Distro/irc20051010

this was originally scheduled for October 11th, 2005.

date: October 10, 2005
checkin time: 5:15 PDT -0700 (8:15pm EDT)
meeting time: 5:30 PDT -0700 (8:30pm EDT)
meet in irc.freenode.net #freegeek-distro

Agenda

commitments

from last meeting:

martin(carryover):

get tested a freekbox3 package release
post instructions to pdx QC process.
finish svn, add user access (either ssh or webdav)

rob:

(carryover) spend 4 hours over next two weeks helping document putting packages packages in repository (michael wanted to help with this "as long as I can poke you enough where I fail to understand")
[8:55pm] <tymp> commit : post log
[8:24pm] <tymp> commit : weigh in on kde vs gnome on the distro list
[8:42pm] <tymp> commit : try submitting to skippy's popcon server
[8:43pm] <tymp> ugh, commit: i will set popcon up on fgoly's office machine
[8:58pm] <tymp> commit : hassle on the list for a facilitator for next meeting and agenda items

michael:

I will email the adoptionteachers list [to arrange a meeting to discuss developing user personae]
get feedback from build teachers on user personae.
[8:24pm] <WormRunner> commit: WormRunner will talk on the list

vagrant:

(carryover) hand-roll package repository scripts until a better method is devised.

matteo:

join the list discution on kde v gnome

ongoing items:

user-oriented-design
1. Someone in adoptionteachers wants to use the personae for teacher orientation
2. Rob (barely) started reading the alan cooper book (cooper takes credit for inventing the persona-driven design thing)
Prognosis (how is this project looking?) & Long-term outlook (including definition of objectives, estimation of required work, and timeline)
hosting our own popularity-contest server
Install CD: multi-profile or otherwise?
svn repository hosted at svn.freegeek.org ?
how is the meeting time/date (tuesdays, 7pm, -0700) for active participants?
packaging & repository
1. Vagrant's packaging class in PDX in Nov ? - Who will attend, or help plan?
2. "mini-dinstall would be most useful for developers making experimental releases and cvs builds, reprepro would be best for standard archive format for our released packages, and our hand-rolled one could be as good as reprepro, but more flexible ... but it would require writing it."

new items:

(posted friday before the meeting or sooner)

[i think these are not too late. :) -- baxrob 03:01, 8 Oct 2005 (PDT)]

How to propose and approve changes to FreekBox3
Shall we Fork?
1. Is it forking? could it be branching?
How can we set goals and deadlines?
How can we help Martin get SVN up?

late items:

(posted saturday before meeting or later) may not have time

late items here

log / summary

meeting content should get posted here. some possible useful categories below.

Present

Summary

Full Log

18:09 -!- vagrantc [n=vagrant@bgp01383936bgs.montbl01.nm.comcast.net] has joined #freegeek-distro
18:09 -!- Topic for #freegeek-distro: next meeting: http://wiki.freegeek.org/index.php/Distro/irc20051011, er 20051010 ?
18:09 -!- Topic set by vagrantc [] [Fri Oct 7 20:47:00 2005]
18:09 [Users #freegeek-distro]
18:09 [ meho] [ skippy] [ stillflame] [ tymp] [ vagrantc]
18:09 -!- Irssi: #freegeek-distro: Total of 5 nicks [0 ops, 0 halfops, 0 voices, 5 normal]
18:09 -!- Channel #freegeek-distro created Tue Aug 16 20:18:10 2005
18:10 -!- Irssi: Join to #freegeek-distro was synced in 40 secs
18:12 * vagrantc made it
18:12 < skippy> huzzah
18:13 < vagrantc> a nasty storm knocked out my internet connection a couple days ago, so i'm over at a freind's house
18:14 < skippy> I got the kids tucked into bed early, and the wife's out at a painting class, so I'm all set
18:15 < vagrantc> got a rambunctious youth vying for my attention tonight.
18:16 < vagrantc> if no one else has volunteered to facilitate, i can do it, too.
18:17 -!- gate [n=gate@cpe-24-210-70-207.columbus.res.rr.com] has joined #freegeek-distro
18:17 < vagrantc> hopefully the new meeting time will be better for everyone
18:18 < gate> its in 12 min, right
18:18 < vagrantc> that's the theory.
18:19 < gate> was not sure what time was finally decided on. I see about 120 unread emails I still need to go tough
18:20 < vagrantc> i think start time 5:30 ... not sure if "checkin time begins earlier"
18:21 < skippy> do we have a facilitator ?
18:21 < vagrantc> i can facilitate, if need be.
18:21 < vagrantc> i'm less involved in most of the discussion, i think.
18:22 < skippy> so gate, how's the job?
18:22 < skippy> digging into fun stuff yet?
18:30 < skippy> and... go!
18:30 < vagrantc> who's present?
18:30 < skippy> Me me me!
18:31 * vagrantc fires up a browser to point at the agenda
18:32 < vagrantc> so... it's kind of confusing to have the agenda on the wrong date :)
18:32 < skippy> that doesn't exactly help, no.
18:32 < vagrantc> tymp: you around?
18:32 < vagrantc> meho: what about you?
18:33 < skippy> gate: ping
18:33 < vagrantc> stillflame: you around?
18:33 * vagrantc wonders if there was some social event in portland tonight
18:34 * meho just got off the buss
18:34 < skippy> today is Columbus Day. Would y'all be celebrating that? ;)
18:34 < vagrantc> doubtful.
18:34 < vagrantc> i mean, freegeek is closed on saturday for columbus day.
18:34 < vagrantc> i guess some people celebrate it, and some people protest it, and some people could care less. both like the time off.
18:35 < vagrantc> shall we review commitments?
18:35 < gate> I'm here, who would have though finding a usb cable would take so long
18:37 < skippy> I had no commitments.
18:37 < vagrantc> tymp: you're the one who nudged this meeting time into existence, where are you? :P
18:38 * meho did not do a good job at "join the list discution on kde v gnome" but managed a fb.oly convertion
18:38 < vagrantc> i basically gave up on the new repository set up for the moment.
18:38 < skippy> gate: did you respond to tymp re: the GNOME takeover business ?
18:39 < gate> nope, I have a lot of email to dig though still.
18:39 < skippy> ok
18:40 < meho> wormrunner did email the list(s), im not sure of what feed back he may of gotten
18:40 < vagrantc> so... all the other commitments are from people not present.
18:42 < vagrantc> any more to say about commitments?
18:42 < vagrantc> i'd basically like to drop my current commitment for the time being.
18:42 < vagrantc> (package repository stuff)
18:43 < skippy> do you care to share a reason (you don't have to )
18:43 < vagrantc> it turned out to be trickier than i thought to get all the features i was hoping for.
18:44 < skippy> fair enough.
18:44 < vagrantc> we could probably implement some other system that's less ideal without too much hassle.
18:44 < vagrantc> and i can continue to work on that.
18:45 < gate> were u still planning on implementing one of the aleratives still
18:45 < vagrantc> it's not my biggest priority, but wouldn't be hard to set up mini-dinstall somewhere, and set up a group to with write access.
18:46 < skippy> vagrantc: care to "commit" dropping the previous action, so the logs can be grepped to catch that ?
18:47 < vagrantc> commit: set up mini-dinstall package repository for freekbox3 packages
18:48 < vagrantc> shall we move on to ongoing items
18:48 < skippy> sure
18:48 < vagrantc> anyone have much to say about user-oriented design?
18:48 < skippy> http://tango-project.org/Tango_Desktop_Project
18:49 < skippy> http://betterdesktop.org/welcome/?q=data
18:49 < skippy> not stictly U-O-D, but similar.
18:49 < vagrantc> care to summarize something about each of them?
18:50 < skippy> Tango Project looks to be an outgrowth from Red Hat's "Bluecurve" unified theme for both KDE and GNOME; with an eye toward freedesktop.org spec interoperability
18:50 < skippy> "The Tango Desktop Project exists to create a consistent user experience for free and Open Source software with graphical user interfaces."
18:50 < skippy> betterdesktop is Novell's project to do some real testing of interface designs.
18:51 < skippy> including videos of actual tests. Even if not directly applicable to us, we might glean some testing methodologies should we ever decide to execute tests on our own
18:52 < skippy> for the graphically-abled, the Tango project might be something to which one could easily contribute: http://tango-project.org/Tango_Icon_Gallery
18:53 < vagrantc> should we try and get someone to commit to looking into these more, and post to the list about it?
18:53 < skippy> well...
18:54 < skippy> 1) if GNOME and KDE are both used for future FreekBox incarnations, Tango might help simplify UI differences between the two
18:54 < skippy> 2) if Tango flops, any effort spent on it might well be wasted effort. =(
18:55 < skippy> I have _zero_ visual / graphical arts capacity, so I can only share the info about the project, and not contribute in any more meaningful way
18:55 < vagrantc> same here.
18:55 < gate> skippy: do you know how much pat looked into it. I though the orig post came from him
18:55 < skippy> gate: All I know is the post that Pat sent, which included nothing but the link.
18:56 < skippy> I can follow up with Pat.
18:56 < vagrantc> i take it this is a post to a list for freegeek columbus?
18:56 < vagrantc> was there much discussion?
18:57 < skippy> vagrantc: yes. no, not much discussion.
18:57 < skippy> http://freegeekcolumbus.org/archives/fgcmh/2005-October/000645.html
18:57 < skippy> http://freegeekcolumbus.org/archives/fgcmh/2005-October/000646.html
18:58 -!- vagrantc changed the topic of #freegeek-distro to: meeting: ongoing items: user-oriented-design | full agenda: http://wiki.freegeek.org/index.php/Distro/irc20051011
18:59 < vagrantc> is there more on the issues? any feedback from the adoption or build lists?
18:59 < vagrantc> meho: ?
18:59 < gate> ok, I think we need to head out to portland and wake them up
19:00 < vagrantc> commit: head to portland and wake them up
19:00 < vagrantc> i'll do that in a few weeks
19:00 < vagrantc> but i may not have it done by next meeting.
19:00 -!- iron_jeff [n=jet@dialup-4.246.24.91.Dial1.SanJose1.Level3.net] has joined #freegeek-distro
19:01 < meho> I asked people on this topic, and the persona topic, there wasnt mutch feed back
19:01 < vagrantc> shall we move on to prognosis and long-term outlook ?
19:01 < skippy> it seems as though folks either "get this", or they don't.
19:01 < skippy> those that don't, don't see it as a particularly useful exercise.
19:02 < vagrantc> skippy: get the user-oriented-design?
19:02 < skippy> vagrantc: yes; but we can move on.
19:02 < vagrantc> and persona.
19:02 * vagrantc agrees with skippy on both counts.
19:02 < gate> i don't recall the email that went out but did it jump right into the persona
19:02 < vagrantc> ok ... so how is the project looking?
19:04 < skippy> As a whole, I am excited that there is interest in the project enough to generate regularly attended meetings!
19:04 < gate> want to jump to the tech stuff for a while, everyone is allreay asleep and I just got the vagrantcs email from yesterday
19:04 < skippy> Even if we don't agree on the nitty-gritty of the project implementations
19:04 < vagrantc> yes.
19:05 < vagrantc> and even if the meetings are only half-attended :)
19:05 < gate> we do have someone new, hello there iron
19:05 * vagrantc thinks it must have been the day before yesterday
19:05 < vagrantc> iron_jeff: are you here for the distro meeting?
19:06 [Users #freegeek-distro]
19:06 [ gate ] [ meho ] [ stillflame] [ vagrantc]
19:06 [ iron_jeff] [ skippy] [ tymp ]
19:06 -!- Irssi: #freegeek-distro: Total of 7 nicks [0 ops, 0 halfops, 0 voices, 7 normal]
19:06 < iron_jeff> Yep! Dave told me about these. I am Jeff Trull.
19:06 < vagrantc> iron_jeff: welcome. a url for the agenda is in the topic.
19:06 < gate> so what freegeek u from jeff
19:07 < skippy> hi iron_jeff.
19:07 < iron_jeff> Thanks. I plan to lurk, basically. Gate: I guess you could say PDX! I was helping with the softmodems.
19:07 -!- vagrantc changed the topic of #freegeek-distro to: meeting: ongoing items: prognosis and long-term outlook | full agenda: http://wiki.freegeek.org/index.php/Distro/irc20051011
19:09 < vagrantc> i think we have a better idea how to work together ... at least with the packages and cd installer work (even though there's some disagreements)
19:09 < skippy> I agree.
19:10 < gate> vargrantc: what idea did you have to overcome what was talked about in the email
19:10 < gate> do we want to start having lots of little post-install scripts for the debian packages
19:10 < vagrantc> gate: i would have given them if i had ideas :)
19:11 < vagrantc> gate: do you mean in the packages? most of what we're doing needs to be done outside of debian packages to remain policy compliant. we could ignore policy, but then we loose a lot of what's good about debian.
19:11 < skippy> should we quickly skim through the remaining on-going items ?
19:11 < gate> ya, the current way kind of goes on the assumption they only a desktop is being installed
19:12 < vagrantc> sure. i'm happy to move on. popularity-contest server ?
19:12 < gate> ok, will head back in a few
19:13 < skippy> it's easy enough to install the CGI script that handles incoming submissions. I've not yet tried to set up an email alias that delivers smtp-submitted data to the popcon queue
19:13 < gate> though we wanted to only deal with cgi though
19:13 < vagrantc> skippy: you mean the debian popcon queue?
19:13 < skippy> vagrantc: no, the local popcon queue
19:14 < vagrantc> skippy: the part that processes them and makes reports and such?
19:14 < skippy> the raw data gets staged in files, and then a separate script processes the collection of all data and generates the summary reports
19:14 < vagrantc> oh. you're just talking about submit methods ?
19:14 < skippy> so, before the summary is run, I call that a queue.
19:14 < vagrantc> got it.
19:15 < skippy> I did not set up (nor try to set up) an email alias; though I imagine it's not terribly difficult.
19:15 < skippy> the HTTP POST mechanism just sends to a perl CGI. Quite easy.
19:16 < skippy> hosting our own popcon server(s) ought not be a problem. Reliably submitting data from adopted computers might be.
19:16 < vagrantc> skippy: do you need anything from the group to continue with it?
19:16 < vagrantc> yeah, reliably submitting will be a challenge.
19:16 < skippy> vagrantc: if folks want to submit data, I can then test the analyzer script, to see how it works.
19:16 < skippy> I have not run the analysis on the two submissions I've had.
19:17 < vagrantc> skippy: i already submitted data from my laptop.
19:17 < skippy> right, and I submitted from my office machine.
19:17 < skippy> If a few more folks can submit, I'll run the analysis and share my experiences.
19:17 < skippy> SUBMIT !!
19:17 < vagrantc> i think the way we need to submit the data will be to put a hook so that when networking becomes available, it will submit once a week.
19:17 < skippy> sounds reasonable.
19:18 < vagrantc> skippy: i can set up submissions on a couple additional machines.
19:18 < vagrantc> meho: would you be able to set it up anywhere?
19:18 < vagrantc> commit: set up a few machines to post to freegeekcolumbus's popcon
19:18 < skippy> so we'll need a token file to keep track of the last submission, and then check that files timestamp on ifup ?
19:19 < vagrantc> that's basically what i was thinking.
19:19 < skippy> after successful posting, touch the token file
19:19 < meho> vagrantc: do you mean a popcon server here
19:19 < vagrantc> meho: not a server, just submitting to skippy's popcon server
19:19 < skippy> meho: no, just post popcon client data to our server
19:20 < meho> commit: set up a few boxes to submit to skippys popcon server
19:20 < skippy> http://freegeekcolumbus.org/cgi-bin/popcon.cgi
19:21 < meho> only freekbox or any box?
19:21 < vagrantc> skippy: i was basically thinking of setting up up for it's normal cron.weekly popcon submission, and having an extra hook in the ifup stuff to ensure it gets sent. it might end up getting submitted twice a week, but that shouldn't be a big deal.
19:21 < skippy> multiple submissions will simply result in the newest data overwriting the older data for that client.
19:21 < skippy> meho: freekboxes, pleas
19:21 < skippy> *please
19:22 < skippy> and it's worthwhile for *us* to note that popcon weights _against_ daemons and long-lived processes because it evaluates binary's atimes.
19:23 < skippy> so a daemon that spawns and doesn't reload, or re-read the binary file, will not be seen by popcon as a "popular" program on the host
19:23 < skippy> for most Freekboxes this is likely a non-issue, but it's worth keeping in mind.
19:23 < vagrantc> skippy: i think for our little experimentation, any machine that gets used regularly should be fine, as popcon ignores really old data.
19:24 < skippy> that's all. Next.
19:24 < vagrantc> skippy: the daemon issue is largely only for machines that don't get rebooted often.
19:24 -!- vagrantc changed the topic of #freegeek-distro to: meeting: ongoing items: install cd
19:25 -!- vagrantc changed the topic of #freegeek-distro to: meeting: ongoing items: install cd | full agenda: http://wiki.freegeek.org/index.php/Distro/irc20051011
19:25 < vagrantc> i've made a few CD builds, and did some work to try and make a new simple-cdd release, so people don't have to grab from subversion.
19:25 < vagrantc> commit: release new simple-cdd packages
19:26 < vagrantc> gate: have you made any CD images ?
19:26 < gate> not in over a week
19:26 < skippy> all the CD images I've made have failed the profile execution.
19:27 < gate> and they were failing on me, yup that where they were blowing up on me too
19:27 < gate> the first ones I made worked though, not sure what changed
19:27 < vagrantc> gate: worked in what sense?
19:27 < gate> it installed the correct profiile and most everything went right
19:28 < gate> the thing that was wrong was kde got installed too
19:28 < vagrantc> heh.
19:28 < gate> but I think that was a mistake on my packing that I think I fixed
19:28 < gate> in your last commit did u change any of the underlying code
19:28 < gate> or just update profiles
19:29 < vagrantc> the base-config menu is, i believe, some debconf questions that are either not present in the gnome preseeding, or some questions that *are* present that cause it to break.
19:29 < vagrantc> but i'm pretty sure it's debconf related.
19:30 < vagrantc> gate: i've updated some of the packaging stuff, but haven't made the move to include your takeover changes, as i'm not sure how to proceed.
19:31 < gate> let me post to the email list. I have some ideas but boy is it just getting more complicated
19:31 < vagrantc> it is hard to get it right, but even harder when you do it wrong, in the long-term.
19:32 < vagrantc> not there there is exactly one right way ...
19:33 < gate> I'm just not sure if we want to segment it so much
19:33 < gate> the debian packages are just logical grouppings
19:33 < vagrantc> i think the takeover handling and packages are two totally different things.
19:34 < gate> which is why I have moved pretty much all of our stuff into the deb packings deps
19:34 < vagrantc> the takeover stuff is very specific- to produce an exact replica of a system type, whereas the packages only provide package selections
19:35 < vagrantc> gate: but some of the package dependencies/recommends are arguably not appropriate to the -desktop package, in my opinion
19:36 < gate> to the freekbox3-gnome-desktop they are
19:36 < vagrantc> since we're a little off-topic, should we change the topic ?
19:36 < gate> yes, any other desktop maybe they should not be there
19:37 < vagrantc> gdm isn't part of the user's desktop, it's part of the system GUI
19:37 < vagrantc> freekbox3-tools have nothing to do with the desktop
19:37 < gate> I though I had that in the takeover package list. that is one of the few packages that I did not want to make a dep
19:38 < vagrantc> it's a recommends, which is a "soft" dependency, basically.
19:38 < gate> which is what it is. but if kdm or xdm is installed it does not remove it, right
19:38 < vagrantc> and there are other packages which are a little unclear, too, like nautilus-cd-burner - not every desktop system will have a CD burner ...
19:39 < gate> other than a little bit of wasted space I see no hard in that.
19:39 < vagrantc> gate: yes, but if someone uses "aptitude install freekbox3-gnome-desktop" to install your software set, it will automatically pull in gdm.
19:40 < vagrantc> my intention with the -desktop packages is to provide packages that can be installable on non-freekbox systems, too. otherwise, there is no real point in separating them.
19:40 < gate> that was not my intent with making it gdm a recommend.
19:41 < gate> I only wanted it to install if xdm, kdm .. what not allready insatlled
19:41 < vagrantc> in my opinion, it should be handled elsewhere- like in the takeover scripts
19:41 < gate> why, can't the packing system handle it
19:42 < skippy> if it's recommended, apt will not automatically install it, will it?
19:42 < vagrantc> there's no technical reason, but i don't see any purpose to having *-desktop packages, then.
19:42 < gate> skippy:, thats what I though
19:42 < vagrantc> skippy: apt will not, but aptitude is recommended, and numerous other tools will as well.
19:43 < skippy> ok, so could we switch to "suggests" instead of "recommends" ?
19:43 -!- everestx [n=chatzill@d149-67-91-146.col.wideopenwest.com] has joined #freegeek-distro
19:43 < skippy> hi chris!
19:43 < vagrantc> suggests would be fine.
19:43 < vagrantc> i made the same mistake with the kde desktop, too.
19:43 < gate> thats really all I was trying to do. I did not want to force it to be installed
19:43 < skippy> vagrantc: do you know if Synaptic automatically installs "recommends" in the same way that aptitude does ?
19:43 < vagrantc> skippy: i don't.
19:43 < skippy> since I think more people will be using synaptic than aptitude
19:44 < vagrantc> it still would be more appropriate as a suggests.
19:44 < skippy> sure, that's' fine.
19:44 < vagrantc> is the freekbox3-services packages fine with folks?
19:45 < everestx> skippy: hi, just going to try to stay up with the convo.
19:45 < gate> only if I can make a freekbox3-bob package
19:45 -!- WormRunner [n=earthwor@209-162-205-24.dq1mn.easystreet.com] has joined #freegeek-distro
19:46 < vagrantc> gate: what, pray tell, would be the purpose of "freekbox3-bob" ?
19:46 < skippy> the level of granularity in packages is getting confusing to me. It seems these other "things" aren't exactly "FreekBoxes" as I understand the term.
19:47 < skippy> I see "FreekBox" as being a unified collection of software applied to the computer given to adopters.
19:47 < vagrantc> skippy: but it is important (at least to me) to have all these functionalities split out into separate packages.
19:47 < gate> vagrantc: he is the master of all meta packages so anything our minds dream of
19:47 < skippy> vagrantc: fine, I don't necessarily disagree with granularity on its own; but are they Freekbox-granular?
19:47 < vagrantc> skippy: and this is important to prevent the need for forking.
19:47 < skippy> gate: did you suddently become a SubGenius ?
19:48 < gate> bob is a little private joke that I only get
19:48 < vagrantc> skippy: the granularity is necessary for the kde/gnome variants, and will also make the package selections more useful for other non-freekbox systems that freegeeks may use.
19:48 < skippy> gate: oh. I thought you meant J.R. "Bob" Dobbs from the Church of the SubGenius
19:49 < gate> their is the planet bob
19:49 * vagrantc finds the off-topic meter off the scale
19:49 < skippy> vagrantc: refresh my memory on what "freekbox-services" includes.
19:49 < vagrantc> one more comment about installer CDs specifically
19:50 < vagrantc> skippy: all the services that the user will want, but does not interact with directly- printing, network autoconfiguration, sound ... cups/xprint, ifplugd, alsa
19:50 < vagrantc> are people happy with having gate (or others) work on a single-profile CD image, while i continue to work on a multi-profile image?
19:51 < skippy> vagrantc: I see most of those as fundamentally necessary to the successful build of a usable workstation for an adopter, so splitting int into "services" seems extraneous for _our_ purposes.
19:51 < gate> it all looks good to me
19:51 < skippy> vagrantc: I'm not unhappy with that separation of labor; and i think much of it can be merged back together at a later date. I know gate is working on a single profile to minimze the number of variables he's dealing with at any one time
19:52 < vagrantc> skippy: for my purposes, it makes it easier to install a freekbox-like set of applications on an application server, without having to install conflicting print system, sound system, or network packages
19:52 < vagrantc> skippy: yes, a single profile reduces variables greatly.
19:53 < skippy> vagrantc: here again we run into the philosphical distinction between what I perceive as the distro group's immediate goal (a single workstation), and your propogation goals of terminal servers and the like.
19:54 < skippy> perhaps refinement of the goals is in order.
19:54 < vagrantc> perhaps.
19:54 < gate> drop lessdisk and use ltsp, right
19:55 < gate> ok maybe not
19:55 < vagrantc> but having the kde and gnome variants requires a certain amount of granularity. i *feel* the additional cost for further granularity is small.
19:55 < skippy> vagrantc: let me ask: what's the benefit to said additional granularity?
19:55 < vagrantc> gate: ltsp isn't in debian sarge, and the packages in debian etch have a long way to go.
19:55 < skippy> I recognize I'm being somewhat two-faced in that question
19:55 < skippy> since I've not adequately addressed the perceived benefits of GNOME's additional packages
19:55 < vagrantc> skippy: the biggest benefit is not having to implement a separate distro for application server images.
19:56 < gate> vargrantc: I'm just joking. I have no desire to do either for a long time
19:56 < skippy> vagrantc: and, for clarity, how many application servers do you see benefitting from the work?
19:57 < vagrantc> skippy: well, freegeek portland has something like 5 (or more) servers.
19:57 < gate> my only objection to the whole multi profile is what happens when it out grows a 650meg CD. But I do like the idea of it.
19:57 < vagrantc> skippy: but one of the barriers to implementing is is a technical challenge. so i work to reduce the technical challenges.
19:58 < gate> but since a profile and be dropped and added with out any real work its a good place to work
19:58 < gate> s/and/can/
19:58 < vagrantc> gate: that's the idea.
19:59 < skippy> okay, so the end result with the addition of freekbox-services is that what I'm calling the "freekbox" (that is, the end-user system) merely requires a few additional packages?
19:59 < skippy> the contents of which would otherwise have been in the freekbox packages ?
19:59 < vagrantc> yes.
19:59 < skippy> and it would be "require" not "recommend" or "suggest" ?
19:59 < vagrantc> yes.
20:00 < vagrantc> (actually, the contents used to be in the takeover scripts, and not in the packages)
20:00 < skippy> it seems like an unnecessary abstraction for the task that most people will expect of the freekbox, but I guess I don't have any other objection than that, at this time.
20:00 < vagrantc> apt-get install freekbox3 will behave identically.
20:01 < vagrantc> and the takeover script will behave identically.
20:01 < skippy> off-topic: can the takeover script be called from a post_install .deb script, so that `apt-get install freekbox3-<foo>` will actually install the freekbox?
20:02 < skippy> where <foo> is some token to indicate consent to the process that will commence ?
20:02 < gate> no
20:02 < vagrantc> not in a policy compliant way.
20:02 < vagrantc> which is why we have the whole takeover mess.
20:02 < skippy> gotcha. thanks. back to schedule.
20:03 < gate> and takeover does package stuff which it can't do because the database is locked
20:03 < gate> we on to next topic yet
20:04 < skippy> svn.freegeek.org resolves. i don't know how complete it is.
20:04 < vagrantc> we could make a debian package that is a policy breaker, to simplify some things, as the takeover scripts are really just technically compliant with debian policy- they don't reflect the spirit of policy.
20:04 < vagrantc> "meow"
20:04 < vagrantc> is what http://svn.freegeek.org gets us.
20:04 < skippy> vagrantc: have a handy link to the policy in question ?
20:05 < skippy> http://svn.freegeek.org/trac/
20:05 < gate> I was going to ask if we are getting to that
20:05 < gate> but I think that has already been answered
20:06 -!- vagrantc changed the topic of #freegeek-distro to: meeting: ongoing items: svn.freegeek.org | full agenda: http://wiki.freegeek.org/index.php/Distro/irc20051011
20:06 < skippy> stillflame isn't participating tonight. Last I knew, it was mostly done.
20:06 < skippy> what's left to implement I don't know.
20:06 < skippy> whether user accounts are available, I don't know.
20:06 < vagrantc> gate: i was wondering if maybe you'd continue to maintain your changes on the old "freebox" svn project ?
20:06 < gate> off topic, whats the max speed of usb1
20:07 < gate> vagrantc; I did that over the weekend
20:07 * vagrantc brings out the off-topic bat
20:07 < skippy> 1.1 M/s ?
20:07 < gate> offtopic, ok, that 40G is going to take some time to dd
20:07 < vagrantc> gate: great!
20:07 < vagrantc> who's going to hassle martin about svn ?
20:08 < vagrantc> i guess i will.
20:08 < vagrantc> commit: hassle martin about svn
20:08 < skippy> I try to offer encouragement and offers to help when I see him...
20:08 -!- everestx_ [n=chatzill@d149-67-91-146.col.wideopenwest.com] has joined #freegeek-distro
20:08 * skippy blinks at the everestxs
20:08 < vagrantc> any more svn discussion?
20:08 < skippy> no
20:08 < vagrantc> what do people think of the new meeting time?
20:09 < gate> i like it, I'm still awake
20:09 < vagrantc> WormRunner: you showed up a little late, no?
20:09 < skippy> 's'okay for now, for me. May change in the future.
20:09 < vagrantc> it's good till the end of the month, for me. maybe a little longer. hard to say.
20:10 * vagrantc hopes to get internet access at home un-borked
20:10 -!- vagrantc changed the topic of #freegeek-distro to: meeting: ongoing items: meeting time | full agenda: http://wiki.freegeek.org/index.php/Distro/irc20051011
20:11 < gate> so was that everything
20:11 < vagrantc> packaging ...
20:11 < vagrantc> when i go to portland, november-decemberish, i hope to teach a packaging class.
20:12 < vagrantc> i don't know that more needs to be said about that.
20:12 < gate> something tells me I will not be there
20:12 < vagrantc> and there's a summary of the various ideas about package repositories, which has been covered in past meetings and on the list.
20:12 < skippy> video tape it, and publish a torrent of the mpeg
20:13 < vagrantc> skippy: excellent idea!
20:13 < skippy> vagrantc: I hope to video tape tomorrow's class at FGCMH
20:13 < vagrantc> so... there's some new items...
20:13 < gate> and get all screen captures
20:13 < gate> over better yet use kde and script the whole thing
20:13 < skippy> How to propose and approve changes to FreekBox3
20:14 -!- vagrantc changed the topic of #freegeek-distro to: meeting: new items: propsing/approving changes to the freekbox3 | full agenda: http://wiki.freegeek.org/index.php/Distro/irc20051011
20:14 < gate> hah, just get commit rights
20:14 < vagrantc> well, that doesn't publish packages.
20:14 < skippy> there ought to be some sort of review process
20:14 < gate> I think all you have to do is post to the list and if you don't get shoot down it will end up in their
20:15 < vagrantc> that's basically how it's gone for the last several years.
20:15 < gate> oss is a big review process. bad ideas normally get removed
20:15 < vagrantc> it doesn't really feel right to me ... but i also don't want to get overly beurocratic about it either.
20:16 < skippy> I think it's reasonable to expect at least a modicum of justification from the person suggesting the change.
20:16 < vagrantc> absolutely.
20:16 < skippy> they should be prepared -- and able -- to field questions from others about the proposal.
20:17 < gate> which is what posting to the mailing list does
20:17 < skippy> the topic is specifically about the FB3. More generally speaking, I think we should use Trac (or *shudder* RT) tickets to monitor new feature requests
20:17 < vagrantc> though only with medium-to-large scale changes... maintenance and bugfixes and such should be ok without an involved process.
20:19 < skippy> who will have commit privileges?
20:19 < skippy> anyone who wants them?
20:19 < vagrantc> that is something that needs a little more review, for sure.
20:20 < vagrantc> at the moment, anyone who has access to any cvs project has access to all projects.
20:20 < skippy> ick
20:20 < vagrantc> at the same time, it is a bad idea to be too restrictive in who has commit access- especially with small projects.
20:21 -!- everestx__ [n=chatzill@d149-67-91-146.col.wideopenwest.com] has joined #freegeek-distro
20:21 < meho> what the coders are talking about http://lists.freegeek.org/pipermail/coders/2005-October/001473.html
20:21 < gate> u can always uncommit if someone goes crazy
20:22 < skippy> woah, keep it up everestx__ ! You'll have the longest nick in history! =)
20:22 < vagrantc> gate: right.
20:22 < vagrantc> gate: and remove access.
20:22 < vagrantc> (if necessary)
20:22 -!- everestx [n=chatzill@d149-67-91-146.col.wideopenwest.com] has quit [Read error: 110 (Connection timed out)]
20:23 -!- everestx__ is now known as everestx
20:23 < skippy> i don't know how much interest there will be for commit access; so it may prove to be a non-issue. But if we don't have guideline, chances are significantly greater that we'll wish we had one.
20:24 < vagrantc> indeed.
20:24 < skippy> since SVN is still unknown, we can probably place this on hold for tonight.
20:24 < vagrantc> to me, the hard part is giving people access without meeting them, or at least having a gpg trust relationship
20:25 < vagrantc> but yes, this can be brought up on the list.
20:25 * vagrantc has an idea for later
20:25 < skippy> vagrantc: several of us in CMH have been tentatively kicking around the idea of a weekend trip to PDX, to see first-hand the operation there. PGP key signing could occur then, if we do it.
20:26 < vagrantc> skippy: that would be excellent. earlier today i had the idea of getting funding for a conference to help with expenses and such.
20:26 < skippy> groovy.
20:26 < vagrantc> basically, i think we should post this to the list.
20:27 < skippy> agreed.
20:27 < vagrantc> next issue is forking/branching ...
20:27 < gate> crap, I need to get a key
20:27 < skippy> FORK YOU!
20:27 < skippy> sorry, I've been waiting all night to say that. ;)
20:27 < gate> I like spoons, they are safer
20:27 -!- vagrantc changed the topic of #freegeek-distro to: meeting: new items: forking/branching | full agenda: http://wiki.freegeek.org/index.php/irc20051011
20:28 < gate> who put that one on
20:28 < vagrantc> basically, i think we're moving more in the direction of a branch than a fork.
20:28 < vagrantc> tymp: you loaded this agenda, moved it to now, and don't even show up? :P
20:29 < skippy> i agree. We're using (and collaborating on) the same fundamental bases: Debian and the core FreekBox services / apps
20:29 < vagrantc> documentation will probably be the hardest part.
20:29 < skippy> most time consuming, at any rate.
20:29 < vagrantc> right.
20:30 < gate> I believe there was some talk of that at the last columbus meeting
20:30 < gate> but I was playing hookie so can't give much more detail
20:30 < skippy> The FB3 manual is very conversational. Much of the manual could be used as is.
20:30 < vagrantc> time consuming enough that it might not get done ... the documentation is arguably not good enough as is ...
20:30 < gate> I still don't like it being in open office
20:30 < skippy> I would prefer to see a more functional manual -- a guidebook.
20:30 < skippy> gate: so export it.
20:30 < gate> makes it harder to diff it out
20:31 -!- everestx [n=chatzill@d149-67-91-146.col.wideopenwest.com] has quit ["Chatzilla 0.9.68a [Firefox 1.0.7/20050915]"]
20:31 < vagrantc> yes, the openoffice decision was basically the main person working on it deciding to switch to openoffice.
20:31 < vagrantc> i think it's a bad decision.
20:31 < skippy> i don't like it being one monolithic file. one file per chapter seems better. That way we could mix-and-match backend differences easier.
20:31 < vagrantc> indeed.
20:32 < gate> the doc person is on the lists so might want to suggest that
20:32 < skippy> writing it in OOo makes it a lot easier for users of all skills to contribute. Writing something that long in plain HTML is cumbersome, and many people don't want to learn docbook
20:32 -!- everestx_ [n=chatzill@d149-67-91-146.col.wideopenwest.com] has quit [Read error: 110 (Connection timed out)]
20:32 < vagrantc> time check: how much longer do people want to meet?
20:32 < skippy> i'd like to wrap up soon.
20:33 < vagrantc> skippy: yes. that's the pro of openoffice.
20:34 < vagrantc> ok ... i'd like to say something about deadlines ...
20:34 < vagrantc> well, goals and deadlines
20:35 < vagrantc> and i already committed to getting some info or action from martin about svn.
20:35 < vagrantc> any more forkbranch discussion?
20:35 < gate> nope
20:36 < vagrantc> as far as goals, i'd like to get new packages, new simple-cdd and a new installer CD built by the end of october.
20:36 < vagrantc> new freekbox3 packages, that is,.
20:36 < skippy> i like that timeframe.
20:37 < vagrantc> i don't know that i could integrate a gnome profile into that image ...
20:37 < skippy> i think we could agree to concurrently make a stand-alone GNOME installation in that same time frame.
20:38 < skippy> I'd like to try to learn more about D-I, and making better use of it to preseed; but I don't know if Oct. 31 is enough time...
20:38 < vagrantc> i am not in a position to test CD images- can anyone offer to test images i build?
20:38 < skippy> vagrantc: sure.
20:39 < skippy> I have a test system I can blow away, and we can always get some systems at FGCMH.
20:39 < vagrantc> thanks.
20:39 < vagrantc> any more goal and deadline issues?
20:40 < skippy> gate: think we can get a GNOME system with installer by the end of the month?
20:40 < skippy> vagrantc: I'd like to try to "brand" the installer process, too.
20:40 < vagrantc> skippy: sure.
20:40 < gate> depends on how many changes we end up making to the install scripts
20:41 < vagrantc> yeah.
20:42 < vagrantc> being able to have the takeover scripts install both profiles could considerably slow down the CD effort- unless it goes way easier than i expect.
20:42 < gate> and we still need to do a little bit of audio clean up. I did not implement any of the stuff we did on the box before the linuxfest
20:42 < vagrantc> audio clean up ?
20:42 < gate> oss, alsa packages
20:42 < vagrantc> yes.
20:43 < gate> audo does not work right in the current gnome setup
20:43 < skippy> is it "better" that a single takeover script handle both profiles; or should each profile use its own takeover script? Or should the takeover script be pluggable in some way ?
20:44 < gate> not sure how to do it plugable
20:44 < vagrantc> i think long-term, takeover should support an arbitrary number of profiles.
20:44 < gate> the "current" form requires maybe 10 more lines for a new profile
20:45 < vagrantc> in short-term, it might be best to fork the takeover scripts, with the intention of next release supporting at least the two variants.
20:45 < vagrantc> gate: by current, you mean your modifications ?
20:45 < gate> yes
20:45 < gate> hence the ""
20:46 < vagrantc> i'd like to drop the direct comparison to packages, since the takeover customizations are separate from the packages...
20:46 < skippy> a pluggable takeover could have the common bits included in the main file, and then it could scan something like /usr/share/freekbox/profiles/ for *.takeover.py scripts to include; and then present a choice for which to use...
20:46 < gate> I'm still going to focus on supporting more than one profile
20:46 < skippy> This way, each profile could bundle its own takeover extensions as needed in the .deb
20:47 < gate> maybe scott just answered the question u posted to the list vag
20:47 < vagrantc> skippy: except you need to know what profile you're installing *before* you install the packages.
20:47 < skippy> package installation is different from takeover, I thought.
20:47 < skippy> takeover would prompt for which profile to "complete" the takeover.
20:48 < vagrantc> skippy: you mean a separate "freekbox3-takeover-gnome" package?
20:48 < skippy> no, I mean the freekbox package would include the core takeover.py; and freekbox-gnome would include a gnome.takeover.py script.
20:49 < gate> I'm thinking have the -desktop package plop a script in a special dir
20:49 < vagrantc> gah.
20:49 < gate> and then have takover scan for them
20:49 < skippy> then running takeover would query for all the available *.takeover.py scripts, and prompt the user for which to install.
20:49 < gate> yes it would cause some fun if both kde and gnome where installed
20:49 < skippy> we could get fancier, and mark in some way which one(s) were already installed
20:49 < vagrantc> the reason the freekbox3-takeover package exists is because it's impossible to do all that right.
20:50 < gate> freekbox3-takeover will still include the main package
20:50 < gate> and will just "source" the other ones if present
20:50 < vagrantc> include it how?
20:50 < vagrantc> aand "source" how?
20:50 < gate> thats what I don't have worked out yet
20:51 < skippy> maybe we should revisit the policy issue. Being completely uninformed of the matter, I don't have much problem violating policy so that freekbox-installer automatically launches the installation process...
20:51 < gate> by include I mean it will contain the install_freekbox and main takeover scripts
20:51 < vagrantc> the takeover scripts need to be run before you install the freekbox3 packages, because you need to do customizations before you install other packages (and their dependencies)
20:51 < vagrantc> such as debconf
20:52 < gate> I'm still thinking u have to run the takeover manually. it will just not contain the profile specifc stuff but if gnome-desktop was installed before hand (say by freekbox3_install) then it will do the setup.
20:52 < gate> the install script can still look for an env var set by the cd installer or prompt
20:52 < vagrantc> it's a chicken and egg problem.
20:53 < vagrantc> making policy-breaking packages might actually even be harder ... not sure.
20:54 < gate> I believe I can make it work. but it still will not auto takeover like skippy would like
20:54 < gate> unless installing from a CD
20:54 < skippy> gate: I'm willing to work with you on the CD / D-I components necessary
20:55 < vagrantc> gate: so, it would *require* installing from a CD, with your proposed idea?
20:55 < gate> no
20:56 < vagrantc> i just don't see how you can grab and tweak customizations before installing the packages themselves.
20:56 < gate> the script will still ask u what u want to install
20:56 < vagrantc> gate: maybe describe it step by step
20:56 < skippy> time for me to go. i'll read the log.
20:56 < skippy> g'night.
20:57 < gate> I'm still rolling it over
20:57 -!- skippy is now known as skippy|away
20:57 < vagrantc> gate: unless you need to go, too.
20:57 < gate> I will toss some things down in code maybe wed
20:57 < gate> and see how it goes.
20:58 < gate> i'm not sure how to pull it off yet
20:59 < gate> ok well I think we need to call it a meeting night
20:59 < vagrantc> alright.
20:59 < vagrantc> skippy|away: you got a log you can post?
20:59 < vagrantc> gate: or do you?
20:59 < gate> I don't think I log
21:03 -!- iron_jeff [n=jet@dialup-4.246.24.91.Dial1.SanJose1.Level3.net] has left #freegeek-distro ["Kopete 0.10 : http://kopete.kde.org"]
21:18 -!- gate [n=gate@cpe-24-210-70-207.columbus.res.rr.com] has left #freegeek-distro []