
|
|
<p>15:29 < hypercubus> it's these people pushing for public announcements</p>
|
|
<p>15:29 < fvw> I doubt you'll get more eepsites with bounties. The people who build them do it because it's fun, I doubt we could pay those who don't find it fun enough.</p>
|
|
<p>15:29 < deer> <soros> dynamic freesites can be updated, but only once a day... </p>
|
|
<p>15:29 < jrandom> probably true fvw</p>
|
|
<p>15:29 < deer> <Pseudonym> I was thinking more of using the general fund to support bounties for apps, not services/eepsites</p>
|
|
<p>15:29 < fvw> nobody's pushing for announcements, it was just discussed briefly.</p>
|
|
<p>15:30 < hypercubus> the project is evolving and growing naturally, have patience</p>
|
|
<p>15:30 < jrandom> ok word Pseudonym. </p>
|
|
<p>15:30 * fvw nods at pseudonym. That might be good yes.</p>
|
|
<p>15:30 < jrandom> what would y'all suggest?</p>
|
|
<p>15:30 < nicktastic> hypercubus: They're just brainstorming ways to grow the network without GROWING the network ;)</p>
|
|
<p>15:30 < jrandom> the entire donation pool is available to be applied wherever we see fit</p>
|
|
<p>15:30 < fvw> though I think small bug or feature bounties have the greatest chance of actually causing stuff to happen as opposed to being a nice gift for the person who happened to do it anyway.</p>
|
|
<p>15:31 < deer> <Pseudonym> small bounties don't seem to be working. how about we push a bunch of money into the MyI2p pot</p>
|
|
<p>15:32 < hypercubus> how about you donate?</p>
|
|
<p>15:32 < nicktastic> jrandom: Well, for swarming file transfer and dds to be useful, we need streams faster than 4kbyte/sec, so two bounties are fairly dependent on the streaming library bounty</p>
|
|
<p>15:32 < nicktastic> jrandom: But from earlier discussion, that sounds rather trivial</p>
|
|
<p>15:32 < cervantes> throwing money at things isn't going to make stuff appear overnight either :)</p>
|
|
<p>15:32 < deer> <Pseudonym> I have donated</p>
|
|
<p>15:32 < deer> <soros> just announce i2p to slashdot</p>
|
|
<p>15:32 < deer> <soros> thats all you need</p>
|
|
<p>15:33 < hypercubus> that is exactly the opposite of what we need</p>
|
|
<p>15:33 < deer> <Pseudonym> not overnight, but maybe somebody will start working on it</p>
|
|
<p>15:33 < jrandom> nicktastic: the streaming lib will be lots of work, but thats the 0.4.3 release :)</p>
|
|
<p>15:34 * nicktastic consults roadmap</p>
|
|
<p>15:34 < jrandom> but I agree with cervantes, $$ doesn't make code, coders make code.</p>
|
|
<p>15:34 < deer> <soros> is i2p listed on freshmeat ?</p>
|
|
<p>15:34 < jrandom> if only there were some magic way to get in touch with hackers without letting general users know ;)</p>
|
|
<p>15:34 < jrandom> not to my knowledge soros</p>
|
|
<p>15:34 < fvw> cross-post to other anonymity-related software mailinglists?</p>
|
|
<p>15:35 < fvw> actually, I think most of the people where already involved with freenet or gnunet, and have become aware of i2p already.</p>
|
|
<p>15:35 < cervantes> hack into their inferior anonymity networks and say "hi come and work for us"</p>
|
|
<p>15:35 < jrandom> we do get a good # of hits from gnunet's links page</p>
|
|
<p>15:35 < jrandom> heh cervantes </p>
|
|
<p>15:35 < deer> <demonic_1> there r some ng's i would think</p>
|
|
<p>15:36 < cervantes> (work for us or we'll give your ip to big brother)</p>
|
|
<p>15:36 < cat-a-puss> you could put refrences to I2p in wikis talking about related things</p>
|
|
<p>15:36 < deer> <baffled> I think one thing we need is some way to get mail in to i2p and anonymously out of it.</p>
|
|
<p>15:36 < jrandom> i think someone has already placed i2p at various spots in wikipedia, though i dont know about iA lately</p>
|
|
<p>15:36 * fvw doesn't see why you couldn't run smtp over a tunnel.</p>
|
|
<p>15:37 < jrandom> agreed baffled, a solid way to do mail *securely* would be great</p>
|
|
<p>15:37 < cervantes> is that possible though</p>
|
|
<p>15:37 < fvw> we must be careful not to spam though.</p>
|
|
<p>15:37 < jrandom> fvw: do you trust your mail client?</p>
|
|
<p>15:37 < jrandom> however, a mixminion/mixmaster outbound gateway would *rule*</p>
|
|
<p>15:37 < jrandom> (so someone go set up a web interface to one of those. please :)</p>
|
|
<p>15:37 < fvw> jrandom: as much as I trust any other client software... Do you trust your IRC client? your web browser? ...</p>
|
|
<p>15:38 < cervantes> you'd have to trust the guy who owns the gateway isn't reading your mail</p>
|
|
<p>15:38 < jrandom> fvw: no.</p>
|
|
<p>15:38 < jrandom> fvw: and thats a problem.</p>
|
|
<p>15:38 < jrandom> fvw: a problem we must fix before we can recommend that people use I2P for anything beyond testing.</p>
|
|
<p>15:39 < fvw> How do you suggest making mail clients "more anonymous"?</p>
|
|
<p>15:39 < jrandom> it'd need to be a local SMTP/POP3 "server" that reads from the client, parses, interprets, and acts accordingly.</p>
|
|
<p>15:39 < cervantes> you'd need a bespoke mail application for a start</p>
|
|
<p>15:39 < jrandom> (there are a few apps out there that do that already)</p>
|
|
<p>15:39 < cervantes> (client)</p>
|
|
<p>15:40 * cervantes apologies for saying "bespoke"</p>
|
|
<p>15:40 < cervantes> *apologises</p>
|
|
<p>15:40 < jrandom> but that gets to one of the points in the weekly status notes - there are just so many important things that need to get done</p>
|
|
<p>15:40 < fvw> jrandom: That'd be very easy, at least on unix. Just hack up a sendmail drop in and something that does fetchmail and you're there. Could be done in shell scripts if you wanted to.</p>
|
|
<p>15:40 < deer> <duck> me hears an echo of his name</p>
|
|
<p>15:40 < jrandom> we need to focus if the bounties are going to be sufficient</p>
|
|
<p>15:40 < jrandom> oh, heya duck</p>
|
|
<p>15:41 < deer> <duck> sorry, I was euh.. drinking'</p>
|
|
<p>15:41 < jrandom> duck: just wanted to check in to see if there was any update on that web gateway thingy? and/or whether it might be something normal i2p users could use?</p>
|
|
<p>15:41 < jrandom> heh, cheers</p>
|
|
<p>15:41 < nicktastic> drunken duck</p>
|
|
<p>15:41 < cervantes> pond water?</p>
|
|
<p>15:41 < jrandom> fvw: get coding :)</p>
|
|
<p>15:42 < deer> <duck> nope, the dev did freeze up. will have to find someone else</p>
|
|
<p>15:42 < jrandom> ok, sorry to hear that</p>
|
|
<p>15:42 < deer> <baffled> We told you not to keep putting them in the closet to protect them.</p>
|
|
<p>15:43 < deer> <duck> my initial specs: http://duck.i2p/files/anonyproxy.txt</p>
|
|
<p>15:44 < deer> <baffled> Is getting mail in/out of i2p as easy as some type of interface web/tunnel to one of these mixmaster thingies?</p>
|
|
<p>15:44 < jrandom> perhaps we can work on a revamp of the spec for that and see if it could serve the needs of normal eepsites (with i2p general funds pitching in)</p>
|
|
<p>15:44 < jrandom> oh ok cool duck, i'll check that out</p>
|
|
<p>15:44 < jrandom> baffled: out of i2p? yes. in to i2p? probably more work</p>
|
|
<p>15:44 < fvw> baffled: Why do you want to add mixmaster? Everything mixmaster offers we already have.</p>
|
|
<p>15:45 < jrandom> fvw: mixmaster has a network of outproxies, plus nontrivial delays</p>
|
|
<p>15:45 < jrandom> ah ok duck, spec glanced over. we may be able to figure something</p>
|
|
<p>15:45 < deer> <baffled> I don't, jrandom suggested setting up a web interface to it not me.</p>
|
|
<p>15:46 < jrandom> (though it seems to have some heavyweight requirements, so maybe not. unsure, we can see)</p>
|
|
<p>15:46 < deer> <duck> its very easy; expectation was 1.5h study of ingredients and then 3-4h patching</p>
|
|
<p>15:46 < fvw> outproxies would be useful yes. As for nontrivial delays, someone who's not already using i2p isn't going to use i2p just for mail when there's mixmaster, whereas someone already using i2p is going to be compromised elsewhere by our lack of delays (if this is possible) anyway</p>
|
|
<p>15:46 < jrandom> right right, plus ship perl, privoxy, and apache duck ;)</p>
|
|
<p>15:47 < jrandom> perhaps fvw. (though i2p 3.0 blah blah blah)</p>
|
|
<p>15:47 < fvw> hehe, I hesitate to say "good point", but I get what you mean.</p>
|
|
<p>15:48 < nicktastic> FYI, JES (Java Email Server) provides SMTP and POP3 servers under the GPL</p>
|
|
<p>15:49 < jrandom> ok, perhaps there should be some more discussion on the list or on the forum about what one or two client apps we should explore focusing on</p>
|
|
<p>15:49 < jrandom> word nicktastic, there's also a kickass one from apache too</p>
|
|
<p>15:49 < nicktastic> Nice, know what its called?</p>
|
|
<p>15:49 < jrandom> http://james.apache.org/</p>
|
|
<p>15:49 < nicktastic> Thanks</p>
|
|
<p>15:50 < jrandom> (nntp too (drooool))</p>
|
|
<p>15:50 < nicktastic> Wow</p>
|
|
<p>15:50 * nicktastic gets a stiffy</p>
|
|
<p>15:51 * fvw has joined #i2p-porn. Or at least it feels like that.</p>
|
|
<p>15:51 < fvw> Ok, next?</p>
|
|
<p>15:51 < jrandom> ok, we can continue client app discussions and strategizing on the list and in the forums</p>
|
|
<p>15:51 < nicktastic> Yes</p>
|
|
<p>15:52 < jrandom> but for now, moving on to 6) ???</p>
|
|
<p>15:52 < nicktastic> Or during non-meeting hours ;P</p>
|
|
<p>15:52 < jrandom> anyone have anything else they want to bring up?</p>
|
|
<p>15:52 * fvw nods. It is worth some discussion on-list.</p>
|
|
<p>15:52 < deer> <duck> little explanation to the www-inproxy: the idea was to get some ISP(s) to offer such a gateway as a service</p>
|
|
<p>15:52 < fvw> nah, the list is good. Gives everyone a chance to chime in, not just those who happen to be awake.</p>
|
|
<p>15:52 < jrandom> word duck, which is quite cool</p>
|
|
<p>15:52 < deer> <duck> so joe i2p-less-sixpack can access it using his MSIE</p>
|
|
<p>15:52 < deer> <duck> but the host is anonymous</p>
|
|
<p>15:52 < deer> <soros> http://it.slashdot.org/article.pl?sid=04/09/14/2226226&threshold=0&tid=172&tid=128&tid=201&tid=218 (ugly new exploit for windows xp)</p>
|
|
<p>15:52 < jrandom> i2pless! heretic! burn them!</p>
|
|
<p>15:53 < deer> <duck> ISP takes part of the risk, hence the whitelist requirement</p>
|
|
<p>15:53 < deer> <duck> and ofcourse payment for domain etz</p>
|
|
<p>15:53 < fvw> hehe. Then suddenly we plaster child porn all over the famous eepsites and watch half the people get arrested and the other half install i2p.</p>
|
|
<p>15:53 < jrandom> heh</p>
|
|
<p>15:53 < deer> * duck calls the AIVD</p>
|
|
<p>15:54 < fvw> duck is dutch? *ponders*</p>
|
|
<p>15:55 < deer> <duck> I think many clientapps are not really killer</p>
|
|
<p>15:55 < jrandom> ok, anyone else have anything they want to bring up?</p>
|
|
<p>15:55 < jrandom> agreed duck, but we need *something* </p>
|
|
<p>15:55 < deer> <duck> some self-made smtp tunnel thing is not going to be a big thing</p>
|
|
<p>15:56 < deer> <duck> myi2p with IOU accounting</p>
|
|
<p>15:56 < deer> <duck> fvw: Bedankt foor die bloumen</p>
|
|
<p>15:57 < deer> <soros> "After complaints to NIC.CX (the regulation authority of .cx domains) by an office worker named Rhonda Clarke of Christmas Island, the site goatse.cx was taken down Friday, January 16, 2004. (Goat.cx and Hick.org/Goat remain active.) A petition has even been launched to bring goatse.cx back. " </p>
|
|
<p>15:57 < deer> <soros> i've lost faith in humanity</p>
|
|
<p>15:57 < deer> <duck> small thing about the site: I2P was added to the <title> of each page for google purposes</p>
|
|
<p>15:57 < deer> <soros> sorry wrong window</p>
|
|
<p>15:57 < jrandom> ah ok duck</p>
|
|
<p>15:57 < deer> <duck> but I dont keep up with the latest google-dance, so it might be worthless.</p>
|
|
<p>15:57 < jrandom> perhaps if there were a way to explicitly not include it?</p>
|
|
<p>15:58 < jrandom> (e.g. so we can say "Welcome to I2P.net" instead of "I2P - Welcome to I2P.net", etc)</p>
|
|
<p>15:58 < deer> <duck> that is ofcourse possible</p>
|
|
<p>15:58 < deer> * duck looks on the fun-o-meter</p>
|
|
<p>15:58 < jrandom> we can always just add title = "I2P - How does it work" to menu.ini</p>
|
|
<p>15:58 < deer> <duck> nope, not today</p>
|
|
<p>15:58 < deer> <thetower> Oh oh, isn't there some way we can get google to crawl i2p? Like some sort of reverse proxy or something?</p>
|
|
<p>15:58 < jrandom> yeah, not worth it</p>
|
|
<p>15:58 < jrandom> thetower with duck's thingamabob, probably.</p>
|
|
<p>15:59 < deer> <duck> yeah</p>
|
|
<p>15:59 < fvw> but as mentioned, you probably don't want to be the one running it.</p>
|
|
<p>15:59 < deer> <thetower> Seems like if eepsites were coming up in google searches that would be good advertising right there.</p>
|
|
<p>16:00 < deer> <duck> fvw: I have contacted an isp who is interested</p>
|
|
<p>16:00 < deer> <duck> but he is not going to build it</p>
|
|
<p>16:00 < jrandom> thetower: perhaps if an ht://dig were hooked up to files.i2p, and if files.i2p exposed the database as big file with html links, that could be mirrored..?</p>
|
|
<p>16:00 < fvw> duck: really? How large and in which country?</p>
|
|
<p>16:00 < cervantes> how about a cache instead of a proxy</p>
|
|
<p>16:00 < cervantes> ah</p>
|
|
<p>16:00 < deer> <duck> 20cm</p>
|
|
<p>16:00 < fvw> If I were an ISP and not afraid of the legal problems, I'd still not be interested until I2P was a lot bigger.</p>
|
|
<p>16:01 < jrandom> a cache would be interesting too, a swarm of squids, etc</p>
|
|
<p>16:01 < deer> <duck> skynet</p>
|
|
<p>16:01 < fvw> that's pretty big. Did you give them a phone book to sit on?</p>
|
|
<p>16:01 < nicktastic> hehe</p>
|
|
<p>16:01 < deer> <duck> fvw: they'll likely scan the site before adding it</p>
|
|
<p>16:01 < deer> <duck> so you'll have to find another place for your nasty stuff</p>
|
|
<p>16:01 < fvw> Once or every update?</p>
|
|
<p>16:02 < fvw> the latter seems a lot of work for very little content.</p>
|
|
<p>16:02 < deer> <duck> every 2nd sunday of a month with an x in it</p>
|
|
<p>16:02 < deer> <duck> geeh</p>
|
|
<p>16:03 < jrandom> ok, we've gone past the two hour mark, does anyone else have something to bring up, or should we continue further discussions on the list / in the forum / etc?</p>
|
|
<p>16:03 < fvw> I just find it highly unlikely that a serious ISP will commit resources to I2P at this point.</p>
|
|
<p>16:03 * cervantes covers his head with a saucepan</p>
|
|
<p>16:03 < deer> <duck> fvw: your emotions are noted.</p>
|
|
<p>16:03 * fvw nods at jrandom. I need a drink. Keep up the good work.</p>
|
|
<p>16:03 < deer> <duck> when will we go for the 24h meeting?</p>
|
|
<p>16:04 < jrandom> perhaps next week duck </p>
|
|
<p>16:04 * jrandom winds up</p>
|
|
<p>16:04 < deer> <duck> oh boy!</p>
|
|
<p>16:04 < fvw> duck: my emotions? You haven't even begun to see my emotions. Let me get a few drinks though.. *grin*</p>
|
|
<p>16:04 * jrandom *baf*s cervantes on the head, closing the meeting</p>
|
|
</div>
|
|
{% endblock %} |