I2P dev meeting, DATE

Quick recap

  • Present:

Brownspider, co, jrand0m, mariesofie, mihi, shardy, w0rmus,

Registro completo do IRC

[22:57] <jrand0m> ok, buenos noches srs y srtas 
[22:57] <jrand0m> agenda: 
[22:57] <jrand0m> 0) welcome 
[22:57] <jrand0m> 1) cvs 
[22:57] <jrand0m> 4) naming service 
[22:58] <co> You forgot 5) questions.
[22:58] <jrand0m> 3) dev status 
[22:58] <jrand0m> 2) spec questions? 
[22:58] <jrand0m> 5) other questions? 
[22:58] <jrand0m> oh, shit, I forgot to reorder those.  ok.  they're numbered incorrectly :)  0 == 0, 1 == 1, 4 ==2, 3 == 3, 2 == 4, 5 == 5 
[22:59] <jrand0m> lets see if I can keep that straight as we go... 
[22:59] <jrand0m> ok, 0) welcome 
[22:59] <shardy> yay for permutation groups!
[22:59] <jrand0m> welcome to meeting 57 
[22:59] <jrand0m> yeah, they're all just symbols anyway 
[22:59] <w0rmus> sweet whatup ;0
[23:00] <w0rmus> I'll help comprise peanut gallery
[23:00] <jrand0m> 1) cvs is still down, after 10+ days.  we're finding a new host. 
[23:00] <jrand0m> sf.net sucks, and I have no reason to believe gnu's nongnu is better. 
[23:00] <co> jrand0m: Why not make that host have the alias "cvs.invisiblenet.net"?
[23:00] <jrand0m> nop is leading the charge on finding the new host. 
[23:01] <jrand0m> sure co, once we get the server 
[23:01] <shardy> what do you need for a host?
[23:01] <jrand0m> shardy> reliable net connection, ssh/cvs access.  and some disk space 
[23:01] <shardy> do you have something lined up?
[23:01] <shardy> because if not. I may be able to assist.
[23:02] <jrand0m> awesome!  I don't know what nop has lined up, but I'll have him check in with you (unless he's here now?) 
[23:02] * w0rmus taps nop
[23:03] <shardy> I've got 1.1 business sdsl. I'd need to find a machine. but as long as you're not using uber amounts of bandwidth I could probably host the server.
[23:03] <shardy> how much disk space would you need?
[23:03] <jrand0m> the repository currently comprises ~ 6Mb.  so probably 50M would handle growth for a good while 
[23:04] <shardy> oh. pfft. that's nothing.
[23:04] <shardy> and the machine wouldn't need to be super fast?
[23:04] <shardy> you wouldn't be doing big compile jobs on it?
[23:04] <jrand0m> naw, a 286 would probably do it. 
[23:04] <jrand0m> no, strictly cvs checking / checkout  
[23:04] <jrand0m> (well, and diff, and log, etc ;) 
[23:05] <jrand0m> "we're java, we don't need compile farms" </fark> 
[23:05] *** Signoff: cohesion (class)
[23:05] <w0rmus> do people concerned with anonymity access CVS using something like JAP?  I've never used CVS
[23:05] <jrand0m> w0rmus> I use cvs through a private series of ssh proxies 
[23:05] <co> jrand0m: Keep in mind that a C or C++ implementation may be likely in the future.
[23:06] <mihi> jap does not allow cvs access AFAIK :(
[23:06] <w0rmus> ssh tunnelling
[23:06] <shardy> let me see what I can do. someone said they were going to give me another drive... if I can get a disk I have a machine I can toss up.
[23:06] <jrand0m> oh, of course co.  I just don't expect we can require a cvs repository to necessarily be a compile farm as well. 
[23:07] <jrand0m> awesome shardy.  there's anything we can do, just say the word. 
[23:07] <co> jrand0m: You are right. They should be separate.
[23:07] <shardy> will do. let me scrounge for a drive, I should be able to get one, and if I do I'd be happy to host the cvs for everyone.
[23:07] <jrand0m> shardy++ 
[23:07] <w0rmus> yay ;0
[23:07] <jrand0m> ok, 4) naming service 
[23:08] <jrand0m> co, how goes?   
[23:08] <co> I am still writing, but would like to say a few words about it.
[23:08] <co> First, to address a question from thecrypto during the last meeting, the NS does not provide notification of someone's being online.
[23:09] <co> It merely says that a person can be contacted via certain methods, such as AIM.
[23:09] <co> Second, the client side.
[23:09] <co> There will be an API which programs can use to make queries to naming servers.
[23:10] <co> The underlying mechanism will read a configuration file of which servers to query, will use the I2P network to obtain the results, and will pass the results back to the caller.
[23:11] <co> The underlying mechanism will also read the destination mechanism for the local router to contact from a file.
[23:11] <jrand0m> the destination mechanism? 
[23:11] <co> I'm sorry, the destination address.
[23:11] <jrand0m> ah 'k 
[23:12] <co> That is all at this time.
[23:12] <jrand0m> cool 
[23:12] <w0rmus> I agree
[23:12] <w0rmus> ;)
[23:12] <jrand0m> any ballpark ideas on timelines for various milestones?   
[23:13] <jrand0m> obviously nothing anyone could hold you to, of course, just wondering 
[23:13] <co> Let's say end of the week for finishing the specification and publishing it and the API.
[23:14] * mariesofie arrives late
[23:14] <jrand0m> ah nice co 
[23:14] <co> Then, I will start implementing it. I am not certain how long that will take, though.
[23:14] <jrand0m> understandable 
[23:15] <jrand0m> anyone else have any naming service questions/thoughts? 
[23:15] <jrand0m> ok, 3) dev status 
[23:16] <jrand0m> dev goes well. 
[23:16] <jrand0m> the java side is up to spec and implements all I2CP and I2NP messages and structures 
[23:17] <jrand0m> the java architecture itself is functional and I'm going to continue stubbing out the various subsystems 
[23:17] <co> Have you tested it?
[23:17] <jrand0m> the messages &amp; structures?  yes via the TestData harness in net.invisiblenet.i2p.data.test 
[23:17] <co> I mean connecting two different computers with I2P.
[23:18] <jrand0m> oh, no, thats requires the full implementation of the communication subsystem 
[23:18] <co> I see.
[23:18] <jrand0m> first I'm building the various subsystems to operate in test mode, then implementing the various subsystems so they can operate in isolation 
[23:19] <jrand0m> we're probably 2 weeks off from a client sending a message to a client on a different router 
[23:19] * mariesofie cheers
[23:20] <jrand0m> there's still a lot of work to be done after that before the alpha, but thats progress  
[23:21] <jrand0m> the datastrucutres and i2np specs need about a dozen small modifications that I've been accumulating during the implementation to address things overlooked.  e.g. "datastructures p11, TunnelSigningPublic/PrivateKey should contain SIGNING Public/Private keys" and "i2np p15, TunnelCreateStatus - add hash of the replying RouterIdentity" 
[23:21] <shardy> man. i really need to read up on the specs.
[23:22] <jrand0m> well, they'll soon be hosted on your machine so it'll be easy :) 
[23:22] <w0rmus> haha
[23:22] <w0rmus> I have not finished specs either
[23:23] <mariesofie> i printed the specs, read them so many times it got worn out and i had to print out another copy
[23:23] <jrand0m> from discussions I've had with various people, I've found the specs aren't that great at conveying how the thing actually works.  they cover the fuzzy stuff and the nitty gritty, but not why the nitty gritty meets the why 
[23:23] <w0rmus> heh
[23:23] <jrand0m> rofl mariesofie 
[23:23] <jrand0m> ok, thats it for 3) dev status 
[23:24] <jrand0m> now for 2) spec questions 
[23:24] <w0rmus> I suppose I'll read them instead of stupid calculus
[23:24] <co> I am thinking a little bit into the future.
[23:24] <co> The python and C or C++ implementations will need to have message data readable by the Java implementation.
[23:24] *** Signoff: mihi (EOF From client)
[23:25] <jrand0m> correct co 
[23:25] <co> How will you accomplish that?
[23:25] <jrand0m> the datastructures spec specifically defines the byte layouts 
[23:25] <jrand0m> and everything is big endian and all numbers are unsigned 
[23:25] <mariesofie> what level of technical knowledge are you targetting the specs for? anyone with decent comp. knowledge? college level CS engineer students?
[23:25] <co> Oh, all right.
[23:25] *** mihi_ (~none@anon.iip) has joined channel #iip-dev
[23:26] <mariesofie> i.e. who is the target audience?
[23:26] *** mihi_ is now known as mihi
[23:26] <jrand0m> mariesofie> well, those specs were really targeted haphazardly.  i2p_philosophy was the "ok, wtf is this all about", but the rest of the specs were targeted towards people interested in actually implementing the system 
[23:26] <jrand0m> we really really need some docs that go in the middle 
[23:27] <mariesofie> i see
[23:27] <mariesofie> I found the API docs are very easy to understand and useful, but ironically I still get mixed up when reading the I2NP specs trying to understand the core architecture
[23:28] <mariesofie> maybe thats reflective on me more than the documentation :)
[23:28] <jrand0m> heh geek :) 
[23:29] <jrand0m> ok, any other questions on the specs?  lets move on to 5) other questions  
[23:29] <jrand0m> anyone have any other questions?  this is our last bullet point on the meeting agenda 
[23:30] <w0rmus> I wonder where thecrypto is with achat
[23:30] <jrand0m> ah, thecrypto is offline for the next three weeks or so 
[23:30] <mihi> what happens with peer review?
[23:30] <w0rmus> or atalk
[23:30] <w0rmus> wow
[23:30] <mihi> is anyone reviewing it?
[23:30] <mariesofie> i thought thecrypto has 2hrs per day
[23:31] <w0rmus> and I can't even see the codes he has :(
[23:31] <jrand0m> mihi> reviews have been sent to various people for review and as feedback comes it will be addressed. 
[23:31] <jrand0m> w0rmus> do you have any questions on ATalk? 
[23:32] <shardy> I will be reviewing it as soon as I have time :)
[23:32] <mihi> i meant, did any feedback come till nowß
[23:32] <w0rmus> I suppose not
[23:32] <mihi> s/nowß/now?/
[23:32] <jrand0m> mihi> largely in the form of discussions and clarifications 
[23:32] <jrand0m> awesome shardy :) 
[23:33] <Brownspider> hapy birthday google
[23:33] <jrand0m> mariesofie> right, but thats not even enough time for him to d/l the java lang docs to continue dev :/ 
[23:33] <w0rmus> ahaha wtf
[23:35] <jrand0m> ok, any other questions / thoughts? 
[23:35] <w0rmus> I guess I should mention that I've never coded outside of school
[23:35] <w0rmus> but I gotta start somewhere anyways
[23:35] <jrand0m> now's a good time to start :) 
[23:35] <jrand0m> word 
[23:35] <w0rmus> ;0
[23:35] <mariesofie> i have questions on the api, but not yet, in a day or two when i can try it out some more
[23:35] <w0rmus> I've taken 2 years of java or so
[23:36] <jrand0m> ok cool mariesofie, just bounce a msg to the list or bounce me a message here whenever 
[23:37] <co> mariesofie: Have you read the discussion on the iip-dev mailing list?
[23:37] <w0rmus> where is it archived?
[23:37] <jrand0m> http://news.gmane.org/thread.php?group=gmane.comp.security.invisiblenet.iip.devel 
[23:37] <jrand0m> (fairly low traffic atm) 
[23:38] <Brownspider> jrand0m wants you code something that can not logicly exist, to tear the world usunder, to end the reign of god.
[23:38] <shardy> my services are still offered if you need any crypto cores or such written or debugged. 
[23:39] *** M123456789 (~no@anon.iip) has joined channel #iip-dev
[23:39] <co> Brownspider: huh?
[23:39] <Brownspider> co, it was on his freesite, nevermind
[23:40] <jrand0m> great shardy, I have a feeling we're going to need some once the routers get up and running, and especially when we get the C/Python APIs back up to spec 
[23:40] <mariesofie> co> i've only read back to about issue #52 or so
[23:42] <jrand0m> ok.  any last words (as I eye the *baf*er...) 
[23:43] *** mihi_backup_ (~none@anon.iip) has joined channel #iip-dev
[23:43] *** Signoff: mihi_backup (EOF From client)
[23:43] *** mihi_backup_ is now known as mihi_backup
[23:43] <jrand0m> ok great.  next week, same bat time, same bat place. 
[23:44] * jrand0m *baf*s the meeting to an end