From Nsnam
Jump to: navigation, search

This is the archive for the IRC chat on 2008/04/29. Full transcript is below, notes follow.

Participating: Florian, George, Joe, Liu, Sam, Tom, Vincent

Major discussion points:

  • Misc
    • General consensus is to try and keep regular chats going


<sjansen> hello
<tjkopena> hi sam, nice to meet you (after a fashion)
<sjansen> likewise
<tjkopena> it looks like everyone I expected to login is here, though hagen is sleeping
<fw> guess i'll start... I've started doing a forward port of nsc 0.3.0 to ns-3
<fw> i.e. i took ns-3-dev and diffed that with ns-3-nsc
<fw> at the moment the tcp-point-to-point example looks ok
<fw> but tcp-large-transfer does not
<fw> havn't had the time to investigate yet
<tjkopena> is school over for you florian?
<fw> after that works i'll look into proper trace support 
<fw> no
<fw> i'm in the final stages of my thesis...
<fw> but hopefully i'll hand it in mid-may
<tjkopena> got it
<tjkopena> liu, when do your classes end for the spring?
<LJ__> end at June
<tom_henderson> Florian (and LJ and HGN), I'd be happy to set you up for access on our server, so you can post your repositories as you work on them
<tom_henderson> I would just need an ssh public key from you (please, use a passphrase)
<fw> i've uploaded my stuff to
<fw> but thank you for the offer.
<tom_henderson> and also, what your preferred username and shell would be
<fw> i'll gladly accept the offer later this year, after i've decided to continue working on the project post-soc
<LJ__> how can we setup our own repo?
<tom_henderson> I can tell you outside of the chat
<fw> LJ__: its easy... the mercurial documentation contains all necessary information
<tjkopena> florian, you really should move that work onto
<LJ__> but need username and password maybe
<tjkopena> make it more clear that it's part of a larger project
<fw> tjkopena: Hrm, i was a bit concerned because im reluctant to upload dirty hacks to nsnam
* vgauthier has quit (Read error: 104 (Connection reset by peer))
* vgauthier (n=vgauthie@ has joined #ns-3
<tjkopena> fw, I don't think that's a huge concern
<tjkopena> better to have people looking at it more readily
<fw> tjkopena: alright.
<tjkopena> on a different note, given that there's a few weeks before gsoc technically begins (end of may), and everyone is still in class, I think it'd be worthwhile to step back for a second
<fw> tom_henderson: ignore what i said earlier, i'll send you my pubkey later...
<tjkopena> the applications, although solid, good definitely stand to be refined into a more concrete plan
<tjkopena> it would be good to develop a better technical picture and set of milestones and get those up on the ns3 website
<tjkopena> keep in mind not everyone has seen the content of the applications, just the blurbs on the ns3 project page
<tjkopena> I would suggest everyone start a section on the wiki, using their application as the initial framework, and then craft it into a more detailed plan & presentation
<tom_henderson> tjkopena-- I agree; here is a wiki link to link from:
<tjkopena> as the work is completed it may then be filled in and subpages created.  by the time we're done, there's a nice little site for each project documenting the summer
<tom_henderson> look, for instance, at what Craig posted last night:
<tom_henderson> everyone could perhaps maintain their own wiki page with the details
<tjkopena> yes, that looks good to me
<tjkopena> for the gsoc projects especially, it would be worthwhile to develop a concrete milestone schedule to go with the kind of description craig has up
<tjkopena> liu, I think mathieu already talked to you about this, but he definitely wants to spend a good bit of time planning out your project rather than diving into coding
<LJ__> yes
<LJ__> we have talk about it early
<LJ__> my job porting Quagga Routing Suite to ns3, it mainly base on mathieu/ns-simu. i have step the first: list all system calls which need to be implemented in ns3 as simu_xxx.
<LJ__> the next is to implement them in ns_simu
<tjkopena> that's my understanding as well
<LJ__> now there is a problem, i donot clear how to run quagga in a single pc
<LJ__> as for testing
<tjkopena> Tom is probably most competent to discuss that; I don't have real experience w/ quagga
<tjkopena> but, I really think you, fw, and hgn should not worry about the coding for a little bit, and instead make a very thorough project plan
<LJ__> i am not familar with how routing deamos runs
<tjkopena> and a framework in which to document what you do as it happens
<sjansen> I think there is a fair amount of experimentation required with fw's work; I don't believe in a project plan being too detailed really.
<tjkopena> sure; I'm not talking about a huge upfront design effort
<tjkopena> I'm just talking about creating a project page much like Craig's that Tom linked to, and filling out a little more what has to get done
<sjansen> Ok fair enough.
<tjkopena> really, the applications do have most of what I'm thinking of
<tom_henderson> LJ__, I can help you in better understanding the quagga port, and in developing some interim milestone steps
<LJ__> that would be great
<tjkopena> just need a bit of cleanup, making the goals more precise and filling in some milestones
<fw> tjkopena: alright, will do.
<tjkopena> to do that, everyone should register an account on the wiki
<tjkopena> tom will approve it, and then you're allowed to edit
<tjkopena> I would start by linking from the project page tom noted above, creating a new page and copying your gsoc application over to there
<tjkopena> sjansen, one key thing then is to record the results of those experiments, ideas, and directions taken on that wiki section so it's all documented
<LJ__> Tom, have you receive my msg?
<sjansen> tjkopena: Yes, I agree with that 100%.
<tjkopena> cool; that's really the main thing I'm looking for here, a framework to record that kind of stuff, track progress, and give information to the outside world
<sjansen> The wiki seems like a good place for it all. I agree that its very useful to maintain transparency here.
<LJ__> maintain our own project wiki, write down what were doing and what will be done?
<tjkopena> well, a set of wiki pages in the ns-3 wiki
<tjkopena> much like the page craig created, but over time filled in with notes and such
<tjkopena> for example, things you learn about quagga in preparing for your project would be good to have up there
<LJ__> got it
<tjkopena> from my end, I think that's something that will be minimal effort in the next few days, but will pay off as the summer goes on
<tjkopena> other thoughts on gsoc topics?
<tom_henderson> LJ__, I have not received any msg from you (if you mean, by email)
<riley> did hagen ever log log in.  I don't recognize his log-in, so can't tell for certain
<sjansen> hrh = Hagen
<sjansen> hgn
<tjkopena> riley, no.  hgn is him, but the account seems to still be sleeping
<LJ__> oh, just now , i sent you msg personally
<riley> ok, I'll email him this transcript of the chat and coordinate with him on the task
<tjkopena> sounds good; he may have class or something else now (I'm not sure what the time is there)
<fw> tjkopena: 5:45 pm
<tjkopena> thx fw
<LJ__> Tom, i will sent you a mail tomorrow, it is 23:45 here
<tom_henderson> LJ__ OK
<LJ__> something about i run quagga
<tjkopena> given that liu should be asleep, I didn't have much more to say about gsoc topics.  everybody seems to be on the mailing list already, and more or less underway on their projects.  any other comments?
<sjansen> Not from my corner.
<riley> None from me.
<tom_henderson> no, thanks for getting this chat going
<tjkopena> how does the group feel about regular chats to get everyone together like this?
<riley> Agree this is useful.  What about each week, same time same station?
<sjansen> Seems reasonable but I probably wont make them all at this time...
<tom_henderson> this time works pretty well for me.  
<tjkopena> this time works for me, and presumably mathieu as well when he gets back from vacation
<tjkopena> we may have to change when classes end
<tjkopena> I was thinking of moving it around though so liu isn't always penalized and up late
<tjkopena> what about assuming this time next week and go from there?
<sjansen> sure
<LJ__> sorry about that, but it was ok for me
* vgauthier has quit (Read error: 104 (Connection reset by peer))
<tom_henderson> OK, I will not be around next week, though
<riley> Next week works for me.
<tjkopena> ok, so let's do that
<tjkopena> I'll send a note to the mailing list covering what we discussed & advertising for next week
<fw> tjkopena: thank you
<tjkopena> tom, I'll also email vincent about the stat framework.  I was hoping to talk about it while he was on here, but he seems to have left & I have a meeting to get to as well
<tom_henderson> OK, I can't stick around past 9 anyway
<tjkopena> liu, if you talk w/ Tom about quagga and you & I talk about moving your application onto the wiki and filling in some stuff, that'll be great for when mathieu gets back next week
* vgauthier (n=vgauthie@ has joined #ns-3
<tjkopena> hi vgauthier
<tjkopena> I was just about to email you
<vgauthier> hi 
<tjkopena> most of us have to run to different meetings, but I was hoping to talk about the stat framework
<sjansen> yeah I gotta get to work, thanks guys
* sjansen has quit ("leaving")
<tjkopena> vgauthier, I just got your last email about the framework.  I have some thoughts on your api sketch proposed earlier that I'll email later today
<vgauthier> yes, off course
<LJ__> tjkopena:ok, it will be fine, there is already a page 
<LJ__> i can edit here or start a new page
<tjkopena> liu, yes, editing from there sounds reasonable to me
<tjkopena> ok; I have to run as well.  thanks everybody
<LJ__> it will be done before mathieu back