Technical-2006-12-05

(03:08:23 PM) sabdfl: sorry i'm a bit late
(03:09:13 PM) shawarma: They require an external power source, so I think they're can shoot quite far.
(03:09:19 PM) mdz: so let's dive right in then
(03:09:21 PM) mdz: everyone is here
(03:09:36 PM) mdz: https://wiki.ubuntu.com/TechnicalBoardAgenda
(03:09:49 PM) Keybuk: mdz: are we going to be doing ubuntu-dev today, or have we delegated that to greyskull now?
(03:10:09 PM) sabdfl: do we yet have an official skeletor?
(03:10:17 PM) sabdfl: i think the delegation is something that should be done here
(03:10:34 PM) shawarma: dholbach kind of implied that hadn't happened yet..
(03:10:35 PM) mdz: note that I've begun to decline outright when folks attempt to join -core-dev who aren't members of -dev, unless there are special circumstances
(03:10:49 PM) mdz: (e.g., prior discussion or someone with a Debian background)
(03:10:56 PM) ***kylem ducks.
(03:11:36 PM) mdz: I assume the rest of you are OK with that
(03:11:37 PM) ajmitch: Keybuk: there's no motu council setup yet, so I doubt there'll be any delegation taking place until it's agreed on by TB/CC
(03:11:45 PM) Keybuk: mdz: I'm fine with that
(03:12:02 PM) mdz: sabdfl: I think that proposal is still under discussion and being revised
(03:12:10 PM) mdz: so probably not ready to take any official decision on the part of TB
(03:12:29 PM) mdz: I believe CC needs to approve the council first
(03:12:54 PM) mdz: then TB can take decisions regarding composition and delegation
(03:12:31 PM) Keybuk: ok, just wanted that clarified
(03:12:51 PM) sabdfl: personally, i'd be happy for the TB to cowboy it
(03:13:10 PM) sabdfl: CC already delegates this stuff to TB
(03:13:22 PM) sabdfl: composition should follow TeamCouncil style rules
(03:13:31 PM) sabdfl: TB should be appointing, or at least nominating
(03:13:36 PM) sabdfl: careful of elections etc etc
(03:14:17 PM) mdz: the current proposal is: https://wiki.ubuntu.com/MotuProcessesSpec
(03:14:25 PM) mdz: (part of it anyway)
(03:14:38 PM) mdz: it calls for a 5-member council
(03:15:45 PM) mdz: I'm not sure that quite that many are needed, but the overall structure seems sane to me
(03:16:03 PM) mdz: the composition of the council is not addressed by that proposal, as I believe there's an existing CC process for it
(03:16:19 PM) mdz: but in this case as you say, similar things are delegated to the TB by CC
(03:16:50 PM) mdz: personally, I don't have enough information to come up with an appropriate list of nominations today
(03:16:38 PM) sabdfl: they call for the TB to ACK new devs
(03:16:54 PM) sabdfl: seems to me that it would be better to appoint Grayskulls who we trust
(03:17:03 PM) sabdfl: TB won't be as familiar with the candidates
(03:17:11 PM) sabdfl: and TB holds the reigns on -core-dev
(03:17:14 PM) mdz: sabdfl: you explicitly asked for TB acks
(03:17:31 PM) sabdfl: i suggested it from a single member, not "the board"
(03:17:42 PM) sabdfl: if there is a TB member on MC then that's sorted
(03:17:48 PM) mdz: Keybuk: you've read it, yes?
(03:17:50 PM) mdz: mjg59: have you?
(03:17:58 PM) Keybuk: mdz: re-reading it at the moment
(03:18:03 PM) Keybuk: I've read it at least thrice
(03:18:17 PM) mjg59: I've read it
(03:18:21 PM) mjg59: But have little to add right now
(03:18:51 PM) mdz: sabdfl: having a TB member on MC is not the same as an ack by a TB member; presumably the council would decide by vote or consensus
(03:18:57 PM) Keybuk: I understood the "need TB ack" was just that the TB would retain ownership of the MOTU team, and a TB member had to do the actual LP lifting
(03:19:17 PM) ajmitch: are you going to appoint people to the motu council, or take nominations from MOTUs?
(03:19:34 PM) sabdfl: either way, i think we should keep it lightweight
(03:19:42 PM) sabdfl: encourage them to review member applications by email
(03:19:50 PM) sabdfl: so that they don't block on IRC meetings
(03:20:10 PM) sabdfl: i'm happy with that too if it cuts down on meetings :-)
(03:20:25 PM) sabdfl: a formal report is nice - for the record
(03:20:37 PM) sabdfl: it gives the council two weeks to respond
(03:19:47 PM) mdz: I understood it as: council provides a report and recommendation to TB, TB member signs off on it and grants privileges
(03:20:22 PM) mdz: the proposal calls for a two-week deliberation period for applications to the council
(03:20:47 PM) mdz: I, too, prefer that both the applications and the TB signoff happen via email
(03:21:00 PM) mdz: and reserve interactive interviews for -core-dev
(03:21:34 PM) mdz: I would like to formally incorporate UbuntuDevelopers as guidelines for the council's decision process
(03:21:44 PM) mdz: and have that document maintained by the TB
(03:21:41 PM) sabdfl: sounds good
(03:22:35 PM) mjg59: I think that sounds reasoanble
(03:22:49 PM) mdz: ok, I can communicate that feedback to dholbach and ask him to revise
(03:23:21 PM) mdz: what about nominations for the council?
(03:23:33 PM) sabdfl: dholbach and pitti have been mooted
(03:23:50 PM) pitti: yes, dholbach wanted to have two core-devs at least initially
(03:23:58 PM) sabdfl: would be nice to have more non-canonical presence
(03:24:15 PM) sabdfl: would potentially be a stepping stone for greater non-canonical TB presence
(03:24:56 PM) mdz: I don't think that any of us are involved enough in ongoing MOTU operations to generate that list off the cuff
(03:25:38 PM) mdz: and I think we should certainly talk to any nominees formally before appointing them
(03:25:55 PM) mdz: so perhaps we should solicit nominations from MOTU itself and invite them to attend a TB meeting?
(03:25:59 PM) sabdfl: pitti: could you and dholbach work up some nominations?
(03:26:37 PM) pitti: we should ask some MOTUs
(03:27:14 PM) pitti: sabdfl: I don't track MOTUs that closely, though; dholbach asked me to join to review reports, packages, and guide people
(03:27:22 PM) pitti: but I'm happy to work with Daniel, sure
(03:27:36 PM) sabdfl: ok, let's ask the MOTU's for nominations
(03:27:58 PM) sabdfl: is it reasonable to hope to get this approved and the council in place at the next TB meeting?
(03:28:18 PM) mdz: seems reasonable to me
(03:28:23 PM) ***ajmitch would think it'd take only a few days if it's not done by drawn-out voting
(03:29:34 PM) sabdfl: we want to avoid voting if possible
(03:29:53 PM) sabdfl: caucusing, yes, voting, not if we can help it
(03:30:02 PM) mdz: agreed
(03:30:24 PM) mdz: pitti: since you'll talk to dholbach anyway, can you communicate the revisions we agreed on above as well?
(03:30:39 PM) pitti: mdz: no problem
(03:30:51 PM) mdz: pitti: thanks
(03:31:11 PM) shawarma: ..so ubuntu-dev applications are deferred until the council is formed?
(03:31:40 PM) mdz: shawarma: good question
(03:32:49 PM) mdz: mjg59,sabdfl,Keybuk: ?
(03:32:58 PM) sabdfl: how many folks came specially for that?
(03:33:02 PM) mdz: there is surely a backlog of ubuntu-dev applications
(03:33:03 PM) Keybuk: mdz: I'm happy for us to review them until the council is formed
(03:33:07 PM) shawarma: I did.
(03:33:07 PM) mdz: though it would be a good way to exercise the process
(03:33:12 PM) Keybuk: there's a fair backlog now, and I think its unfair to make them wait longer
(03:33:24 PM) vil: me too
(03:33:31 PM) sabdfl: agreed
(03:33:32 PM) shawarma: I personally wouldn't mind too much. I've only just applied 5 hours ago. :-)
(03:33:48 PM) sabdfl: perhaps we can do a fast cut through the list
(03:33:58 PM) mjg59: Well, can we find out how many people are actually here?
(03:34:06 PM) mdz: vil applied 2006-11-17
(03:34:06 PM) ***shawarma raises his hand.
(03:34:09 PM) sabdfl: folks with (a) really good wiki pages, and (b) lots of LP evidence of contribution get a quick +1?
(03:34:13 PM) mjg59: If there's 20 candidates, it's not going to be practical to get through them all
(03:34:22 PM) mdz: (c) sponsor
(03:34:26 PM) sabdfl: that too
(03:34:39 PM) geser: is it to late to add myself to the list?
(03:34:39 PM) mdz: it looks like we have two in attendance
(03:34:58 PM) mdz: geser: not if you have everything prepared (a, b, c above)
(03:35:21 PM) mdz: vil: are you prepared?
(03:35:27 PM) geser: crimsun: would you vouch for me?
(03:35:34 PM) vil: sabdfl: how much is lots (for LP)?
(03:35:43 PM) vil: mdz: sort of
(03:36:11 PM) mdz: given that sponsored uploads, perhaps the most significant evidence of contribution, aren't tracked in LP...
(03:36:12 PM) crimsun: I'll gladly sponsor geser, having worked with him on numerous syncs and merges
(03:37:01 PM) crimsun: he has been working in -motu to teach prospective MOTU to merge, too
(03:37:13 PM) Keybuk: I've seen a fair amount of geser's work
(03:37:38 PM) sabdfl: let's start with the folks who were on the list
(03:37:44 PM) sabdfl: vil, URL for your wiki page?
(03:37:55 PM) sabdfl: sponsor details?
(03:37:59 PM) sabdfl: launchpad home page URL?
(03:38:58 PM) Keybuk: Dec 05 15:35:38 <doko_> mdz, Keybuk, mpt: Vladmir Lapacek (vil) will be at the T
(03:38:58 PM) Keybuk: B meeting tonight, applying as a developer. I cannot attend the meeting tonight,
(03:38:58 PM) Keybuk: some comments: he is active in java packaging (new packages, and updating packa
(03:38:58 PM) Keybuk: ges) and eclipse packaging (eclipse and eclipse-pydev) over the past nine or ten
(03:38:58 PM) Keybuk: months. I'd appreciate if he becomes developer.
(03:39:59 PM) mdz: can we agree that such testimonials should be sent via email to technical-board@, and not on random IRC channels?

(03:39:38 PM) vil: https://wiki.ubuntu.com/VladimirLapacek
(03:39:40 PM) mdz: oh yes, I saw that briefly but didn't make the association
(03:39:47 PM) vil: https://launchpad.net/people/vil
(03:40:09 PM) vil: yes, I worked mainly with doko
(03:40:45 PM) sabdfl: doko's is a good reference
(03:41:22 PM) mdz: mailed doko accordingly
(03:41:49 PM) mdz: for -dev
(03:42:18 PM) sabdfl: vil: do you think eclipse will be a reasonable proposition for main soon?
(03:43:01 PM) vil: sabdfl: currently eclipse is not so stable to put it there
(03:43:15 PM) vil: sabdfl: but you know, doko makes his best
(03:42:22 PM) mdz: vil: how long have you been working with him?
(03:44:01 PM) vil: mdz: doko contacted me after an upload to REVU some 9 - 10 months ago
(03:44:12 PM) vil: mdz: don't remember exactly
(03:45:19 PM) mdz: vil: is that when you first started to learn packaging?
(03:46:09 PM) vil: mdz: started some 3 months earlier because the first package (eclipse-pydev) took a lot of effort
(03:46:33 PM) vil: mdz: it was not the easiest one to do... lot of patching eclipse / java build
(03:46:21 PM) mdz: vil: do you also work with the MOTU team, or only with doko?
(03:46:55 PM) vil: mdz: i would say rather with doko
(03:47:36 PM) mdz: vil: are you aware of anyone in MOTU working on java-related packaging?
(03:48:17 PM) vil: mdz: not really, though i work with a few people from debian-java
(03:48:34 PM) vil: mdz: this should be pretty close
(03:49:00 PM) vil: mdz: among them are tashiro, man-di
(03:49:14 PM) vil: mdz: again people around eclipse
(03:49:06 PM) mdz: vil: apart from making your own uploads to universe, what would be different about your involvement if you were a member of ubuntu-dev?
(03:50:04 PM) vil: mdz: no sure if this is the right answer, but the uploads would be the biggest change
(03:50:43 PM) mdz: vil: do you think there is the potential to get more people involved in Java within the MOTU team?
(03:51:08 PM) mdz: with Java becoming more open, it would be good for Ubuntu to be a place where interested developers could contribute
(03:51:34 PM) vil: mdz: now with open-sourced sun java, i would guess that more people will be interrested in it
(03:51:39 PM) vil: mdz: so yes!
(03:51:04 PM) LaserJock: I think a MOTU Java team would be appreciated
(03:52:19 PM) mdz: vil: are there other parts of Ubuntu where Java support could be improved, beyond continuing maintenance of Eclipse?
(03:52:21 PM) vil: also sun java packages need a lot of care
(03:52:53 PM) mdz: what do they need specifically?
(03:53:46 PM) vil: mdz: jvm and compiler are already open-sourced and several people are trying to put it together with GNU classpath
(03:54:07 PM) vil: mdz: still waiting for OS sun run-time libraries
(03:54:42 PM) vil: GNU classpath has a devel branch, that is not yet here in Ubuntu
(03:54:59 PM) vil: tons of libraries from sf.net
(03:56:12 PM) mdz: mjg59,Keybuk: questions for vil?
(03:56:39 PM) Keybuk: none from me
(03:56:45 PM) mjg59: Nope
(03:57:37 PM) sabdfl: +1 from me, based on doko's reference and an obvious interest in seeing the java packages well done
(03:57:56 PM) Keybuk: +1 from me on doko's recommendation, and general in-aweness of anyone who understands java
(03:57:59 PM) mdz: +1, likewise
(03:58:07 PM) mjg59: +1 (not that it matters at this point)
(03:58:26 PM) mdz: vil: congratulations
(03:58:27 PM) sabdfl: mjg59: vote sooner :-)
(03:58:30 PM) vil: looks good, thank you all
(03:58:33 PM) ogra: congrats vil
(03:58:34 PM) sabdfl: congrats and welcome aboard
(03:58:37 PM) tsmithe: well done!
(03:58:46 PM) pitti: vil: welcome

(03:58:47 PM) Keybuk: +1 on kylem, +1 on keescook :p
(03:58:51 PM) ***Keybuk speed votes

(03:55:52 PM) BenC: Sorry to interrupt, but if the TB has time to get kylem for upload privs quickly, I'd appreciate it...he has an appoint to leave for soon
(03:58:41 PM) sabdfl: now, who is this kylem character?
(03:58:53 PM) BenC: sabdfl: Some weirdo we picked up at allhands
(03:59:07 PM) BenC: he claims to work for you :)
(03:59:30 PM) BenC: I'll vouch for his burp ability, and toss in a technical recommendation
(04:00:56 PM) mdz: kylem: the usual procedure is that you tell the board a bit about yourself, your work in Ubuntu, and the person who has been sponsoring your uploads speaks on your behalf
(04:01:27 PM) mdz: kylem: Debian experience is also relevant
(04:02:16 PM) kylem: ok, so i'm kyle, i work for canonical on the ubuntu kernel. recently i've been taking care of dapper security and proposed, and edgy security and proposed. i also am on the debian kernel team maintaining hppa, am the upstream guy for parisc-linux, and wrote a couple drivers.
(04:03:01 PM) Keybuk: kylem: how can we improve the process of updating drivers in stable ubuntu releases?
(04:03:16 PM) kylem: Keybuk, we can't. it's a really hard problem.
(04:03:38 PM) kylem: i assume you mean backporting from recent linus kernels, to old linus kernels.
(04:04:09 PM) kylem: there's very little that can be done to make it less painful. things frequently get renamed in the api, functions deprecated and moved around, calling conventions changed.
(04:04:36 PM) Keybuk: is there no useful solution to it?
(04:05:00 PM) pitti: the mysterious thing 'standard driver interface' that many people cry out for?
(04:05:17 PM) kylem: not really, it's a hard problem. what i've done since i've done this for sky2 and tg3 so far, is written a header to make a bit of it less painful that wraps things.
(04:05:27 PM) BenC: I think Keybuk's means more in regards to how we provide them to users, rather than how much work it is for us to backport it
(04:05:27 PM) mjg59: /usr/src/linux/Documentation/stable_api_nonsense.txt
(04:05:29 PM) mjg59: Sadly
(04:05:53 PM) siretart: what problems do occur when compiling kernel packages from feisty in edgy or dapper? this seems to be what ppl are actually doing
(04:06:52 PM) kylem: siretart, old compilers is a possible problem, the need to backport build deps like udev
(04:07:03 PM) ***pitti sees things break when using 2.6.17 on feisty due to udev incompatibilities; that might also affect the other direction
(04:08:00 PM) Keybuk: mjg59: any questions for kylem?
(04:08:16 PM) kylem: Keybuk, in terms of providing drivers to users? i'm not quite sure what you mean, i suppose we could split out drivers and divert the modules if we didn't want to provide whole new kernels for one updated driver.
(04:08:55 PM) mjg59: We have the infrastructure to do so for the installer, so it wouldn't be implausible
(04:10:10 PM) kylem: mjg59, yeah, kind of hard if the driver you're trying to update to support hw is the network though.
(04:10:53 PM) Keybuk: are we ready to vote for kylem?
(04:11:05 PM) BenC: you guys don't want to hear from me?!
(04:11:19 PM) Keybuk: BenC: sure :)
(04:11:24 PM) BenC: I had a whole speech prepared
(04:11:52 PM) BenC: well, over the past two weeks, kylem has done work with dapper/edgy, in regards to security and proposed updates
(04:12:23 PM) BenC: eh, probably pointless for me to talk anyway :)
(04:12:51 PM) BenC: but all-in-all, I think he's a good candidate, has done a fine job reviewing patches and assuring things compile
(04:07:13 PM) sabdfl: oh, alright, the t-shirt gets me. +1
(04:11:52 PM) mjg59: I've done enough work with Kyle already to +1 him from personal experience
(04:13:22 PM) Keybuk: that's +1s from everyone spread out there
(04:13:24 PM) Keybuk: kylem: congrats
(04:13:30 PM) kylem: neat.
(04:13:51 PM) sabdfl: ok, well done kylem
(04:13:59 PM) pitti: kylem: welcome!
(04:14:02 PM) kylem: pitti, thanks :)

(04:13:50 PM) Keybuk: keescook: you're next; please introduce yourself, and let us know what you'll be doing in Ubuntu
(04:13:58 PM) keescook: Hi! I'm Kees Cook. I've been doing security updates to main packages in all the stable releases, as well as having multiple uploads sponsored into the development releases. Other core-devs that have done the sponsoring include crimsun, pitti, and seb128. I'd love to be able to upload to feisty/main directly, as it makes the merges go much faster. :)
(04:14:03 PM) keescook: LP: https://launchpad.net/people/keescook Wiki: https://wiki.ubuntu.com/KeesCook
(04:14:07 PM) sabdfl: BenC, for future ref it would be good to have canonical folks prepped with wiki pages etc
(04:14:11 PM) sabdfl: thanks keescook :-)
(04:14:23 PM) keescook: my pleasure! :)
(04:14:26 PM) sabdfl: MythTV ROCKS ;-)
(04:14:27 PM) ***pitti waves the fanboy flags for Kees
(04:14:31 PM) ajmitch: keescook: you forgot me! ;)
(04:14:56 PM) keescook: ajmitch: sorry! My listed sucked. :) ajmitch too. :)
(04:14:58 PM) pitti: Kees has done an awesome job on security and other distro-related work so far, has learned stuff at an incredible pace
(04:15:08 PM) pitti: and he is very cautious when it comes to new things
(04:15:14 PM) ajmitch: keescook: and I _was_ here all ready to cheer for you :)
(04:15:15 PM) pitti: and asks really good questiosn
(04:15:20 PM) ***keescook hugs ajmitch
(04:15:28 PM) Keybuk: keescook: for the sake of disclosure, you're an employee of Canonical, but not a member of the distro team?
(04:16:17 PM) keescook: Keybuk: correct, I'm part of the IS dept, but still doing some amount of merge work (for security updates) and for packages in main I have a familiarity with (inkscape)
(04:16:34 PM) shawarma: IS?
(04:16:51 PM) keescook: shawarma: information systems, or IT, infrastructure, etc.
(04:16:52 PM) pitti: shawarma: Information systems, the guys that care for hw infrastructure
(04:17:15 PM) Keybuk: keescook: is this so elmo gets a sneak preview of incoming kernel holes? :p
(04:17:34 PM) keescook: heh. only if the embargo is over (sorry elmo)
(04:17:35 PM) ***pitti would like to see keescook more involved with distro development, too
(04:18:17 PM) sabdfl: keescook: what sort of automated security checking do you think we could do in ubuntu?
(04:18:29 PM) sabdfl: what's the roadmap for hardening the distro, do you think?
(04:18:36 PM) sabdfl: and what do you think about turning avahi on by default?
(04:18:55 PM) keescook: sabdfl: I think a certain level of automatic code auditing is possible, but only catches the really obvious stuff
(04:19:25 PM) keescook: as far as distro hardening, I think we're on the right track.
(04:19:23 PM) pitti: https://blueprints.launchpad.net/distros/ubuntu/+spec/memory-protection, btw (didn't get discussion in Mountain View, but kees and I discussed it a bit)
(04:19:54 PM) keescook: that spec hasn't been drafted as it was never officially scheduled, but I have some extensive notes
(04:20:08 PM) sabdfl: pitti: needs a drafting session, though :-)
(04:20:10 PM) keescook: we talked to doko about the toolchain changes needed, though
(04:20:59 PM) keescook: Our stack protection is already proving very nice; I've had at least one vuln I tested that just refused to work in edgy.
(04:21:32 PM) sabdfl: that's a win
(04:22:00 PM) sabdfl: what about network service exposure, what's your opinion about having the distro listen on some ports by default, as opposed to nothing-by-default?
(04:22:00 PM) keescook: Anything to make the window of opportunity smaller, and the executable relocation stuff (PIE) is going to be another big win.
(04:22:57 PM) keescook: I'm pretty conservative about ports. I think the newly spec'd port policy is solid.
(04:23:09 PM) keescook: pitti and I have also been working on some automated tests for updated packages
(04:23:27 PM) keescook: testing the updates tends to take the longest amount of time, so any amount of automated testing helps
(04:23:53 PM) keescook: most packages don't have any kind of test suite, so we've been working with iwj and lifeless to put together some of our own test suites.
(04:23:53 PM) pitti: these should help with catching regressions of old security fixes (but of course not for new vulns)
(04:24:42 PM) keescook: lately "fuzzers" have proved a useful tool for doing some blockbox testing. (See the "month of kernel bugs", for example)
(04:24:44 PM) sabdfl: where is the newly spec'd port policy?
(04:25:03 PM) pitti: https://wiki.ubuntu.com/DefaultNetworkServices
(04:26:03 PM) keescook: also, the addition of apport has allowed for much more detailed crash reports
(04:26:23 PM) keescook: the recent koffice update I did was a result of an audit near a reported crash, which had security implications.
(04:27:31 PM) sabdfl: interesting
(04:27:37 PM) sabdfl: that's a good success story, pitti
(04:27:55 PM) pitti: apport FTW \o/ :)
(04:28:18 PM) sabdfl: ok, +1 from me on the back of existing work done in security, and commitment to ongoing contribution
(04:28:33 PM) Keybuk: +1 from me
(04:28:34 PM) Keybuk: mjg59: ?
(04:28:40 PM) pitti: (just for the record, Kees can already upload to main for stable-security)
(04:28:50 PM) sabdfl: fwiw i found kees to be a great contributor to discussions i've participated in with him
(04:29:23 PM) mjg59: +1 also
(04:29:36 PM) mjg59: Great deal of personal recommendations
(04:29:47 PM) ***pitti hugs keescook and congratulates him for his shiny new core-dev badge
(04:29:50 PM) keescook: yay! thanks guys. :)
(04:29:51 PM) ajmitch: congrats, keescook
(04:29:56 PM) sabdfl: well done keescook
(04:29:59 PM) sabdfl: and welcome aboard
(04:30:00 PM) ogra: congrats keescook
(04:30:06 PM) ***keescook hugs everyone

(04:30:34 PM) Keybuk: ok, that's the core-dev candidates
(04:30:29 PM) sabdfl: mjg59: did you get much of a look into https://wiki.ubuntu.com/DefaultNetworkServices before it was approved?
(04:30:52 PM) mjg59: sabdfl: Nope
(04:31:12 PM) sabdfl: mjg59: i think it's something that should be discussed in this forum
(04:31:18 PM) ***pitti agrees
(04:31:34 PM) pitti: at UDS I made sure to get it past many people
(04:31:46 PM) pitti: but the more review, the better
(04:31:47 PM) sabdfl: i think it's very well written
(04:31:53 PM) sabdfl: thank you for that pitti
(04:32:06 PM) sabdfl: i know mdz is a big supporter, hopefully he will get back shortly to comment
(04:32:17 PM) sabdfl: i'm a little sceptical of two things
(04:32:34 PM) sabdfl: first, how it is supposed to be possible to detect network services without actually talking to them
(04:32:45 PM) pitti: sabdfl: they talk to you
(04:33:00 PM) ogra: thats what avahi is for
(04:33:02 PM) pitti: i. e. avahi broadcasting info about new IPs and such
(04:33:12 PM) sabdfl: pitti: inevitably, that means listening, and beng vulnerable to an attack
(04:33:13 PM) pitti: but of course there is a certain amount of communication
(04:33:16 PM) sabdfl: and second, how we can realistically promise always to separate local, configured services from detected ones
(04:33:21 PM) pitti: sabdfl: right, that was our major concern
(04:33:35 PM) sabdfl: that seems to be a commitment to diverge from upstream if they don't provide that separation
(04:33:57 PM) pitti: sabdfl: either a permanent diversion, or we refuse to enable it by default
(04:34:21 PM) pitti: we specifically reserved the right to say 'no, we don't feel good about that thing'
(04:34:29 PM) pitti: sabdfl: one example are remote printers
(04:34:39 PM) pitti: the current print dialog doesn't give you a clear separation
(04:34:49 PM) pitti: until that's settled, I won't enable browsing by default
(04:36:06 PM) pitti: sabdfl: btw, for your first case, getting a DHCP address is no different from zeroconf network layout communication
(04:36:22 PM) Keybuk: (or, indeed, getting a reply to a DNS query)
(04:36:26 PM) keescook: (almost more dangerous, actually)
(04:37:04 PM) sabdfl: true
(04:37:09 PM) sabdfl: true, true, true!
(04:37:20 PM) pitti: the other thing is, people do enable that shi^Wstuff
(04:37:29 PM) pitti: so we should rather concentrate on making it reasonably safe in that mode
(04:37:30 PM) sabdfl: i just have this thing about slippery slopes :-)
(04:37:39 PM) pitti: sabdfl: and rightly so!
(04:38:11 PM) pitti: I must say I have always been proud of our (alleged) no open ports policy
(04:38:18 PM) pitti: well, 'almost, but not quite' (dhcp, dns)
(04:38:26 PM) ogra: ++
(04:38:46 PM) sabdfl: me too
(04:38:50 PM) pitti: but communication on the IP level about IP addresses and host names is inherently untrusted
(04:39:17 PM) pitti: and we never expose information on the app level (like automatic sharing of music, etc.)
(04:39:27 PM) pitti: ... by default, of course
(04:38:53 PM) sabdfl: mjg59: thoughts?
(04:39:42 PM) mjg59: I have some qualms about the ease of implementation, and I'd like to have more of an idea about what the UI for this stuff is going to look like
(04:39:47 PM) mjg59: But overall, the spec seems sound
(04:39:57 PM) sabdfl: is the "listening" code very, very well audited?
(04:40:10 PM) pitti: the threat right now is: people broadcast a malicious mp3 which is automatically picked up and peole play it back without checking, and it triggers an exploit of the mp3 codec
(04:40:12 PM) sabdfl: is it written in a language like Python, where there's less likelihood of some sort of overflow?
(04:40:29 PM) sabdfl: that's no worse than linking to it on the web
(04:40:46 PM) pitti: sabdfl: right, or sending by mail (and it's only the local network)
(04:40:53 PM) pitti: sabdfl: avahi in this case is C
(04:41:14 PM) pitti: but it confines its privileges in a good way
(04:41:19 PM) pitti: non-root, chrooted, etc.
(04:41:30 PM) sabdfl: ok
(04:41:34 PM) pitti: i. e. maximum impact is injection of false network information
(04:41:46 PM) pitti: but people can do that anyway by just sending you bogus stuff from their machines
(04:41:55 PM) sabdfl: bring on the day
(04:41:56 PM) pitti: i. e. we have to treat this information as untrustable anyway
(04:42:53 PM) sabdfl: the spec does not say which network services will be available by default
(04:42:59 PM) sabdfl: it talks about listening by default
(04:43:06 PM) sabdfl: and exposing mac, ip, hostname etc
(04:43:15 PM) sabdfl: and "availability of network-facing services"
(04:43:16 PM) Keybuk: that was deliberate, other specs cover that
(04:43:28 PM) Keybuk: e.g. the libnss one (zero configuration networking?)
(04:43:37 PM) pitti: right
(04:43:40 PM) sabdfl: i would like the TB to have to ack new open ports explicitly
(04:43:49 PM) sabdfl: dunno how others feel
(04:43:50 PM) pitti: it was meant to be a policy document, no tech changes
(04:44:08 PM) pitti: sabdfl: I randomly picked ubuntu security team for now, happy to change to TB
(04:44:17 PM) pitti: sabdfl: or both even
(04:44:44 PM) sabdfl: what's the plan for feisty, w.r.t. ports listening by default?
(04:44:56 PM) pitti: sabdfl: we already have libnss-mdns for resolving .local host names
(04:45:08 PM) pitti: sabdfl: and avahi on by default for automatic service discovery
(04:45:16 PM) sabdfl: that listens on an IP port? not just localhost?
(04:45:26 PM) pitti: sabdfl: cups browsing depends on whether I or anyone else manages to get the printer dialog right
(04:46:08 PM) pitti: sabdfl: right, avahi opens an UDP port, much like DNS or dhcp
(04:46:24 PM) pitti: no TCP ports so far, and none planned
(04:46:44 PM) sabdfl: but dns and dhcp are listen-for-responses-to-sent-messages only
(04:46:55 PM) sabdfl: as opposed to listen-to-broadcast-traffic
(04:46:49 PM) pitti: I don't have any plans for any other default services for now
(04:47:02 PM) sabdfl: ah well
(04:47:06 PM) Keybuk: sabdfl: for UDP, the distinction is artificial
(04:47:21 PM) Keybuk: as to get responses, you have to listen to any traffic you receive, and filter it to decide what's actually a response
(04:47:30 PM) Keybuk: (as apposed to TCP where the filtering is down at the network layer)
(04:47:35 PM) sabdfl: ok
(04:47:56 PM) sabdfl: will the server edition have this stuff on or off by default?
(04:48:12 PM) pitti: sabdfl: avahi and nss-mdns make little sense for servers IMHO
(04:48:18 PM) pitti: thus I'd vote for not install them by default
(04:48:19 PM) Keybuk: server edition hasn't even got this stuff installed :p
(04:48:25 PM) pitti: this stuff is mainly for desktop bling
(04:48:34 PM) pitti: automatic pr0n sharing is less interesting on servers
(04:48:42 PM) pitti: and usually people want to be in full control of their IPs and such
(04:49:12 PM) sabdfl: ok, can that be made explicit in the policy? server edition is nyet nyet nyet?
(04:50:29 PM) pitti: sabdfl: I'll change 'security team' to 'TB && security team' for signoff then, too?
(04:50:35 PM) sabdfl: sure
(04:50:50 PM) sabdfl: any further commentary? mdz, want to ack you are happy with this?
(04:53:12 PM) mdz: sabdfl: I reviewed the spec at UDS and was satisfied with it, yes
(04:53:35 PM) mdz: how did we get onto a discussion about network services anyway?
(04:53:50 PM) mdz: there are two discussion items remaining on the wiki agenda
(04:53:51 PM) pitti: mdz: over the approval of keescook and the discussion for it
(04:53:53 PM) sabdfl: mdz: keescook applied
(04:53:58 PM) sabdfl: asked his opinion
(04:54:06 PM) mdz: I saw keescook and kylem were approved, both have my ack as well
(04:54:14 PM) sabdfl: turns out the TB hadn't been consulted on this significant change in policy ;-)
(04:54:37 PM) Keybuk: sabdfl: the TB approved it (me and mdz, at least)

(04:51:50 PM) sabdfl: Keybuk: what's next?
(04:52:15 PM) Keybuk: PPC
(04:52:21 PM) shawarma: There's stilla a few ubuntu-devs left..

(04:52:55 PM) ryanakca [n=ryan@unaffiliated/ryanakca] entered the room.

(04:53:20 PM) Keybuk: shawarma: I don't see you in the list
(04:53:23 PM) ajmitch: ogra: yes, I was surprised to hear he wasn't :)
(04:55:15 PM) shawarma: Keybuk: My name is Soren Hansen. I should be there.

(04:55:58 PM) sabdfl: ok, shawarma, you're up
(04:55:41 PM) shawarma: https://launchpad.net/people/shawarma
(04:55:50 PM) shawarma: https://wiki.ubuntu.com/SorenHansen
(04:56:20 PM) shawarma: This turned out a bit longer than anticipated:
(04:56:25 PM) shawarma: The personal stuff: I'm 25, male, soon to be married, mathematics/computer science student/FLOSS consultant living in Denmark. I work well both alone and in teams. People tend to laugh when they're around me, and I tell myself that they're laughing with me rather than at me, but who's to tell? I've worked as a programmer and systems administrator in a Linux environment for 6 years.
(04:56:31 PM) shawarma: Since I was 4, I've been pulling things apart to figure out how or why they work. I'm also the kind of person that actually reads a manual, rfc or other type of specification from cover to cover to fully understand how things are *supposed* to be working. My work on Debian and Ubuntu is no different. If I stumble upon a bug I can get really stubborn in trying to figure out what's causing it and trying to fix it. Also, I've of course read the Debian
(04:56:40 PM) shawarma: Stuff, I'd like to work on in Ubuntu: I've as of yet not settled on a particular niche that I want to fill in. I currently study mathematics at a university so math related stuff like TeXlive and Axiom gets more attention than certain other bits. I'll be going back to studying computer science next summer and at that point I'll probably be more focused on stuff like exotic programming languages. Meanwhile, I work as sysadmin (on a consultancy basi
(04:57:18 PM) shawarma: LaserJock did most of my sponsoring.
(04:57:27 PM) ogra: we all met shawarma in paris
(04:57:54 PM) sabdfl: looks like lots of LP interaction
(04:58:06 PM) sabdfl: I think LaserJock wrote in support of your application, to the TB
(04:58:41 PM) Keybuk: shawarma: I notice that there's a fair gap in your work (basically November)
(04:58:52 PM) Keybuk: and again for most of August and September
(04:59:15 PM) Keybuk: is this a side-effect of studenthood? (I don't know the danish academic calendar)
(04:59:30 PM) shawarma: Keybuk: That's right. I've been tied up in other things. Also, it was very tedious to get anything uploaded during november.
(05:00:17 PM) shawarma: Keybuk: Among other things, yes.
(05:00:56 PM) shawarma: Keybuk: I don't have a whole lot of time available, but the little I do have, I'd like to be able to spend efficiently on Ubuntu.
(05:01:16 PM) LaserJock: shawarma did a lot of good work during Dapper especially
(05:01:18 PM) shawarma: Keybuk: Having upload privs is a bit catalyst for that.
(05:01:29 PM) ogra: shawarma, could you imagine helping out with edubuntu app selection for higher math grades ?
(05:02:06 PM) shawarma: ogra: Honestly, no. :-)
(05:02:23 PM) LaserJock: ogra: don't worry I'll rope him into it ;-)
(05:02:25 PM) shawarma: ogra: sorry. :-)
(05:02:58 PM) sabdfl: what's an MMSC?
(05:03:15 PM) shawarma: sabdfl: It's the software that enables mobile phones to send porn to other phones.
(05:03:29 PM) shawarma: sabdfl: Well, and other pictures, but it's probably mostly porn.
(05:03:36 PM) shawarma: Everyknow knows that.
(05:03:40 PM) shawarma: Everyone, even.
(05:03:43 PM) sabdfl: what else would a camera phone be for?
(05:03:51 PM) shawarma: exactly.
(05:04:20 PM) sabdfl: shawarma: can you tell me what would be the thing you'd most like to change about the proceses for new developers in ubuntu?
(05:05:03 PM) shawarma: sabdfl: Well, the mentoring process is a really good idea. Other than that, I think Ubuntu is a really good place for new people. Lots of encouragement all around.
(05:05:45 PM) shawarma: sabdfl: Also the MOTU school is a really good idea. I can't really think of anything concrete, I'd like to change.
(05:04:52 PM) mdz: shawarma: who has been sponsoring your uploads? LaserJock?
(05:05:09 PM) shawarma: mdz: Mostly, yes.
(05:05:50 PM) mdz: LaserJock: anything other than merges?
(05:06:05 PM) LaserJock: I did a review of rawstudio I believe
(05:06:09 PM) LaserJock: on REVU
(05:06:45 PM) LaserJock: most recent stuff was merge and syncs
(05:06:32 PM) shawarma: Right. libmms was mine, as well. I think it got included back in breezy.
(05:06:48 PM) mdz: shawarma: is rawstudio something you packaged from scratch?
(05:06:52 PM) shawarma: mdz: Yes.
(05:07:14 PM) shawarma: mdz: I've done a lot of that, actually, for a company I used to work for. Internal, proprietary stuff, but still.
(05:07:11 PM) mdz: shawarma: was there anything unusual or challenging about packaging that particular app?
(05:07:19 PM) shawarma: mdz: Not at all.
(05:07:32 PM) shawarma: mdz: I know upstream, so I told them how to make my job easy.
(05:08:07 PM) shawarma: It uses cdbs, so the debian/rules is only like two lines or so. :-)
(05:08:45 PM) shawarma: I am pretty familiar with the build system though. Also when it comes to exotic packages. Axiom, for instance, was a nightmare.
(05:08:52 PM) sabdfl: +1 from me on the back of a long track record of contribution, lots of bug interaction, and good packaging experience with refs
(05:09:32 PM) mdz: +1
(05:09:36 PM) mjg59: If we're voting, +1 from me (for the same reasons)
(05:09:40 PM) Keybuk: +1
(05:09:52 PM) mdz: shawarma: congratulations and good luck
(05:09:59 PM) shawarma: ogra: Yes, that was a good one.
(05:10:02 PM) ogra: congrats shawarma
(05:10:11 PM) ajmitch: shawarma: congrats & welcome :)
(05:10:15 PM) shawarma: sabdfl, mdz, mjg59, Keybuk: Thanks, guys!
(05:10:25 PM) LaserJock: shawarma: congrats
(05:10:35 PM) shawarma: Keybuk: *G*

(05:11:48 PM) Keybuk: geser: you're up; introduce yourself, etc.
(05:11:58 PM) geser: wiki: https://wiki.ubuntu.com/MichaelBienia
(05:11:58 PM) geser: LP: https://launchpad.net/people/geser
(05:12:06 PM) geser: My name is Michael Bienia. I'm 27 years old and live in Dortmund, Germany.
(05:12:11 PM) geser: I'm using Ubuntu since the development cycle of breezy but only started in august 2006 to also contribute to Ubuntu.
(05:12:17 PM) geser: Since than I was fixing packages with unmet deps and doing sync and merges for packages in universe. crimsun sponsored my uploads and also reviewed my sync requests.
(05:12:40 PM) Keybuk: universe is falling way behind on merges at the moment -- in fact, it's making hardly any headway
(05:12:42 PM) geser: I'm also helping in #ubuntu-motu as far as I can
(05:12:52 PM) Keybuk: what do you see as the problem there?
(05:13:40 PM) geser: I also try to get through the merges list for universe
(05:14:53 PM) pitti: one problem is certainly the feedback of changes to Debian, to make packages syncable
(05:15:08 PM) sabdfl: geser: that's an impressive package maintainance list. thank you for a big contribution already!
(05:15:45 PM) LaserJock: yes, the geser-crimsun team account for probably a majority of MOTU work right now
(05:16:15 PM) Keybuk: indeed
(05:16:22 PM) Keybuk: a good percentage of the sync requests I process come from geser
(05:16:43 PM) mdz: Keybuk: all well-formed?
(05:16:44 PM) shawarma: <crimsun> I'll gladly sponsor geser, having worked with him on numerous syncs and merges
(05:16:47 PM) shawarma: <crimsun> he has been working in -motu to teach prospective MOTU to merge, too
(05:17:03 PM) Keybuk: mdz: yes
(05:17:27 PM) geser: mdz: a list of my filed sync requests [http://tinyurl.com/yanzaq]
(05:17:27 PM) mdz: geser: which tools and resources do you use to find unmet dependencies?
(05:17:54 PM) geser: mostly the output of apt-cache unmet -i
(05:19:49 PM) sabdfl: geser: from a community perspective, what can we do to improve debian acceptance of ubuntu changes?
(05:22:28 PM) ogra: crimsun, some extra words about geser ?
(05:22:46 PM) sabdfl: geser, did you see my Q re debian / ubuntu maintainer interactions?
(05:23:07 PM) geser: sabdfl: yes, I've seen it
(05:24:33 PM) sabdfl: geser, do you want to duck the Q, or are you framing a response?
(05:24:53 PM) geser: I've currently no idea how to improve it
(05:25:01 PM) sabdfl: ok
(05:25:22 PM) crimsun: geser: beyond even making a more concerted effort to push patches to BTS as appropriate?
(05:25:43 PM) LaserJock: crimsun: yeah, I think BTS is the key
(05:25:46 PM) sabdfl: crimsun: i believe we automatically mail all changes now
(05:26:00 PM) elmo: sabdfl: to the PTS, not the BTS
(05:26:10 PM) LaserJock: sabdfl: opening bugs seems to work much better then putting it on the PTS
(05:27:36 PM) pitti: filing Debian bugs with patches generally works well for me at least, and I could request syncs for many packages due to that
(05:25:26 PM) sabdfl: well, +1 from me on the back of a LOT of sync and merge work
(05:25:43 PM) Keybuk: +1 from me
(05:25:58 PM) mdz: +1 based on development contributions and crimsun's testimonial
(05:27:31 PM) ogra: geser, hey, congrats !
(05:27:31 PM) geser: sabdfl, Keybuk, mdz: thanks
(05:27:33 PM) sabdfl: that's quorum - mjg59?
(05:27:59 PM) mjg59: +1
(05:28:10 PM) sabdfl: ok welcome aboard geser
(05:28:15 PM) pitti: geser: welcome
(05:28:21 PM) geser: thanks to all

(05:28:12 PM) sabdfl: PPC!
--(03:31:02 PM) mdz: PowerPC background is at https://wiki.ubuntu.com/PowerPCReview--
(05:28:21 PM) Keybuk: can I suggest we table PPC for another time
(05:28:26 PM) Keybuk: this meeting is already two and a half hours long
(05:28:23 PM) sabdfl: floors open, its a very well commented spec
(05:28:42 PM) mdz: I made my position clear at UDS and prior (proposed the spec in the first place)
(05:28:42 PM) sabdfl: Keybuk: can we give it 10? i suspect it may be fast.
(05:29:13 PM) mdz: sabdfl: any notable news?
(05:29:38 PM) sabdfl: none that would change the community decision here, though we should frame the result carefully so as not to burn any bridges
(05:29:42 PM) sabdfl: big, blue bridges
(05:29:56 PM) ***ogra is voting against it, simply because the majority of US classrooms is equipped with iMacs
(05:30:00 PM) sabdfl: timing isn't great, there is work afoot, etc etc but we have to make our call
(05:30:20 PM) ogra: would be a big loss for edubuntu
(05:30:20 PM) mdz: ogra: having spent a lot of time in US classrooms, that doesn't match my observations
(05:30:42 PM) mjg59: My understanding is that PS3 will work with a pretty standard PPC userland
(05:30:54 PM) mjg59: So the /potential/ installed base is going to get much bigger very soon
(05:31:00 PM) elmo: mjg59: it only has 256Mb of memory
(05:31:24 PM) mjg59: elmo: It's certainly not an ideal machine, but yes
(05:31:11 PM) Keybuk: mjg59: fsvo standard, but yes
(05:31:47 PM) mc44: mjg59: assuming they actually find some blue diodes...
(05:32:07 PM) sabdfl: PS3 is really interesting
(05:32:27 PM) sabdfl: i might, or might not, have my feet on a PS3 development box right now :-)
(05:32:28 PM) mdz: why any more interesting than, say, Xbox?
(05:32:43 PM) elmo: mdz: sony are actively supporting running Linux on it
(05:32:48 PM) mjg59: mdz: xbox was never plausible
(05:32:51 PM) elmo: without the need for e.g. hardware mods or hacks
(05:33:00 PM) mjg59: It requires hardware modifications to boot unsigned code
(05:33:05 PM) Keybuk: ibm are actively filing bugs in LP about problems with Ubuntu on the PS3
(05:33:07 PM) Keybuk: which is nice
(05:33:07 PM) mjg59: Ditto ps2
(05:31:00 PM) ogra: well from there i get the majority of Mac requests and questions, and according to our educators based there its true nowadays
(05:31:07 PM) LaserJock: my lab has gone all iMac, half of them g5s half intel
(05:31:19 PM) mdz: ogra: unless those schools are attempting to purchase support contracts, they're not facing a big loss in the reclassification of PowerPC as community-supported
(05:32:32 PM) ogra: mdz, well, the community wont build official CDs
(05:33:16 PM) mdz: ogra: how does the word 'official' change the reality of that situation at all?
(05:33:43 PM) elmo: ogra: you realise we build CDs for architectures on ports.u.c, right?
(05:33:48 PM) mdz: the isos aren't mirrored as widely, but in reality I don't think that presents much of a problem
(05:33:59 PM) ogra: well, if we turn away from ppc, i dont think i'll demote time into ppc testing first place anymore ...
(05:34:17 PM) mdz: download figures show there's very little demand for them, as documented in the proposal
(05:35:01 PM) mdz: it's important to note, both here and in any announcement, that we're talking about a reclassification of the port, not dropping it
(05:35:06 PM) sabdfl: it's not hugely expensive for us to support
(05:35:12 PM) ogra: iMacs make also very nice thin clients ...
(05:35:23 PM) pitti: we have a lot of ppc specific problems ATM, and few people have the time to look into them and fix them
(05:35:26 PM) mjg59: mdz: Have any mirror operators complained about the size of the archive?
(05:35:27 PM) sabdfl: if sufficient community interest means the packages all build and ISO's work and get tested, the impact is minimal
(05:35:31 PM) elmo: mjg59: yes
(05:35:36 PM) mdz: mjg59: ->elmo, but yes, is my understanding
(05:35:44 PM) mjg59: elmo: Would the reclassification of PPC be sufficient to resolve this?
(05:35:48 PM) elmo: lots. the archive is 200Gb
(05:36:15 PM) elmo: mjg59: that and removal of older distros would help alleviate the problem. you're never going to resolve it entirely
(05:36:29 PM) Mithrandir: sabdfl: experience shows that pre-release community testing is way, way smaller than what we'd like it to be. Ports even less.
(05:37:02 PM) elmo: Mithrandir: to be fair none of the current port architectures HAVE a community to speak of
(05:37:13 PM) elmo: I think we hire or have hired every existing hppa user in the world ever.
(05:37:13 PM) mjg59: elmo: Well, it's arguable about whether PPC does
(05:37:23 PM) ogra: i'd rather think the ppc community has more endusers than devs by design ...
(05:37:29 PM) mjg59: At least, in terms of the development community
(05:37:35 PM) sabdfl: we may need to have some official arch's that are mirrored only by explicit decision of the mirror
(05:37:43 PM) sabdfl: sparc, for example
(05:37:41 PM) Keybuk: elmo: not to mention the entire Herd community
(05:37:43 PM) mjg59: Even Debian appears to have difficulty obtaining sufficient PPC input
(05:38:23 PM) elmo: well as a first step we need to stop sending distros we EOLed 6 months ago to mirrors, but yes
(05:38:46 PM) mjg59: I think the mirror space argument is effectively separate from the support argument
(05:39:02 PM) elmo: mjg59: I heard that there was some problems with personality conflicts and debian powerpc developers? something to do with d-i? not sure of the details, tho
(05:39:07 PM) mjg59: Changing the status of the architecture would be one way of forcing a resolution of that issue, but we could achieve the same without doing so
(05:39:45 PM) mjg59: elmo: I don't think there's been any real movement on finding a replacement for the previous porter
(05:39:51 PM) sabdfl: main concerns for me are: live-cd / framebuffer / X / ubiquity / boot manager issues complicating Colin's life
(05:39:59 PM) mjg59: Anyway, I think we should ignore the mirror space argument
(05:40:13 PM) sabdfl: - risk of apps FTBFS slowing down adoption of new upstream goodness
(05:40:23 PM) mdz: Even the "petition" in the wiki has few names on it, and most of them seem to be reacting to a belief that the port will go away and/or they won't be able to use it on their systems anymore
(05:40:30 PM) Mithrandir: sabdfl: casper doesn't really have any extra code for supporting ppc so it's not a big problem there at least.
(05:41:21 PM) Keybuk: perhaps we should describe it as a "universe architecture" ? or similar
(05:41:31 PM) Keybuk: make it more obvious that it'll be supported by the community
(05:41:33 PM) ogra: ltsp has, and we are currently the only distro supporting ppc-ltsp at all ...
(05:41:40 PM) sabdfl: ok, i promised keybuk we would keep it tight
(05:41:44 PM) Keybuk: and it'll only really affect people with conanical support contracts
(05:41:56 PM) mjg59: I don't think this conversation is doing a good job of quantifying the costs
(05:42:17 PM) mjg59: Nothing that anyone has said so far has helped me form an opinion
(05:42:25 PM) mdz: * FredericBages -- France -- Ubuntu is the best OS for my PCs and iBook G3. As I try to keep the same OS on my PCs and laptop I would have to change my PCs OS if ubuntu becomes unsupported on my iBook :-( Would it be at least possible to keep the "stable" release supported ?
(05:42:36 PM) Keybuk: I don't really have an opinion here either
(05:42:45 PM) mdz: perhaps we haven't done a very good job at presenting the nature of the decision here
(05:42:49 PM) Keybuk: Ubuntu has never fully worked on my PowerBook
(05:43:16 PM) mjg59: I don't think the small userbase is an issue. It's not (in itself) a cost.
(05:43:34 PM) sabdfl: testing at crunch time is
(05:43:35 PM) mdz: mjg59: my reasoning is that PowerPC has many fixed resource costs, e.g. testing of CD images, relative to its miniscule user base
(05:43:36 PM) mjg59: The costs of maintaining PPC support are, as far as I can see:
(05:43:39 PM) mjg59: a) testing
(05:43:47 PM) mjg59: b) fixing PPC-specific bugs
(05:43:52 PM) mdz: PowerPC costs us as much to validate as i386, but provides much less value to the community
(05:44:04 PM) mjg59: But I see absolutely no numbers that tell me how much developer time is currently spent on these things
(05:44:31 PM) ogra: during release testing 1/3 of the time ?
(05:44:33 PM) Mithrandir: mjg59: do you want back-of-the-envelope calculations or proper numbers?
(05:44:39 PM) mjg59: So I'm faced with some nebulous downsides to keeping it, along with nebulous positives
(05:44:42 PM) mdz: mjg59: I'm sure you realize the complexity of assigning numbers in that way
(05:44:53 PM) mdz: mjg59: call it 25% of the release testing effort
(05:44:57 PM) mjg59: mdz: Absolutely. But right now, I have absolutely no clue at all.
(05:45:06 PM) mdz: somewhat less than 25% of mirror space
(05:45:18 PM) mjg59: As I said before, I think the mirror space argument is a side issue
(05:45:25 PM) mjg59: That can be resolved in any case
(05:45:41 PM) mdz: I don't agree; the mirror will always be that much larger on account of powerpc
(05:45:59 PM) mjg59: Splitting the archive can be done without changing the status of the port
(05:46:02 PM) mdz: there's no imminent mirror crisis as a result of powerpc, but it's a cost we carry
(05:46:04 PM) pitti: mjg59: in edgy I might have spent 10 hours total on ppc related issues, and I think I might even be a developer who cares relatively much about ppc; our bug fixing efforts are way too little to sustain a good quality of the port IMHO
(05:46:17 PM) mjg59: Ok, that's a good point
(05:46:30 PM) mjg59: Does the relatively poor quality of the PPC port damage our reputation?
(05:46:45 PM) mdz: I don't think it has enough users to have any such effect
(05:46:51 PM) pitti: well, it's still the best Linux distro you can get, I thihnk
(05:47:25 PM) pitti: it just doesn't keep up with the quality of i386
(05:47:32 PM) ogra: right, i think its a prestige question as well ... it wont damage our reputation, but keep it up if we continue supporting it
(05:48:02 PM) mdz: we have a number of developers who work on powerpc, but their PowerBooks are dying one by one and can't be replaced
(05:48:10 PM) mdz: they will be replaced with x86 systems
(05:48:11 PM) Keybuk: the numbers game interests me
(05:48:18 PM) Keybuk: I haven't seen a Linux PPC user who isn't using Ubuntu
(05:48:28 PM) ogra: right, but as long as they are there we could go on ...
(05:48:31 PM) Keybuk: do PPCs really not last that long?
(05:48:35 PM) mdz: Colin couldn't even find a replacement power supply for his in SF
(05:48:40 PM) ogra: heh
(05:48:50 PM) ogra: he could have bought one on the vending machine
(05:48:52 PM) Mithrandir: Keybuk: my ppc runs Debian.
(05:48:56 PM) mjg59: There's a spare sitting on my desk at work. Absolutely no idea who owns it...
(05:48:56 PM) pitti: I'll continue using mine as long as it works well
(05:48:56 PM) ogra: he just didnt want to
(05:48:58 PM) elmo: Keybuk: the laptops are terrible, reliability wise, IME
(05:49:10 PM) mjg59: But anyway
(05:49:18 PM) mdz: Keybuk: I've seen as many dead ones as live ones in the past 6 months
(05:49:32 PM) mdz: if I count elmo's, which played dead
(05:49:36 PM) mjg59: Right now, the spec doesn't sway me either way
(05:49:40 PM) pitti: but the spec should at least make it clear that we continue support for the released stables (since that guy was concerned about it)
(05:49:44 PM) Keybuk: I thought elmo's played "missing, presumed STOLEN" ?
(05:49:57 PM) elmo: Keybuk: that was powerbook #1, he means powerbook #3
(05:49:58 PM) sabdfl: pitti: +1, could you add that?
(05:50:02 PM) mdz: pitti: we'll make a proper announcement, not point people to the spec
(05:50:03 PM) Keybuk: what happened to #2 ?
(05:50:04 PM) elmo: #2 had a logic board failure and was replaced
(05:50:10 PM) pitti: mdz: right
(05:50:26 PM) pitti: sabdfl: I can add it anyway, I figure, just for clarity
(05:50:28 PM) mdz: especially considering it's been wikified
(05:50:41 PM) elmo: (the [original] spec is my fault)
(05:50:56 PM) sabdfl: there are good, high profile dev's who use Ubuntu for its PPC support
(05:51:02 PM) ***ogra makes a note to wear his hotpants next time he meets elmo
(05:51:03 PM) sabdfl: we will lose some of them, certainly
(05:51:19 PM) sabdfl: though, flip side, it might encourage them to get more organised to contribute to the port
(05:51:24 PM) mdz: sabdfl: developers who contribute to Ubuntu?
(05:51:34 PM) sabdfl: if they think it will have a material impact on the iso's
(05:51:44 PM) Keybuk: mdz: a fair number of ubuntu's core developers are PPC users
(05:51:47 PM) sabdfl: mdz: other than canonical, NAFAIK
(05:51:49 PM) mdz: I know of a few who use it and may be advocates, but don't participate
(05:52:11 PM) mdz: Keybuk: point taken, I meant externally
(05:52:12 PM) mjg59: mdz: I think it's overly simplistic to think of the benefit from developers using Ubuntu purely in terms of the contributoins they make back to Ubuntu
(05:52:14 PM) sabdfl: surely core folks would have made more comments on the spec?
(05:52:29 PM) sabdfl: mjg59: agreed
(05:52:31 PM) Keybuk: pitti: random question
(05:52:43 PM) mdz: mjg59: certainly, but it's misleading to characterize it as losing developers
(05:52:44 PM) Keybuk: if this spec is approved, would you replace your PPC with an i386 or amd64 based machine?
(05:53:02 PM) pitti: Keybuk: no, I wouldn't, not immediately
(05:53:14 PM) pitti: Keybuk: we have edgy and dapper, and they work well
(05:53:15 PM) mdz: mjg59: if they're only with Ubuntu because of PowerPC, then when their PowerPCs die, they might just as likely switch to another distro then anyway
(05:53:17 PM) mjg59: mdz: Do you have any feeling for how the edgy release would have been affected if we hadn't released PPC as an official architecture?
(05:53:18 PM) Keybuk: pitti: so you'd, in fact, still be spending time on the PPC port; simply to keep your primary development box up and running
(05:53:35 PM) pitti: Keybuk: it's not my primary box, but I would probably apply some bug workarounds
(05:53:38 PM) mdz: mjg59: yes, my team would have slept more
(05:53:48 PM) pitti: Keybuk: but I would definitively not buy any new powerpc hardware any more
(05:53:48 PM) elmo: slackers
(05:53:52 PM) sabdfl: pffst
(05:53:55 PM) sabdfl: sleep
(05:54:05 PM) mdz: mjg59: it takes the better part of a full day just to do a quick validation cycle on all of the ISOs we ship in an official release
(05:54:08 PM) pitti: in fact, I bought that ppc in the first place to have all supported arches at home :)
(05:54:17 PM) mdz: mjg59: powerpc is a multiplier on that
(05:54:18 PM) ogra: likewise
(05:54:32 PM) pitti: mdz: 7 hours for me, on the 'all goes well' case
(05:54:33 PM) mjg59: mdz: If you feel that the current release testing schedule is impractical while maintaining sanity, and that the easiest way to rectify this is to drop PPC, then I'm happy to go with you on that.
(05:54:59 PM) mdz: mjg59: we're pursuing a number of strategies to make it less painful, including more automation and more community involvement
(05:55:00 PM) Keybuk: pitti: would you buy new PPC hardware if we continued to support it?
(05:55:02 PM) mjg59: mdz: But I'd like to see an argument for why other alternatives are impractical
(05:55:08 PM) sabdfl: mdz: if we had three community folks commit to being very active at crunch time, would you consider it less of an issue?
(05:55:14 PM) mdz: mjg59: PowerPC is a very simple one which can be executed immediately with immediate benefit
(05:55:15 PM) mjg59: Such as throwing more people at it, for instance
(05:55:25 PM) Keybuk: PowerPC is the only architecture we don't test mostly in vmware <g>
(05:55:30 PM) mdz: mjg59: people = money, that's just converting the cost from one unit to another
(05:55:43 PM) pitti: Keybuk: not a PS3; if Apple would produce a G5 laptop, then I'd consider it (for the same argument: arch coverage at home, and these things also have a certain appeal)
(05:55:48 PM) mdz: sabdfl: I can't see that working; it isn't reasonable to expect three volunteers to go through what we do at release time
(05:55:57 PM) mdz: it's weeks of solid full-time work
(05:56:00 PM) mjg59: mdz: Yes. But it's much easier to think in terms of "Is PPC support worth ($x)"
(05:56:08 PM) pitti: Keybuk: in warty/hoary times, the OOTB quality of the ppc port was pretty amazing, too
(05:56:24 PM) pitti: since then, the quality didn't really drop, but expectations rose :)
(05:56:34 PM) sabdfl: mjg59: we have a hard number on that, it's about $750k p.a. all in, if you factor in security support over time
(05:56:39 PM) mjg59: mdz: If what you're saying is that you don't believe your team can realistically continue to support PPC, then, as I said, I'm absolutely happy to go with that
(05:56:49 PM) mdz: mjg59: manpower would address a part of the issue, but there are other bits which don't parallelize the same (e.g., the time to roll a complete, new set of images)
(05:57:16 PM) mjg59: But that's not the sort of impression I get from the spec, and I think the arguments the spec makes are much weaker
(05:57:17 PM) sabdfl: mdz: iso building should parallelize across architectures, surely?
(05:57:21 PM) mdz: mjg59: I don't think it's practical for the existing team, as staffed, to support it while maintaining the standard of quality we want for more mainstream architectures
(05:57:26 PM) elmo: sabdfl: no
(05:57:30 PM) elmo: sabdfl: there's only one IO path in the machine
(05:57:31 PM) Mithrandir: sabdfl: no, we build all the ISOs on one machine.
(05:57:43 PM) Mithrandir: (we build the livefs-es in parallell, which helps a lot)
(05:57:43 PM) sabdfl: interesting. that seems fixable
(05:58:03 PM) ***elmo cries quietly in the corner
(05:58:13 PM) elmo: this spec wasn't meant to get me MORE machines
(05:58:14 PM) sabdfl: we have buildd's, and they are explicitly not building packages at the time we are rolling iso's
(05:58:18 PM) mdz: mjg59: I agree that the spec doesn't cover all arguments; we discussed at great length at UDS
(05:58:42 PM) mjg59: mdz: Sadly, time constraints meant I wasn't able to participate in the UDS sessions to any great extent
(05:58:48 PM) mdz: sabdfl: building a complex distributed architecture to replace a simple centralized one has its own cost
(05:58:51 PM) elmo: sabdfl: you don't want to be building IOs on our buildds. it's super ISO intensive
(05:58:57 PM) mdz: both in development and upkeep
(05:58:58 PM) Mithrandir: sabdfl: I'd much rather have an even faster cd build machine than multiple build machines. The former makes my job more complex. (Though that discussion is tangential to the discussion at hand).
(05:59:02 PM) mjg59: mdz: But I'd like to see the spec make the arguments that are considered compelling
(05:59:12 PM) elmo: (and like mdz says, it's a much more complicated architecture to develop/maintain)
(05:59:20 PM) mdz: mjg59: I think different folks find different arguments compelling, but of course I agree that the spec should be comprehensive
(06:00:10 PM) mdz: mjg59: elmo made reference to another similar multiplier cost; sysadmin resources
(06:00:23 PM) Keybuk: ok, three hours
(06:00:35 PM) mdz: it's 4xN machines which have certain resource requirements, rather than 3xN
(06:00:36 PM) Keybuk: we don't appear to be reaching any form of consensus here :-/
(06:00:43 PM) sabdfl: sorry, keybuk, didn't foresee the fireworks
(06:00:49 PM) mjg59: Ok. My opinion is that there are technical arguments for and against keeping the port, and neither convinces me.
(06:00:58 PM) mjg59: There are also commercial arguments.
(06:01:07 PM) mjg59: I consider the commercial arguments to be stronger than the technical ones.
(06:01:19 PM) mjg59: And so, personally, I feel that this is a decision best left to the management of the distribution.
(06:01:25 PM) mdz: I don't think there are technical arguments
(06:01:31 PM) mdz: it doesn't inherently suck as an architecture
(06:01:38 PM) mdz: it's a matter of resources as far as I'm concerned
(06:01:47 PM) Keybuk: I don't really see any arguments for tossing the port out other than mdz's desire to reduce his team's workload
(06:01:49 PM) mdz: we should spend our resources where they provide the best value to Ubuntu
(06:02:01 PM) sabdfl: mjg59: you are the management of the distribution :-)
(06:02:02 PM) Keybuk: the technical ones are all largely invented to suit the spec
(06:02:03 PM) mdz: Keybuk: it's not reducing the workload so much as spending it more wisely
(06:02:16 PM) sabdfl: in the sense that, if TB wants to allocate resources to PPC, I'm fine with it
(06:02:34 PM) sabdfl: but the result will be some fewer resources will go to other options
(06:02:21 PM) mjg59: sabdfl: We get to decide funding-requiring decisions? Sweet! When do I get an entertainments budget?
(06:02:42 PM) sabdfl: mjg59: i sing, i dance...
(06:02:46 PM) Mithrandir: mjg59: I'm sure you can get a PS3 to do acpi hacking on. ;-P
(06:02:47 PM) ***Mithrandir hides
(06:02:58 PM) Keybuk: I don't see that the TB has any decision about how mdz allocates his resources ... we can certainly say we'd like it to be an Ubuntu architecture; but that requires a team to want to do it
(06:03:08 PM) Keybuk: up until now, that's been the canonical distro team
(06:03:12 PM) sabdfl: well
(06:03:15 PM) sabdfl: true
(06:03:16 PM) mdz: the two are not really separable
(06:03:17 PM) Keybuk: if they want to cease, that's no different than all the motu resigning
(06:03:25 PM) mdz: Canonical is deeply involved in core Ubuntu development
(06:03:32 PM) mdz: especially in the areas hardest hit by PowerPC (e.g., sysadmin, release process)
(06:04:16 PM) sabdfl: if the TB decides to keep PPC, i won't argue that we don't have the resources
(06:04:29 PM) mjg59: If it's practical, I think it would be nice to see us do one official release that supports the PS3
(06:04:32 PM) mdz: resourcing is never that simple
(06:04:51 PM) sabdfl: IOW, i see this as a "what's best for ubuntu" decision, and trust the TB w.r.t. considering the alternative uses of those resources
(06:05:05 PM) mjg59: Because I think that's the only real hope of there being a significant change in the PPC userbase trend
(06:05:08 PM) sabdfl: my own view is that i would rather be devoting more resources to the arch's which have huge user bases
(06:05:20 PM) sabdfl: people for whom the economics of free software are potentially vital
(06:05:20 PM) mdz: mjg59: do you feel that official status is important in that equation?
(06:05:23 PM) pitti: mjg59: although PS3 is certainly an interesting target to hack on, do you reckon it would have an actually interested user base for putting Linux on it? (I seriously don't know)
(06:05:35 PM) mjg59: pitti: I have absolutely no idea
(06:05:36 PM) sabdfl: i think most PPC linux users are entirely voluntarily so, and probably have multiple machines
(06:05:42 PM) sabdfl: that's not true of i386
(06:05:43 PM) mdz: mjg59: people use Debian on PowerPC widely, and it has similar characteristics to a community-driven Ubuntu port
(06:05:43 PM) Keybuk: I don't think PS3 is really that interesting (I don't have one :p)
(06:05:47 PM) elmo: I was super convinced by the PS3 argument. until i discovered it had 256Mb of memory
(06:05:47 PM) Keybuk: PS2 never has been an interesting arch
(06:05:49 PM) mjg59: I think we have a better chance if there's an official release, sure
(06:05:51 PM) sabdfl: so, philanthropically, i don't think PPC is a win
(06:05:57 PM) Keybuk: most people will play games on them, not Linux
(06:05:57 PM) mdz: sabdfl: ++
(06:06:06 PM) Keybuk: tuxracer is not what a PS3 was meant for
(06:06:11 PM) elmo: Keybuk: PS2 was an entirely different chipset and machine and Sony weren't pushing Linux on it
(06:06:18 PM) elmo: I'm not sure it's a fair comparison
(06:06:25 PM) Keybuk: elmo: Sony pushed Linux fairly hard to the developer community, and got nowhere
(06:06:33 PM) mjg59: I think the main argument in favour of PS3 support is that it gives something that can get into the hands of kids who wouldn't be allowed to install Linux on their parents PCs
(06:06:42 PM) Keybuk: "look, Linux on my PS2! sweet! now what?"
(06:06:45 PM) elmo: Keybuk: not in anything like the same way
(06:06:56 PM) elmo: Keybuk: they've been leading the charge for the PS3, contributing ports to gcc, binutils etc.
(06:06:57 PM) mjg59: Keybuk: Linux on the PS2 always involved extra hardware
(06:06:59 PM) mdz: unless Cell takes off and desktop application software starts to take advantage of it, I don't think PS3 is very interesting for Ubuntu or Linux
(06:07:03 PM) Keybuk: elmo: we do actually support 256MB of memory
(06:07:21 PM) elmo: Keybuk: yes, have you tried using it in anger? I have, and I upgraded my RAM as a result
(06:07:25 PM) Keybuk: elmo: yes
(06:07:30 PM) Keybuk: I deliberately test with it
(06:07:31 PM) ***pitti had 256 MB on his iBook until a month ago, worked fine (apart from OO.o)
(06:07:35 PM) Keybuk: you can't use oo.o
(06:07:37 PM) elmo: ...
(06:07:38 PM) Keybuk: but other than that, it's not too bad
(06:07:44 PM) elmo: because no one uses an office app, right?
(06:07:52 PM) mjg59: Well, probably not on a PS3
(06:07:53 PM) Keybuk: *shrug* abiword was fine
(06:07:55 PM) Keybuk: and so was gnumeric
(06:08:02 PM) pitti: latex+gnumeric for me
(06:08:13 PM) lifeless: kobodl on ps3 ;) nice image.
(06:08:30 PM) mjg59: But I still don't think we're really getting anywhere
(06:08:41 PM) Keybuk: I'd propose the following:
(06:08:48 PM) sabdfl: if sony decides to push linux on ps3
(06:08:48 PM) Keybuk: - mdz should allocate the resources of his team how he best feels
(06:08:56 PM) sabdfl: they will pick a distro partner
(06:09:01 PM) sabdfl: so far, it looks like YDL
(06:09:01 PM) Keybuk: - we'd like to keep PPC as an Ubuntu architecture, whether officially or community supported
(06:09:16 PM) ogra: PS3 would make a wonderful ltsp fat-client :)
(06:09:16 PM) ogra: even with oo.o
(06:09:22 PM) Keybuk: - if the canonical distro team cease official work on it, we'd welcome a community team to take over
(06:09:43 PM) mdz: Keybuk: we can't claim an official port without any developers backing it; Canonical is the backbone of the port
(06:09:56 PM) Keybuk: mdz: that's why I didn't say "official architecture"
(06:10:18 PM) pitti: but then we should be honest and call it 'port' right away IMHO
(06:10:22 PM) sabdfl: if we drop it as an official architecture, i think we should emphasise how the community can keep the port ticking
(06:10:25 PM) mdz: Keybuk: that's the same as saying "we'd like to keep PPC as a community supported architecture"
(06:10:29 PM) sabdfl: - builds continue
(06:10:33 PM) Keybuk: mdz: that's what I said :)
(06:10:34 PM) sabdfl: - iso builds continue
(06:10:44 PM) sabdfl: - testing schedule is public so community folks can sign off
(06:10:48 PM) Keybuk: my wording was perhaps not clear
(06:10:56 PM) bgoldstein left the room (quit: "Ex-Chat").
(06:10:58 PM) sabdfl: - we'll include it in release announcements if it's baked on time
(06:11:00 PM) sabdfl: etc
(06:11:05 PM) mdz: given that everyone seems to think that we're talking about removing the port entirely, we could present this as a decision to rescue the powerpc port as a community-driven project ;-)
(06:11:16 PM) Keybuk: yes
(06:11:16 PM) sabdfl: mdz: spin artist
(06:11:17 PM) mdz: ...and the crowd goes wild
(06:11:26 PM) mdz: sabdfl: you hired me
(06:11:34 PM) sabdfl: mdz: where were you last night?
(06:11:40 PM) Keybuk: Xubuntu manages to be an ubuntu derivative without having any kind of officially supported status
(06:11:51 PM) Keybuk: I don't see why PPC can't have the same success within the community
(06:11:54 PM) mdz: sabdfl: in bed asleep
(06:11:57 PM) sabdfl: at the death of transparency
(06:12:00 PM) sabdfl: you presided
(06:12:03 PM) sabdfl: ah well
(06:12:15 PM) sabdfl: i think this is a chance for the PPC folks to step up
(06:12:17 PM) sabdfl: you never know
(06:12:32 PM) mdz: Keybuk: agreed
(06:12:42 PM) sabdfl: a few good folks and PPC could be in great shape for feisty
(06:13:21 PM) mjg59: Entirely coincidentally, I've actually just switched on my PPC dev box for the first time in a year
(06:13:29 PM) Keybuk: mjg59: any objection to that position for the TB?
(06:13:42 PM) mjg59: Keybuk: I think that's actually the best we can come up with.
(06:14:15 PM) sabdfl: ok
(06:14:22 PM) sabdfl: i think we have consensus
(06:14:30 PM) sabdfl: - ppc becomes a community port
(06:14:45 PM) sabdfl: - we communicate strongly that existing stable releases continue to get maintanance
(06:14:52 PM) vil left the room (quit: Read error: 110 (Connection timed out)).
(06:14:59 PM) sabdfl: - we invite PPC team to form around the port
(06:15:08 PM) sabdfl: - we continue to make iso's etc as long as they are buildable
(06:15:19 PM) mdz: https://launchpad.net/people/ubuntu-powerpc
(06:15:19 PM) sabdfl: - we include PPC in release announcements if the PPC team has signed off on it
(06:15:28 PM) Keybuk: - we continue to provide PPC buildds as long as the hardware works
(06:15:54 PM) sabdfl: yes, and if the community is active we can replace those when the time comes too
(06:15:58 PM) pitti: installer work commitment?
(06:16:01 PM) sabdfl: no problem funding hardware
(06:16:05 PM) mdz: PPC buildds and such may not be expected to meet the same service level as official architectures, but they are redundant
(06:16:11 PM) sabdfl: pitti: community
(06:16:18 PM) Keybuk: pitti: the installer should probably be maintained by the community/team
(06:16:27 PM) Keybuk: the colin clone farm crop was lost
(06:16:33 PM) sabdfl: mdz: i wouldnt want the community team to feel they are working against the odds
(06:16:52 PM) sabdfl: so, give them the horsepower
(06:17:11 PM) sabdfl: i'm sure we can get donated hardware, for a start
(06:17:12 PM) mdz: that means nothing would change for elmo; if that's ok with him then sure
(06:17:28 PM) sabdfl: elmo just got a raise ;-)
(06:17:48 PM) mdz: if e.g. hppa craps out, we don't wake elmo
(06:17:57 PM) ogra: "find *donated* PPC hardware" is a raise ? :)
(06:18:15 PM) elmo: it's fine by me
(06:18:33 PM) sabdfl: ok
(06:18:40 PM) elmo: [well, not the donated hardware part without restrictions (i.e. I don't want to run a farm of laptops in our DC), but we can discuss those later]
(06:19:00 PM) sabdfl: elmo: i had quality donations in mind, from the manufacturer
(06:19:01 PM) sabdfl: any dissenting opinions?
(06:19:15 PM) mdz: I'm comfortable with those bullet points
(06:19:48 PM) sabdfl: mjg59? Keybuk?
(06:19:54 PM) Keybuk: fine by me
(06:19:57 PM) mjg59: I'm good
(06:20:20 PM) sabdfl: ok. we may get to revisit this in future, if resources are made available. for now i think it is the right decision.
(06:20:26 PM) sabdfl: phew
(06:20:29 PM) sabdfl: mdz?
(06:20:39 PM) mdz: sabdfl: shall I draft the announcement? "we are pleased to announce that powerpc has been updated to community status!"
(06:20:36 PM) siretart: next (and last) agenda item?
(06:20:43 PM) Keybuk: siretart: you snuck that in, that's naughty
(06:20:52 PM) mdz: Keybuk: he cleared it with me
(06:20:55 PM) mdz: about 5 minutes into this meeting
(06:21:01 PM) mdz: but perhaps he will have mercy and defer to the next meeting
(06:21:31 PM) sabdfl: mdz: easy on the spin, it needs to be sincere, not everyone will get the in-jokes
(06:21:33 PM) mdz: siretart: would that be OK with you?
(06:21:36 PM) siretart: mmmh, it would have been nicer if that would have been more clear a few hours earlier
(06:21:44 PM) mdz: sabdfl: apparently not everyone gets my jokes even here
(06:21:50 PM) siretart: I don't want big discussion about that anyway
(06:21:54 PM) pitti: Seveas: just another proof of brainmelt :)
(06:21:57 PM) siretart: I'm rather interested in what the problems are
(06:22:08 PM) sabdfl: siretart: juris-my-diction crap
(06:22:19 PM) sabdfl: anyone, anyone...
(06:22:31 PM) siretart: so I suggest I paste my lines and hear 3 minutes of comments about my proposal
(06:22:45 PM) sabdfl: siretart we can't possibly do it justice in that time
(06:22:47 PM) sabdfl: it's a minefield
(06:22:57 PM) pitti: what's the topic again?
(06:23:04 PM) ogra: pitti, patents
(06:23:05 PM) lifeless: Inclusion of packages in main with potential patent issues - ReinhardTartler
(06:23:09 PM) sabdfl: there are no patent issues in linux
(06:23:11 PM) siretart: it is a question about policy rather
(06:23:18 PM) siretart: I wanted to ask what the current policy regarding potentially patented packages for ubuntu main is.
(06:23:26 PM) siretart: Please note that we have a lot of packages, which are indeed patented, but not actively enforced. We seem to have no problem with that. However, there is big uncertainity when it comes to multimedia packages.
(06:23:30 PM) siretart: I'm asking about multimedia packages like ffmpeg. While they are DFSG free and in debian main, this package has recently been demoted from main to universe in breezy.
(06:23:30 PM) mjg59: Aren't we broadly aligned with Debian here?
(06:23:33 PM) mdz: summary: there exist free software programs that aren't appropriate for Ubuntu to include because of potential litigious issues
(06:23:33 PM) siretart: I have some plans with xine. in detail, I want to undo the split of the source package that was done in the xine package. I want to arrange means by splitting the unified source package in seperate binary packages, so that no mp3 decoder needs to be in the ship seed (and therefore on the desktop cd)
(06:23:36 PM) pitti: libmad | 0.15.1b-2.1 | http://de.archive.ubuntu.com feisty/main Sources
(06:23:38 PM) siretart: But I wanted to ask on what to look after while preparing my request.
(06:23:55 PM) siretart: so in fact, my question is if there are patent problems with ffmpeg having in main
(06:23:58 PM) siretart: that's basically all
(06:24:00 PM) Keybuk: siretart: ffmpeg is the subject of patents being actively enforced
(06:24:05 PM) Keybuk: which is why it's not in main
(06:24:11 PM) siretart: Keybuk: and why does it exist in debian/main then?
(06:24:16 PM) siretart: sorry, I don't get it at all
(06:24:23 PM) mjg59: In that we'll ship anything if we don't think there's a realistic chance of us or our users being sued
(06:24:24 PM) Keybuk: because debian don't usually discriminate based on patents
(06:24:30 PM) sabdfl: siretart: it's quite possible there are inconsistencies
(06:24:35 PM) mjg59: Debian may have different ideas about how realistic a threat is
(06:24:35 PM) siretart: Keybuk: which patents? the mp3 encoding patent? that is already disabled
(06:24:36 PM) mdz: in some cases the issues are entirely unclear, in some they are just clear enough that some concern is justified (and we shouldn't ship them), and in others they are clear as a heart attack
(06:25:11 PM) mjg59: I was under the impression that our patent position had always been clear - if that's not the case, then we should clarify that
(06:25:17 PM) mjg59: But otherwise, it's a package by package issue
(06:25:25 PM) siretart: Keybuk: even in debian (and we ship a very similar package)
(06:25:38 PM) mjg59: A package's position in Debian is indicative, but doesn't define our approach to the same package
(06:25:56 PM) siretart: mjg59: to me the patent position is very unclear to me. espc. the differences to debian's policy
(06:26:05 PM) mjg59: siretart: There are no differences to Debian's policy.
(06:26:08 PM) siretart: and the fact that we had ffmpeg in main for hoary
(06:26:15 PM) Keybuk: we don't ship anything in main if we believe there's a reasonable chance either we, or our users, would get sued
(06:26:18 PM) mjg59: But even with the same policy, we can reach different outcomes.
(06:26:21 PM) sabdfl: by and large, if there is a clear patent problem with free code, we make the code availabel via network repository for people to whom the patent issue does not apply
(06:26:25 PM) siretart: mjg59: in this case, I conclude that there wouldn't be any problems with ffmpeg in main
(06:26:45 PM) siretart: mjg59: but there seems to be be problems. you see why I'm highly confused?
(06:26:48 PM) mjg59: siretart: If you have issues with a specific package, then write a main inclusion report sufficiently good to counter any arguments over patents
(06:26:54 PM) sabdfl: if the patent is not enforced, or not applicable in most jurisdictions, then we've shipped code in the past
(06:27:12 PM) mjg59: siretart: If there are no actively enforced patents covering our ffmpeg package, it can go in main.
(06:27:30 PM) mdz: siretart: shipping questionable software to countries which enforce that sort of thing is widely considered to be a bad idea
(06:27:30 PM) siretart: mjg59: I wanted to have this cleared before I start to write an inclusion request. espc for the highly disputed package ffmpeg
(06:27:49 PM) mjg59: But just as us having a package in main doesn't mean that Debian will, a package's location in Debian may not match ours
(06:27:58 PM) mdz: we are in a very different position than Debian where potential litigation is concerned
(06:27:59 PM) pitti: mdz: 'shipping' applies to 'put on CD', I figure?
(06:28:04 PM) sabdfl: main is different to ship
(06:28:08 PM) mdz: pitti: or, *cough* DVD
(06:28:10 PM) mjg59: I don't think there's anything else the TB can say in this respect
(06:28:11 PM) siretart: mdz: the thing is that to the research of the debian maintainer, there are no patents known to be enforced, which would be applicable to the ffmpeg package
(06:28:17 PM) siretart: TTBOMK, at least
(06:28:18 PM) mdz: sabdfl: not by much
(06:28:25 PM) sabdfl: yeah, forgot about the DVD
(06:28:29 PM) pitti: mdz: we don't press DVDs?
(06:28:43 PM) sabdfl: give me 5 minutes, i'll have forgotten about it again and be happy :-)
(06:28:48 PM) siretart: the thing is this: I want either ffmpeg in main, or xine being demoted
(06:28:51 PM) Mithrandir: pitti: amazon sells Ubuntu DVDs.
(06:29:01 PM) mdz: I am legally protected from participating in this conversation in this country
(06:29:01 PM) mjg59: siretart: That's not a TB decision
(06:29:05 PM) siretart: the problem with the demotion is, that the kubuntu and xubuntu folks would be very pissed
(06:29:11 PM) lifeless: mdz: protected ?
(06:29:15 PM) siretart: because they rely heavily on libxine
(06:29:27 PM) mdz: lifeless: protected.
(06:29:33 PM) lifeless: wow
(06:29:34 PM) mjg59: Or, rather, it's not a TB decision as framed in the current topic
(06:29:44 PM) siretart: mjg59: well, since it affects both xubuntu and kubuntu, why isn't this a TB topic?
(06:29:53 PM) mjg59: siretart: Because you said you were going to talk about patents
(06:30:23 PM) siretart: mjg59: yes. with the goal to save much PITA regarding xine packaging
(06:30:33 PM) mjg59: I believe that our patent policy is clear, and that we've made it clear how you should proceed on this issue.
(06:30:53 PM) siretart: mjg59: okay. I will continue with an main report, and refer to logs to this discussion
(06:31:35 PM) siretart: mjg59: I still suggest that we have a wiki page with some patent policy statement
(06:31:57 PM) Mithrandir: in the entirely hypotetical situation that an archive admin is unsure about whether something is actually suitable for main, should he bring it up for the tech board then?
(06:32:27 PM) mjg59: Mithrandir: It sounds more like a legal decision than a technical one...
(06:32:57 PM) Mithrandir: I'd not like to be the one promoting ffmpeg if it ends up causing Canonical to be sued.
(06:33:07 PM) mc44: "not actively enforced" would seem to be a very legal quesion
(06:33:25 PM) siretart: does it really matter in what directory in the archive we ship it?
(06:33:29 PM) sabdfl: if a company claims to have a patent, they are welcome to approach us to discuss it
(06:33:30 PM) ***pitti would have no clue how to answer/verify this when processing a MIR
(06:33:36 PM) sabdfl: so far, AFAIK that has never happened
(06:33:40 PM) sabdfl: thanks all
(06:33:41 PM) mjg59: siretart: To our users? Yes.
(06:33:51 PM) Keybuk: siretart: yes, in most legal juristictions it makes a big difference whether we send it on CD, or a user chooses to download it
(06:33:53 PM) sabdfl: wow
(06:33:54 PM) siretart: mjg59: this is a legal discussion
(06:34:05 PM) sabdfl: three hours, 5 new developers, 1 patent policy, one less architecture
(06:34:08 PM) sabdfl: that was fun :-)
(06:34:09 PM) siretart: Keybuk: I plan to arrange that ffmpeg does NOT need to be on the cd
(06:34:10 PM) Mithrandir: sabdfl: do we have a legal consel we can talk to or should the ubuntu-archive team members not care about this problem until we're approached?
(06:34:23 PM) siretart: Keybuk: I want it just in main, not on the live cd
(06:34:24 PM) Keybuk: sabdfl: closer to four hours now ;)
(06:34:31 PM) siretart: Keybuk: I need it as build dependency for xine
(06:34:47 PM) sabdfl: Mithrandir: use reasonable discretion, don't panic, don't ship DeCSS
(06:34:51 PM) mdz: seeking legal counsel for archive changes is not a reasonable approach for us
(06:34:53 PM) Keybuk: siretart: as mjg59 says, file a MIR
(06:35:06 PM) sabdfl: there's always OIN
(06:35:19 PM) sabdfl: btw, folks we just asked to join FFII
(06:35:20 PM) mdz: DeCSS is an entirely separate issue (copyright circumvention), no?
(06:35:25 PM) elmo: mdz: yes
(06:35:28 PM) Mithrandir: sabdfl: and in the cases we're not sure? I'm not talking about the obvious cases, they're not very interesting.
(06:35:29 PM) Keybuk: sabdfl: ironically, Mithrandir can ship DeCSS :) he's in Norway
(06:35:29 PM) sabdfl: formally funding the fight against software patents in europe
(06:35:34 PM) mc44: sabdfl: that doesnt work for mp3 patent holders though :)
(06:35:37 PM) sabdfl: an announcement will come soon
(06:35:39 PM) elmo: Keybuk: he can't ship it on OUR servers
(06:35:42 PM) siretart: Keybuk: I interpret this as "the tb does not see principal or legal problems with having ffmpeg in main". that is the answer I wanted to hear
(06:36:07 PM) Keybuk: siretart: the TB has clarified our patent policy, and has no opinion either way on ffmpeg
(06:36:13 PM) Keybuk: we haven't investigated it
(06:36:15 PM) sabdfl: Mithrandir: it's up to someone to enforce their patents
(06:36:24 PM) Mithrandir: sabdfl: I'll be happy to take this to email since everybody is tired of the meeting now.
(06:36:30 PM) sabdfl: that's the law, i believe
(06:36:36 PM) ogra: Mithrandir++++
(06:36:37 PM) siretart: ok. let's end this.
(06:36:40 PM) sabdfl: thanks everyone, very good meeting
(06:36:46 PM) siretart: gn8 everyone
(06:36:47 PM) pitti: siretart: thanks for bringing this up
(06:36:50 PM) sabdfl: i'll be happy when we have the MC in place to deal with approvals
(06:36:58 PM) sabdfl: it's fun to talk through the policy stuff though
(06:37:04 PM) sabdfl: night / afternoon / morning all
(06:37:09 PM) pitti: thanks everyone
(06:37:15 PM) mdz: we'll have more time to spend arguing about this sort of thing
(06:37:19 PM) mdz: thanks all
(06:37:22 PM) ogra: thanks and night ...

MeetingLogs/Technical-2006-12-05 (last edited 2008-08-06 16:21:53 by localhost)