[00:05] persia: Surely you jest. At >5000 daily you are ... well ... it must be in the realm of hyperbole. [00:06] persia: That's uh... well... the math doesn't work. === irvy is now known as irv [00:35] troy_s, That's the bottom-end number for a single factory line running for 8 hours. [00:35] Mind you, it's not every day, but it's not uncommon. [00:37] Last time I was contracted to an OEM to work on a preinstall device, the first month's run was 40,000 units (just over a week's output for a line). I wasn't renewed, so don't know numbers for follow-up months. [01:39] persia: And those are all Ubuntu Studio installs? [01:39] No, _Desktop Linux_ installations [01:40] Far as I know, nobody is selling preinstalled Ubuntu Studio in volume (although some folks are selling it piecemeal) [01:41] (and nobody is going to run a factory line if they aren't selling something) [01:41] persia: But those aren't done by a live CD for the love of all things good. [01:41] Well, hacked, but kinda. [01:41] * persia tries to find the code Dell instructs their factories to use, which uses an LiveCD as input [01:41] persia: Which brings us back to the value and need for the Live disk. I would bet the need is less than 0.1% of current users, and current users is... well... let's say low. ;) [01:42] http://en.community.dell.com/support-forums/software-os/w/linux/building-base-ubuntu-factory-iso.aspx [01:43] troy_s, I agree. That said, I believe that if someone (e.g. holstein) wants to spend several hours a week working on it, I have no business telling them not to do so unless it actively disrupts something I'm doing. In the worst case, it motivates the person doing the work to also fix related bugs (which affect users in more interesting ways) [01:44] persia: Yes. Agree. [01:44] persia: It is hard to deduce value in any component given the fact that no one really uses the bits. [01:45] I know a fair number of people who use the bits in the repos. I'm less confident about the number of users who use the install images. [02:17] persia: The problem that I'd think most have (in my case especially) is that the actual useful bits are horrifically out of date with the nature of the development speed. FFMPEG for example. This greatly reduces the actual 'real-world' usage of the various PPAs. [02:17] persia: uS might do well to adopt a more bleeding edge GIT / SVN daily / weekly repository script setup to bring value to those that need the bits. [02:18] My mother would disagree with that. She doesn't like change. [02:18] persia: uS isn't for your mother likely. [02:18] Funny. She uses it daily. [02:19] persia: Although with the clear lack of audience, it is difficult to tell exactly what it is for. [02:19] persia: What does your mother use it for? [02:19] persia: (It is also why I suggested perhaps daily / weekly flavours. I know that I'd happily hope on a PPA system if it were available) [02:19] Recording, songwriting, preparing stuff for students, email, browsing, minor accounting, mastering. [02:20] That mostly requires someone motivated to keep that up to date (and with the time to do so) [02:20] If someone did that, getting stuff into Ubuntu Studio for the 6-monthly cycle would likely be smoother. [02:22] persia: Some of those components must be somewhat scriptable no? [02:22] persia: Like some of the other daily / nightly builds on Launchpad? [02:23] Indeed, and there's a few scripts about that do it. I think fta has the most mature ones. Still, he spends nearly 3 hours a day just watching *one* package to have dailies available. [02:23] (including support, bug tracking, upstream coordination, etc.) [02:24] persia: Brutal. [02:24] persia: There has got to be a better way. [02:25] Not really. If you build it, they will come. If you don't have a plan to receive them, there will be fires and riots. [02:25] lol [02:25] persia: That's not really true actually. [02:25] persia: PT was a little off course. [02:25] persia: ;) [02:26] persia: If you build it you can struggle like mad to have it built and then watch nothing happen. [02:26] persia: Case in point, watch blog stats. ;) [02:26] * persia would claim you failed to build the right thing in that case :) [02:26] LOL [02:26] That's a cheat. [02:26] But, yeah, in the case where nobody uses the dailies, they are easy to support, but one then as to ask: what's the point? [02:27] Avoiding the breakage in "Build it and they will come" and marking it as "Works as expected. Won't Fix" [02:27] If folks use them, they become harder to support, etc. [02:27] that's not quite in the spirit of the CoC: something about respecting others. [02:27] If folks want unsuppported stuff, they ought pull upstream directly. [02:28] persia: It just takes me to the usefulness of the various bits. I know that for example, I might be someone that could long shot be interested in using uS, but I wouldn't because of the nature of Ubuntu's packages. [02:28] persia: Just way too damn old. [02:29] I guess. Speaking for myself, I don't care about the age, so long as they work. [02:29] I can wait a few months for them to work better. [02:47] persia, what kind of music does your mother make? which applications does she use? is she using jack? [02:48] Yes (she's an ethnomusicology consultant), mostly hydrogen/muse/qsynth, yes. [02:48] (well, and openoffice and firefox, but we don't maintain those) [07:08] * abogani waves [07:08] ScottL: I am here now. [13:30] abogani: good morning [13:30] scott-work, To you. [13:31] abogani: are you working with the UKT to get the -rt kernel into maverick? [13:31] * scott-work doesn't think that it's morning in Italy though ;) [13:32] scott-work, You are thinking well ;) [13:32] after talking to Jfo i got the impression that they were expecting you (or as it was referred to as "community") to provide the 2.6.35 (?) -rt kernel [13:32] i'm presuming the 2.6.35 kernel is in maverick, i don't recall at the moment [13:33] It is [13:34] Jfo even mentioned that the kernel team was probably going to consider the -preempt or -latency kernels because they felt that they would be duplication of work considering the inclusion of the -rt kernel [13:34] ARGH [13:34] probably NOT going to consider [13:34] my impression from Jfo was that they felt the -generic and -rt kernels would cover all use cases [13:34] * scott-work is not sure he agress with that logic though [13:35] scott-work: I maintained and offered via kernel.ubuntu.com and PPA -lowlatency kernel. If they don't want accept it I can't really do more. [13:35] They do, but if they want to rely on -rt, they need to be more inclusive of abogani [13:36] persia: that would my concern as well and i voiced that to Jfo (who is really only a messenger in most of our conversations) [13:36] Right. [13:36] abogani, What are you doing the last week of October? [13:36] scott-work: The -rt/-realtime stick on .33. Non one in the world have an more recent version of it. [13:36] abogani: according to Jfo it would seem that the UKT would like (and is expecting) the -rt kernel from you [13:37] Can't, but -lowlatency is possible. [13:37] persia: Sorry? [13:37] are we saying that there isn't an upstream -rt patch for 2.6.33 ? [13:37] abogani, If you're free the last week of October, maybe it would help to meet UKT people in person. [13:37] scott-work: Which exactly the version they expect? 2.6.35? [13:37] scott-work, Rather there isn't an upstream -rt patch for 2.6.35 : only for .33 [13:37] * scott-work apologizes, hasn't really kept up with which kernel versions has -rt patches available [13:38] argh, yes, and .35 is going into maverick, right [13:38] scott-work: I made -lowlatency for two reasons: [13:38] well, then i expect more conversation with Jfo and UKT is in order [13:38] Indeed. [13:39] 1) It is simple to keep in sync with -generic and a lor of people are happy with -generic [13:39] scott-work, I'd suggest talking to smb and ogarasawa about it. smb has previously expressed interest in -rt, and ogarasawa is managing the mainline kernel sequencing this release cycle. [13:40] * abogani give up [13:41] persia: i shall then, i have already pinged Jfo in #ubuntu-community-team but will move onto #ubuntu-kernel then and see if i can talk to them, i've been warned that they are sometimes slow to come up out of code [13:41] abogani: please don't do that, we will get this sorted [13:42] i would like to believe that if we talk to a few select, directly responsible people then we can clear any confusion up, establish clear expectations, and develop a framework in which to achieve those expectations :) [13:44] scott-work: Really I *can't* do different. I sold all my computers and my car and now I have a little job for three months (after that probably I'll return unemployed). Moreover I don't want explain to UKT *every time* how -rt releases works. I'm really exhausted. [13:45] -lowlatency is probably a better target for the meantime. [13:45] -rt is just too hard without stronger support from upstream. [13:47] abogani: would it be possible to attend UDS in October ? (i realize you said you had 3 months of work) [13:47] you could talk directly to the kernel team and the face to face communication would probably clear up a lot of problems ;) [13:49] scott-work: I don't be able to speak English so well. :-) [13:49] abogani: i would expect that sponsorship might be worked out to get you there [13:49] abogani: bah, i think your english is better than a metric tonne of American's IMO [13:50] scott-work: Write and speak are very different things. [13:50] and it's not like there will not be many, many non-native English speaking people there as well [13:51] persia, for example. :-) [13:51] wait, i said that last sentence wrong...LOL i used too many double negatives and confused myself [13:52] there will be many people there who are not native Englishy speakers as well :-) [13:52] Ahhh [13:52] abogani: please really consider it [13:53] i don't think the UKT has us in their minds as a first priority and tend to forget what we have discussed previously [13:53] i think it will take persistance and perserverance with them to get ubuntu studio into their collective conciousness and psyche [13:54] and that might take a trip to UDS and presenting ourselves to their faces in person [13:54] if the -rt kernel isn't the way to go because of the fickle nature of upstream then we should dicsuss it with them once and for all [13:54] and decided a CONSISTENT direction forward for both the UKT and ubuntu studio [13:55] if the -rt isn't going to be always available then we go -lowlatency then all the time and make UKT understand how important it is and why [13:55] but if we can establish it then we can move forward without all this mucking about ...hopefully :) [13:55] (of course we can always offer -rt kernel in PPA when it is available though) [13:57] but anyway, consider UDS this October... as i said, i imagine we could secure sponsorship for you [13:57] abogani: ^^^ last sentence is important [14:00] scott-work: I'll help in any way I'll can but my "energy" is almost run out. [14:04] i can only imagine your frustration and weariness of trying to maintain a kernel for ubuntu studio abogani [14:05] from my short experience with the team i witnessed some of it and i expect that it certainly has been happened well before i came here [14:05] s/happened/happening [14:06] does that mean you'll go to UDS? :) [14:09] scott-work: Sorry I don't know. :-( I would prefer that Ubuntu Studio project find a replace for me. [14:23] abogani: i understand if you feel that you don't have the energy to keep doing this (i think you've been doing this for almost four years?) but i don't think anyone is available to replace you anyways [14:58] abogani: what kinda problems the rt kernel has? [14:59] abogani: the proprietary driver compilation and what else? [14:59] falktx: It isn't aligned with Maverick -generic. [14:59] abogani: ah, so fixes won't come as usual, right? [15:04] falktx: -rt is died, at least on official archives. [15:05] abogani: that's sad... [15:06] abogani: anyway, thanks for your work on this [16:04] abogani: That's interesting, as not many days ago there was a kernel team member here talking about interest towards -rt. [16:05] Haven't followed the discussion though, so it might not have happened after all. [16:06] astraljava: :-) [16:07] Jeremy something, nick JFo. [16:08] jeremy@ubuntu.com [17:46] holstein says his name is jeremy foshee [17:47] https://launchpad.net/~jeremyfoshee