I2P dev meeting, March 2, 2004 @ 13:00 PST

Quick recap

  • Present:

duck, jrandom, kaji, madman, nick, ughabugha, wilde,

Registo Completo do IRC

13:07 < jrandom> 0) hi
13:07 < jrandom> 1) Dev status
13:07 < jrandom> 2) Cascades
13:07 < duck> I'll stop
13:07 < jrandom> 3) Roadmap
13:07 < jrandom> 4) Website
13:07 < jrandom> 5) ???
13:07 < jrandom> 0) hi
13:07  * jrandom waves to the first over-i2p i2p dev meeting :)
13:07 < ughabugha> nick is Janonymous.
13:08 < ughabugha> Ok.
13:08 < duck> hi
13:08 < jrandom> weekly status notes are posted to the mailing list (online at http://i2p.net/pipermail/i2p/2004-March/000155.html)
13:08 < ughabugha> hi.
13:08 < jrandom> as usual, we'll be following that as a guide
13:08 < nick> hello
13:09 < jrandom> jumping into 1) Dev status
13:09  * jrandom repeats mantra of "Progress is being made"
13:10 < jrandom> 0.2.5 has some Good Stuff, and we're finding long hidden bugs
13:10 < jrandom> latest one is db related, but thats not fixed up yet, so no need to track CVS HEAD
13:11 < jrandom> echo tests show pretty good results, but there's still issues to be worked out wrt irc and snark
13:12 < jrandom> how has eepsite retrieval been for people?
13:12 < jrandom> (and/or squid?)
13:12 < nick> here and there
13:12 < duck> generally okay with janonymous or ugha
13:13 < duck> sometimes janonymous goes down etc
13:13 < ughabugha> Yeah.
13:13 < duck> <mihi_backup> jrandom: you *really* know how to make me angry...
13:13 < jrandom> d'oh
13:13 < nick> I've been putting new content
13:13 < jrandom> whats up mihi?
13:13 < jrandom> nice nick
13:14 < jrandom> er janonymous
13:15 < nick> eh?
13:15 < nick> :)
13:15 < ughabugha> IIP <ughabugha> I started a two-way relay now.
13:15 < jrandom> ah cool
13:15 -!- nick [~Janonym@localhost] has quit [Client closed connection]
13:15 < jrandom> whats up mihi?
13:15 < ughabugha> IIP <ughabugha> Automatic.
13:16 < ughabugha> IIP <ughabugha> :)
13:16 < ughabugha> :)
13:16 < ughabugha> IIP <ughabugha> Oops, a bug.
13:16 < ughabugha> Oops, a bug.
13:16 < jrandom> heh is he still there or is there anything else wrt dev status?
13:17 < duck> lets go on
13:17 < jrandom> 'k
13:17 < jrandom> jumping to 2) Cascades
13:17 < ughabugha> IIP <ughabugha> * mihi_backup is now known as mihi_away
13:17 < ughabugha> * mihi_backup is now known as mihi_away
13:17 < ughabugha> IIP <ughabugha> Argh, relaying for my own text doesn't work.
13:17 < ughabugha> Argh, relaying for my own text doesn't work.
13:17 < jrandom> ;)
13:18 < ughabugha> IIP <ughabugha> No, wait.
13:18 < ughabugha> No, wait.
13:18 < madman> then don't speak :)
13:18 < jrandom> mix cascades are one of the two big styles of low latency mix nets, and while we don't use them in i2p, if some people think they're useful, they can tweak their router to get the same effect
13:19 < ughabugha> IIP <ughabugha> Test
13:19 < ughabugha> Test
13:19 < ughabugha> IIP <jrandom> mix cascades are one of the two big styles of low latency mix nets, and while we don't use them in i2p, if some people think they're useful, they can tweak their router to get the same effect
13:19  * jrandom senses an incoming recursive echo...
13:19 < ughabugha> IIP <ughabugha> Damn!
13:19 < ughabugha> Damn!
13:20 < jrandom> anyway, I think its always good to ask fundamental design questions, to poke at i2p and see why we do things the way we do
13:21 < jrandom> i've got the feeling we'll be hearing more about cascade-like systems in the future, so hopefully the description in the email will help explain i2p's relation to cascades
13:21 < ughabugha> Ok, it should work _now_.
13:21 < ughabugha> IIP <ughabugha> Say something.
13:21 < ughabugha> IIP <Janonymous> I move to not go after a mix net implementation within the current roadmap, and leave that for later
13:21 < ughabugha> Yeah, it works now.
13:21 -!- wilde [~anon@localhost] has joined #i2p
13:22 < jrandom> janymous> well, i2p /is/ a mixnet, just not a mix cascade
13:22 < madman> i'm going offline
13:22 < jrandom> 'k, ttyl madman
13:22 < jrandom> heya wilde
13:22 < madman> so bye
13:22 < ughabugha> IIP <wilde> finally
13:22 < ughabugha> IIP <wilde> hello meeting
13:22 < jrandom> but I concur, I don't think mix cascade functionality needs to be on the roadmap
13:23 < ughabugha> IIP <madman2003> bye everyone
13:24 < jrandom> ok, anything else on cascades, or should we move to 3) Roadmap?
13:26 < ughabugha> Hi, wilde@IIP
13:26 < ughabugha> Bye, madman@IIP
13:26 -!- madman [~a@localhost] has quit [( www.nnscript.de :: NoNameScript 3.8 :: www.XLhost.de )]
13:26 < ughabugha> IIP <Janonymous> Just seems like it could be implemented later, like a DHT might be. High Wilde
13:26 < ughabugha> IIP <Janonymous> bye Madman
--- Log closed Tue Mar 02 13:27:07 2004
--- Log opened Tue Mar 02 13:27:52 2004
13:27 -!- jrandom [~jrandom@localhost] has joined #i2p
13:27 -!- Irssi: #i2p: Total of 3 nicks [0 ops, 0 halfops, 0 voices, 3 normal]
13:27 -!- wilde [~anon@localhost] has joined #i2p
13:27 < jrandom> back
13:28 < ughabugha> Uhoh, jrandom quitted.
13:28 < ughabugha> IIP <Janonymous> bah
13:28 < jrandom> ok last I saw was 13:26:08 < ughabugha> IIP <Janonymous> bye Madman
13:28 -!- Irssi: Join to #i2p was synced in 36 secs
13:28 < jrandom> (irssi missed a ping so it dropped it)
13:29 < ughabugha> You missed IIP <Janonymous> Roadmap
13:29 < ughabugha> IIP <Janonymous> ah.. We all agreed to go onto the roadmap ;)
13:29 < jrandom> w3rd
13:29 < jrandom> ok, the roadmap change is likely why mihi hates me now
13:30 < ughabugha> Uh, then you missed alot more than that.
13:30 < ughabugha> Just a second.
13:30 < ughabugha> [23:23 39] <ughabugha> IIP <wilde> oh we have to camps now, the iip gang and the I2P hood
13:30 < ughabugha> [23:23 47] <ughabugha> IIP <wilde> two
13:30 < ughabugha> [23:23 50] <ughabugha> Hehe.
13:30 < ughabugha> [23:23 54] <ughabugha> IIP <Janonymous> :) got a relay going
13:30 < ughabugha> [23:24 22] <ughabugha> This is a temporary script I hacked together quickly. For future meetings we should think of something better.
13:30 < ughabugha> [23:24 44] <ughabugha> IIP <Janonymous> It works
13:30 < ughabugha> [23:25 18] <ughabugha> Ok, concentrate on I2P now, not the relay.
13:30 < ughabugha> [23:25 39] <ughabugha> IIP <Janonymous> So, mix cascades could feasably implemented quite well over i2p
13:30 < ughabugha> Sorry for the flood.
13:30 < duck> this is chaos
13:31 < ughabugha> IIP <Janonymous> but i think we can emphasize /over/
13:31 < ughabugha> Ok, 3) Roadmap
13:31 < ughabugha> duck: Pretty much.
13:31 < jrandom> this aint nothing compared to meeting 67 ;)
13:31 < jrandom> ok, on to 3
13:31 < jrandom> anyone have any thoughts wrt the roadmap?
13:32 < jrandom> the changes / views / concerns?
13:32 < ughabugha> IIP <Janonymous> Multi-tunneling
13:32 < jrandom> janonymous> we've already got that
13:32 < jrandom> (as of 0.2.5)
13:32 < jrandom> if I understand you correctly
13:33 < ughabugha> IIP <Janonymous> but, as in, sending one file over two tunnels to accelerate transmission?
13:33 < ughabugha> IIP <Janonymous> at the same time
13:34 < jrandom> i2p doesn't deal with files, but yes, each individual i2p message can now go down multiple tunnels
13:34 < ughabugha> IIP <Janonymous> I think that would be a great addition for the 2.0 area
13:34 < jrandom> e.g. first 32kb sent down tunnel X, next 32kb sent down tunnel Y
13:35 < ughabugha> IIP <Janonymous> right.. thats what I'm getting at.. that seems very necessary to me
13:35 < jrandom> but i2psnark with i2cp support would be able to maximize things
13:35 < jrandom> janonymous> we do that now
13:36 < ughabugha> IIP <Janonymous> oh. I2PSnark will utilize multiple tunnels for point to point communication?
13:36 < jrandom> all messages can go down multiple tunnels.
13:36 < duck> theoretically
13:37 < jrandom> not just theoretically - if a message takes > 15s, its sent down the other available lease
13:37 < jrandom> and if it takes > 30s, the leaseSet is dropped and refetched, with subsequent messages going down found leases
13:37 < jrandom> BUT
13:38 < jrandom> i2ptunnel (any anything else that uses i2p's mode=guaranteed) waits until each message is delivered before sending the next one
13:38 < jrandom> native i2cp apps don't need to do that
13:38 < jrandom> (nor will any apps that use the socket library, once the socket library is both implemented and supports SACK)
13:38 < ughabugha> IIP <Janonymous> ok.. I just think that will be a great method for these tunnels in the future.. for speed and anomymity
13:38 < ughabugha> IIP <Janonymous> and keeping strain off individual tunnels
13:39 < jrandom> agreed
13:39 < jrandom> ok, anything else on the roadmap?
13:40 < jrandom> (anyone going to bitch me out for dropping the socket lib?  mihi?  :)
13:41 < ughabugha> IIP <Janonymous> I used an analogy earlier today to describe multi-tunneling to someone.. and I said it was like adding lanes to a road
13:42 < jrandom> pretty much
13:42 < jrandom> (though one's on-ramp is always the same number of lanes ;)
13:42 < duck> if mihi is angry he can do the socket api cant he?
13:43 < jrandom> sure, and/or anyone else.  the socket lib is Good
13:43 < jrandom> (but hard, and imho not functionally essential to the operation / security of the network)
13:44 < jrandom> i just wish i had the time to do it and keep moving on the core i2p code
13:44 < jrandom> but, c'est la vie
13:45 < ughabugha> IIP <Janonymous> looks like mihis not here
13:45 < ughabugha> Ok, i'm back now.
13:45 < jrandom> coo'
13:45 < jrandom> ok, moving on to 4) website
13:46 < ughabugha> IIP <Janonymous> big on ramp == cable / small on ramp == dialup ??
13:46 < ughabugha> Hmm
13:46 < ughabugha> What was the socket library going to do?
13:46 -!- nick [~Janonym@localhost] has joined #i2p
13:46 < jrandom> right nanonymous
13:46 < jrandom> er, janonymous
13:46 < jrandom> (no matter what, i2p cant make your local net connection faster)
13:46 < jrandom> ughabugha: http://wiki.invisiblenet.net/iip-wiki?I2PSocketLibrary
13:47 < ughabugha> I mean a socket library for Java? Don't you already have one?
13:47 < jrandom> the socket lib factors out the TCP-esque code out of i2p, letting i2p specialize in IP-like messages
13:47 < nick> yup
13:47 -!- nick [~Janonym@localhost] has quit [Client closed connection]
13:48 < jrandom> ah, right, yes, but this would let applications stream data over i2p much more efficiently (if/when the socket library supports selective ACK, rather than requiring an ACK after each message like it does now)
13:49 < jrandom> i'm not comfortable with implementing SACK within the router, since it can safely go outside of it (into the socket lib)
13:49 < ughabugha> But why drop it? Does it really take that much work?
13:49 < jrandom> yes, to get right
13:49 < jrandom> there's some code thats part way implemented, but i dont have time to maintain and test it
13:49 < ughabugha> Ok. You're the man.
13:50 < jrandom> well, $devWhoImplements it is the man ;)
13:50 < jrandom> anyway, moving on to 4) website
13:50 < ughabugha> :)
13:51 < ughabugha> Any volunteers?
13:51 < ughabugha> IIP <Janonymous> research on implementations of anonymous p2p
13:51  * jrandom echoes ughabugha's question :)
13:51 < jrandom> hmm janonymous?
13:51 < ughabugha> Janonymous: This will be covered under 5) ???
13:51 < jrandom> :)
13:52 < ughabugha> IIP <Janonymous> wll its content that goes on the site
13:52 < jrandom> ah, yeah, I agree
13:52 < jrandom> (see item 7 on http://i2p.net/pipermail/i2p/2004-February/000133.html)
13:53 < jrandom> and item 8
13:53 < jrandom> or is that not what you mean?
13:53 < jrandom> I'll probably post up the truckload of papers i dug through last summer when researching and designing i2p
13:53 < jrandom> (or at least pointers to their citeseer entries)
13:54 < ughabugha> IIP <Janonymous> Ok. Were we going to discuss the next CMS for I2P?
13:54 < ughabugha> jrandom allready chose the CMS.
13:55 < jrandom> yes/no - rather than researching the pros and cons of CMSes for another month or two, we'll just go with drupal for now
13:55 < ughabugha> IIP <ughabugha> /topic #i2p
13:55 < ughabugha> IIP <Janonymous> Ok.. well, as far as what to put there.. We need a presentation
13:55 < jrandom> if we need to migrate to another one, wilde assures me its simple enough to export content
13:55 < jrandom> a presentation?
13:56 < ughabugha> IIP <Janonymous> with lots of illustrations and a step by step introduction to I2P
13:56 < jrandom> we do need the graphic design implemented
13:56 < jrandom> ah right
13:56 < jrandom> a user's intro
13:56 < jrandom> the wiki intro is generally a technie intro
13:56 < ughabugha> IIP <Janonymous> Almost like a multimedia presentation
13:56 < ughabugha> IIP <jrand0m> w0ah
13:56 < ughabugha> IIP <Janonymous> right
13:57 < ughabugha> IIP <ughabugha> Yay!
13:57 < jrandom> ok, i think we can get that together, but we'll probably want to wait on producing that content until we have both a Real installer, and a GUI control system
13:57 < ughabugha> IIP <Janonymous> In it, there should be more pictures, than words. :)
13:57 < jrandom> right
13:58 < ughabugha> IIP <Janonymous> righto
13:58 < jrandom> but we don't have a real installer yet, and (as much as i2pmgr and i2pmole are great) i think there's still work to be done on a control panel
13:58 < ughabugha> This is not the top priority right now.
13:58 < ughabugha> IIP <Janonymous> certainly not. but for those of us not programming...
13:59 < jrandom> right.  so we need volunteers to work on 1) designing what content needs to be on the i2p website 2) designing the graphics / css / layout for the i2p website 3) people to work on creating content for the i2p website
14:00 < ughabugha> IIP <Janonymous> I've got a month to spare. Think it would be time well spent.
14:00 < jrandom> w00t :)
14:00  * jrandom marks Janonymous down as a volunteer... for content design &amp; content creation?
14:00 < wilde> back
14:00 < ughabugha> Well, I could do HTML, CSS and the technical stuff, but I'm not much of a writer, nor a designer.
14:01 < jrandom> r0x0r
14:01 < ughabugha> IIP <Janonymous> right.. I'm not so good at the designing part yet
14:01 < ughabugha> IIP <Janonymous> I could try to write, but not without some good critique.. I've never writen editorial type things
14:01 < ughabugha> I can also use Photoshop and other tools as long as I'm given specific instructions on what to do. ;)
14:01 < wilde> I'll take care of drupal and features
14:02 < ughabugha> IIP <Janonymous> I've got some good ideas for you ughabugha
14:02 < jrandom> right, by content design i don't mean layout, but more "ok, we need a user intro page, a tech intro page, a faq" etc
14:02 < ughabugha> Heh, ok. :)
14:02 < jrandom> r0x0r0r
14:02  * jrandom marks down wilde and ughabugha as volunteers :)
14:02 < ughabugha> Looking forward to them.
14:02 < ughabugha> IIP <Janonymous> there's a shaby picture I made that can be found linked to the new I2POverview doc on I2p
14:03 < jrandom> word, yeah janonymous, that pic is pretty good, some minor tech issues with it, but quite useful
14:03 < ughabugha> IIP <Janonymous> It would look alot nicer in photoshop I'm sure
14:04 < ughabugha> Heh.
14:04 < ughabugha> Janonymous: Let's discuss that privately tomorrow.
14:04 < ughabugha> IIP <Janonymous> eh, actually I may have accidentally deleted it :/
14:04 < jrandom> ok, anything else for the website, or can we move on to 5) ???
14:04 < ughabugha> IIP <Janonymous> ok
14:07 < wilde> ok one thing:
14:07 < ughabugha> Anything else on website?
14:07 < wilde> what is the first feeling you should get on the site?
14:07 < wilde> keywords please
14:07 < jrandom> wilde> see http://i2p.net/pipermail/i2p/2004-February/000133.html (the "rom a branding perspective" paragraph :)
14:08 < jrandom> i do like the anonymous bit by bit thing
14:08 < ughabugha> I suppose not. Should we go on to 5) ??? ?
14:08 < ughabugha> IIP <Janonymous> should we make a more detailed user roadmap?
14:08 < ughabugha> IIP <Janonymous> one describing the long term goals in more detail
14:08 < ughabugha> wilde: Only positive emotions.
14:09 < jrandom> janonymous> agreed, the current roadmap is really just tech notes for tech tasks ;)
14:09 < jrandom> ok, 5) ??
14:09 < jrandom> anything y'all want to bring up?
14:10 < wilde> itoopie isn't really in line with simple and secure
14:10 < wilde> it's more of a cartoon feeling
14:10 < wilde> that's why i asked
14:10 < ughabugha> IIP <Janonymous> yea, might want to build some more anticipation
14:10 < ughabugha> IIP <Janonymous> for the users to get involved
14:10 -!- kaji [~booky5@localhost] has joined #i2p
--- Log closed Tue Mar 02 14:11:08 2004
--- Log opened Tue Mar 02 14:12:12 2004
14:12 -!- jrandom_ [~jrandom@localhost] has joined #i2p
14:12 -!- Irssi: #i2p: Total of 6 nicks [0 ops, 0 halfops, 0 voices, 6 normal]
14:12 < jrandom_> back
14:12 < ughabugha> IIP <Janonymous> like, it should work for them without them even knowing its there
14:12 < jrandom_> wilde> I'm open to suggestions
14:12 < ughabugha> * jrandom_ (~jrandom@localhost) has joined #i2p
14:12 < jrandom_>  <Janonymous> yea, might want to build some more anticipation
14:12 < jrandom_> hmm?
14:14 < ughabugha> You missed these:
14:14 < ughabugha> [00:11 43] <ughabugha> IIP <Janonymous> I'd just like to reemphasize.. I just think all multi-tunneling methods should be transparent and available to all client apps
14:14 < ughabugha> [00:12 07] <ughabugha> IIP <Janonymous> like, it should work for them without them even knowing its there
14:14 < jrandom_> janonymous> already implemented.
14:14 < jrandom_> i2p already transparently balances end to end communication over multiple tunnels
14:15  * jrandom_ kicks jrandom
14:15 -!- jrandom [~jrandom@localhost] has quit [Ping timeout]
14:15 -!- Irssi: Join to #i2p was synced in 231 secs
14:15 < wilde> participation?
14:15 < ughabugha> IIP <Janonymous> even if we profile the fastest most stable tunnels.. we can still use the other slow tunnels for extra throughput if we need it
14:15 < kaji> is iip up?
14:15 -!- kaji [~booky5@localhost] has quit [Client closed connection]
14:15 < ughabugha> IIP <Janonymous> And we may want to distribute the load anyway.. and that should all be transparent to the client apps
14:15 -!- You're now known as jrandom
14:16 -!- protocol [~iip@localhost] has joined #i2p
14:16 < ughabugha> IIP <Janonymous> ok
14:16 < ughabugha> IIP <Janonymous> cool
14:16 < jrandom> :)
14:17 < ughabugha> jrandom: Don't use ACTION, it's not relayed ;)
14:17 < jrandom> hah ok sorry
14:17  * jrandom says something they cant see
14:17 < jrandom> ;)
14:17 < jrandom> ok, anyone else have anything else?
14:18 < jrandom> i think after the current netDb bugs are fixed we may want to try the i2psnark tests again
14:20 -!- kaji [~booky5@localhost] has joined #i2p
14:20 -!- wilde [~anon@localhost] has quit [Ping timeout]
14:20 < ughabugha> But does the 15-second wait really distribute the load?
14:20 < ughabugha> That's not how I see it.
14:20  * protocol says iip is for lusers
14:20 < ughabugha> The way I see it, it should be 100% simultaneous, the node should put packets through the tunnel as fast as the destination can handle them.
14:20 < ughabugha> Through all the tunnels, I mean.
14:20 -!- nick [~Janonym@localhost] has joined #i2p
14:20 -!- kaji [~booky5@localhost] has quit [Client closed connection]
14:21 < jrandom> ughabugha: each message is ideally only sent over one tunnel, but each individual message is balanced over all of them
14:21 < ughabugha> IIP <Janonymous> me and duck ran a test over I2PSnark
14:21 < ughabugha> IIP <Janonymous> earlier today.
14:21 < jrandom> ughabugha: if we sent it over all tunnels always, that'd be a significant amount of wasted traffic
14:21 -!- nick [~Janonym@localhost] has quit [Client closed connection]
14:21 < ughabugha> IIP <Janonymous> Is everyone happy with the "I2P" name?
14:21 -!- kaji [~booky5@localhost] has joined #i2p
14:21 -!- wilde [~anon@localhost] has joined #i2p
14:22 < jrandom> janonymous> i2p is the name.
14:22 < jrandom> you can call it betty, but i2p is the name ;)
14:22 < ughabugha> kaji: Why are you blinking like this?
14:22 < jrandom> ughabugha: kaji likes messing with us
14:22 < ughabugha> jrandom: Yeah, I understand that. I guess it works just as I imagine it.
14:23 < kaji> i dont know
14:23 < kaji> is iip up?
14:23 < ughabugha> IIP <Janonymous> cool
14:23 < ughabugha> IIP <Janonymous> is for me
14:23 < ughabugha> It is for some people, and it's not for others.
14:23 < ughabugha> So I'm running a relay.
14:24 < jrandom> ughabugha++
14:24 < ughabugha> To connect the networks.
14:24 < kaji> cool
14:24 < jrandom> the details of the parallel/serial sending is in net.invisiblenet.i2p.router.message.OutboundClientMessageJob
14:24 < ughabugha> :)
14:24 < jrandom> (for anyone who wants to know more details of how it works)
14:25 < jrandom> ok, anything else people want to bring up?
14:25 < kaji> so how is i2p dev? :) (Mirc sez 'lag=30 seconds')
14:25 < jrandom> kaji> we're making progress ;)
14:26 < jrandom> irssi here has bounced between 80s and 1s lag
14:26 < jrandom> (two disconnects in the last 90 minutes)
14:26 < ughabugha> IIP <Janonymous> is there any more ideas on content for the new site?
14:27 < ughabugha> IIP <kaji> sweet
14:27 < jrandom> beyond http://i2p.net/pipermail/i2p/2004-February/000133.html I think we'll want to go with one of drupal's forum modules
14:27 < ughabugha> No disconnects for me for 80 minutes.
14:27 < jrandom> nice ughabugha 
14:27 < kaji> hmm now the lag is down to a few seconds
14:27 < jrandom> yeah, it varies kaji
14:28 < ughabugha> IIP <Janonymous> oh, me and ugha talked on my chat room over eep today :)
14:28 < jrandom> nice1!
14:28 < ughabugha> Janonymous: I wouldn't call that talking. ;)
14:28 < ughabugha> It was more like shouting over a distance of a few kilometers.
14:28 < jrandom> wait, y'all did voice?
14:29 < ughabugha> IIP <Janonymous> :) it was one message
14:29 < ughabugha> :)
14:29 < ughabugha> Well, I caused all the lag.
14:30 < ughabugha> Anyway, if somebody has a proposition for discussion, do it now, because I have to go.
14:30 < jrandom> word, 90m is a good meeting length to end at too...
14:30 < jrandom> anything else can be taken up on the mailing list
14:30 < jrandom> (and/or iip/i2p later)
14:31  * jrandom winds up the *baf*er...
14:31  * jrandom *baf*s the meeting closed