=== udsbotu changed the topic of #ubuntu-uds-community-1 to: Track: Community | Testing ubuntu on new platforms | Url: http://summit.ubuntu.com/uds-1303/meeting/21669/community-1303-quality [14:55] hello [14:55] hey everyone [14:55] smartboyhw, \o [14:58] #ubuntu-uds-community-2 [14:58] hey everyone.. this should be a fun first session [14:58] #ubuntu-uds-community-2 [14:58] oops [14:59] baloons please give me updates of session as I'm on mobile and not enough bamdwidth for Hangout [14:59] jono_, jajaja [14:59] smartboyhw, ahh [15:00] jono_ what's room 2 on about? [15:00] zyga-uds, mmrazik u guys wanting to join in the hangout? [15:00] balloons, we're going to have a fun session :D [15:00] anyone else? trying to make sure we have proper folks in the session [15:00] balloons: yes [15:00] what are the proper folks ? :) [15:00] balloons: I'm interested to know how this overlaps with hardware certification [15:00] I would lov to if I had my computer here............. [15:00] lol [15:00] balloons: I can if there are enough hangout slots but I don't have to [15:00] eh? [15:01] dholbach, balloons hi! [15:01] mmrazik: I think we have enough... [15:02] uhoh zyga-uds left [15:03] plars, ! what a beard !!! [15:03] lol [15:03] ogra_: :) [15:03] So who's on Hangout? [15:03] smartboyhw, I'm on the left ;-) [15:04] where i can get hangout adress? [15:05] VencaCZ_, http://summit.ubuntu.com/uds-1303/meeting/21669/community-1303-quality-testingstrategy/ [15:05] ty [15:05] dholbach I"m on mobile how can I see? [15:06] smartboyhw, http://www.youtube.com/watch?v=D7LpRcllyYA [15:06] dholbach 's video feed disappeared [15:07] diwic, can you see the other people's feeds? === tomk is now known as Guest82130 [15:07] dholbach, just you [15:07] (are gone) [15:07] dholbach There was a problem while playing [15:07] ah, back [15:07] Touch to retry..... [15:07] dholbach is back [15:08] do you see whoever's speaking? [15:08] is that testing stack public? how can other developer use it? [15:08] dholbach, yup [15:08] I just momentarily stopped my video feed to try out something else [15:08] you don't need to see me ;-) [15:09] I'm not really part of this discussion :) [15:09] dholbach I can't get to it.... [15:09] smartboyhw, it's just a normal youtube video? [15:09] dholbach, but you have the funniest headphones ! [15:09] dholbach, ok, just made sure you were aware [15:09] diwic, thanks [15:09] ogra_, I like them :) [15:09] ogra_, shall I use these instead? [15:09] +1 exploratory testing [15:09] * ogra_ didnt say he doesnt :) [15:10] what kind of tests are you running? is that testing the software stack more or the hardware enablement? [15:10] hi [15:10] hi [15:11] dholbach: There was a problem while playing. Tap to retry [15:11] smartboyhw: it sounds like you are having trouble with your mobile and playing hangouts. there isn't really anything we can do about htat [15:12] yeah [15:12] * smartboyhw condemns his ARMv6 A [15:12] Samsung Android [15:13] zyga-uds, did your question get answered? [15:13] plars: where is the dashboard? [15:13] anyone else who wants to be part of this discussion? who wants to be part of the "fishbowl"? :) [15:13] balloons: I don't think so [15:14] zyga-uds: reporst.qa.ubuntu.com I believe is what he is referring to [15:14] dholbach: I would love to be [15:14] zyga-uds: http://reports.qa.ubuntu.com/ [15:14] reports too.. ^ [15:14] cjohnston: you don't know the URL yet? :P [15:14] I don't know how to spell [15:14] zyga-uds, sent you a link in an irc query [15:15] and I don't use the URL [15:15] took an external USB cam :) [15:15] mmrazik, might want to mute yourself while typing ;-) [15:15] mmrazik: mute yourself while your playing [15:15] dholbach: it works but I don't know how to attend, I guess I have to reload -- typing in private does not work [15:15] balloons: srry. It must be when I plugged the camera in. it has yet another mic [15:15] zyga-uds, does http://summit.ubuntu.com/uds-1303/attend_meeting/21669/ work? [15:16] mute the dog ;) [15:16] reloading didn't help, sorry [15:16] I lost the previous message [15:16] zyga-uds: what are you wanting? [15:16] we are still working on the views for memory and eventstat [15:16] those views are not in production yet [15:17] zyga-uds, ! [15:21] anyone else wish to join in the hangout.. or have questions? [15:22] balloons, well, i'D really like to know how we plwan to handle the android parts, they are an essential bit of phablet (no need that i'm in the hangout though) [15:22] u can pop in for a moment if you wish to ask [15:23] balloons, ogra_ would need to get dressed first :-P [15:23] lol [15:23] lol, no, i'm showered and dressed, but couldnt say much beyond that [15:23] where i can get source for ubuntu mobile? i want part it for my phone [15:23] port [15:23] I'd just like to say I really appreciate the community resources we have for testing; e g before releasing a new PulseAudio version into 13.04 we were able to have some community people testing it in a PPA. Thanks to everyone helping out with this! [15:23] VencaCZ_, in #ubuntu-tablet :) [15:24] ok ty [15:24] LOL, ok [15:24] someone paste me a link [15:25] hmm, the join button just hangs :/ [15:25] ogra_: are you referring to the fact that the phablet image currently uses Android stuff, and are we going to keep that in the image or remove it? [15:26] cjohnston, we will keep it [15:26] was taking a shot in the dark to see if I could interpret the question [15:26] * ogra_ canzt join it seems, sorry [15:28] ogra_, ahh.. well try rephrasing the question then ;-) [15:29] well, how do we make sure we get testing of the android layer in the phablet images, automated and by manual community testing [15:31] does the hangout still work? for me it just broke [15:31] broke here too [15:31] googgle hangout just broke [15:31] broke for me too [15:31] down here too [15:31] broke for me [15:31] bah [15:31] Yup [15:31] uh oh [15:31] same here [15:31] try reloading [15:31] works for me [15:31] yep, works again [15:31] whoa.. a bit a hiccup [15:31] reload it [15:31] Back [15:32] we're stress testing google hangout today! [15:32] lol [15:32] note that ubuntu just runs in a container on top of a minimal android [15:32] its not just the kernel [15:32] its a full minimal android [15:33] the complete HW stack is android [15:33] (way more than the kernel) [15:33] it is waht we currently have and will very likely also keep [15:34] we need that part since we dont have any other way of getting drivers [15:34] ogra_: do you want to try to join the hangout again? [15:34] gema_uds_, i tried over and over, the "join" button is greyed out [15:34] ogra_: you need to click on the acceptance button that you don't mind your image to be broadcasted [15:34] did you click on thata? [15:34] the driveras also require the userspces framework [15:35] gema_uds_, i didnt have that [15:35] * ogra_ tries again [15:35] bah ! [15:36] but the android kernel has not been tested with our user space, so it makes sense to test it [15:37] mmrazik, balloons^ [15:37] hey daniel :D [15:37] hi Limurx :) [15:37] cking, yes, i agree [15:38] testing all layers is good, testing each layer is good, in the end we'll need all kind of tests [15:38] but starting with the autopilot ones, testing all layers at least give us some guarantee that things will run for the user [15:38] devs will prefer to know which layer is at fault when something breaks though... [15:39] uds-community-1: 5 minutes left in this session! [15:40] uh oh [15:40] uds-community-1: 4 minutes left in this session! [15:40] vila, exactly, i just wanted to make aware that we need code for that ... i.e. we dont use the ubuntu input layer at all in the phablet images [15:40] are there virtual environments for those w/o HW to use for testing? [15:40] ogra_, muted you while typing [15:40] joe-uds, we will have a discussion about VM and emulator at 7pm in the SDK session [15:40] ogra_, perhaps a work items is for you to figure out what needs testing [15:40] dholbach, thanks [15:41] cking, sure [15:41] uds-community-1: 3 minutes left in this session! [15:41] ogra_: good point, I think someone mentioned that the autopilot session later will need to care about that [15:41] * ogra_ wont be able to attend many more testing session, thats why i'm here and thourgh i shoudl bring it up :) [15:42] *thought [15:42] uds-community-1: 2 minutes left in this session! [15:42] mmrazik, we might just build x86 android/phablet images [15:42] to run in any VM you like [15:42] ogra_: yup. might be one of the options [15:44] * joe-uds claps [15:44] uds-community-1: This session has ended. === udsbotu changed the topic of #ubuntu-uds-community-1 to: Currently no events are active in this room - http://summit.ubuntu.com/uds-1303/community-1/ - http://ubottu.com/uds-logs/%23ubuntu-uds-community-1.log [15:45] Gonna be a hot discussion right now ;) [15:48] * mainerror takes seat [15:48] These walks from room to room are exhausting ... === mitya57|uds_ is now known as mitya57|uds [15:49] dholbach: can you invite me to the hangout for community-1303-rolling-release? (assuming you're the one running it) [15:49] stgraber: if you are required then the link is available for you in summit [15:49] stgraber, above the video there should be link [15:50] cjohnston, dholbach: right, refreshing fixed it [15:50] Who is running this session? jono ? === udsbotu changed the topic of #ubuntu-uds-community-1 to: Track: Community | Should Ubuntu adopt a monthly cadence/rolling release? | Url: http://summit.ubuntu.com/uds-1303/meeting/21683/community-1303-rolling [15:52] FunnyLookinHat_, it's dholbach [15:52] YoBoY, salut mon ami [15:52] bonjour, comment ça va ? :) [15:53] dholbach, CarlRichell and I are here from System76 and we'd love to be pulled into the Hangout if there's room [15:53] YoBoY, stgraber nous parlons en français pout cette session? [15:53] très bien - et toi - il fait beau ici - beaucoup de soleil à Berlin [15:53] hi FunnyLookinHat_ [15:53] * philipballew feels language isolation [15:53] rickspencer3, Yo [15:53] rickspencer3, avec plaisir :) [15:53] dholbach, can you invite the System76 guys? [15:53] heya CarlRichell! [15:54] heya Jorge! [15:54] pgraner: Can you include me in the hangout for this one? [15:54] if you need more people to give a "loCo" perspective, I can join the hangout here as well. [15:54] dholbach, un peu malade, mais il fait beau aussi :) [15:54] dholbach: ^- or if it's you [15:54] Go ahead and invite Carl on Gplus [15:54] he's driving the webcam [15:54] cjohnston, https://plus.google.com/hangouts/_/f087d7d5109cb66b16cbbdb4b4c7153a846ee9f6 [15:54] Shame that this session has conflated rolling release and monthly cadence, which really have a bunch of separate issues associated - I'd suggest that people try to work separately [15:55] pgraner: do you mean me? [15:55] cjohnston, no cjwatson ^^^^^^^^ [15:55] ta [15:55] cjohnston, change your nick damn it [15:55] 1 2 3 tab [15:56] pgraner: no. [15:56] * mitya57|uds tries to fix his microphone [15:57] does anybody hear me [15:57] ? === timrc_ is now known as timrc [15:58] not currently mitya57|uds [15:58] ha people are already using Mir... why would you do that this early? [15:59] oh those crazy bleeding-edge folks [16:00] * mitya57|uds hopes he fixed it [16:00] does it work for you too? [16:00] if "it" == video, then not yet [16:00] i had to refresh [16:00] works now [16:00] video is great [16:01] perfect [16:01] video is great on this end too [16:01] finally got video :D [16:01] pitti/Laney - did you want to join?> [16:02] sure [16:02] o/ [16:02] dholbach: can you invite Laney? [16:02] ok, looks like it still doesn't work, so I'll just sit in IRC [16:02] jonobacon, yep [16:02] mitya57|uds: reload the page [16:03] how is this any different than os x that comes out every 18-24 months? [16:03] Windows often goes years between versions. OS X recently switched from new versions every 18-24 months to a new version every 12 months. [16:03] people marked as bold at the bottom of the page should have a link saying "join hangout" or some such above the video [16:03] CarlRichell: what would be your ideal cadence then? Six months still? Or something else? [16:03] gema_uds2: well, anyway I don't have much to say :) [16:03] How does this affect LoCo's who need to advocate the rolling release? Or should we be showcasing the lts release only? [16:03] mdeslaur: people can install new software on old OSX/Windows easily, not so much in ubuntu [16:03] hmm, I certainly wouldn't agree that every Ubuntu release has been better than the previous [16:03] mdeslaur: if you want a newer libreoffice for example, you're either in backport or PPA land. [16:03] Is it weird that I can actually picture adconrad happily frolicking in fields of time, money, and cheesecake? [16:03] jcastro_: that's a different problem, and one that will kind of solve itself if everyone is using LTS [16:03] mdeslaur: on other OSes that's just a quick download for users [16:04] they may be more fit to purpose at the time they come out due to changes in the requirements for software [16:04] but I don't think each Ubuntu release has been "better" [16:04] windows releases approximately every 5 years [16:04] dobey: I suspcet that may speed up now [16:04] System76Chi: are you providing security support beyond 18m cycle? or do you upgrade all of their users to next non-lts? [16:04] rickspencer3: we (in canonical) are working with HP,DELL and Lenovo PCs and Laptops only using LTS releases already [16:05] System76Chi: i don't want to buy laptop which will star saying it's obsolete in less than 12months. [16:05] I think if we have less releases putting more work into LTSs would be vital [16:05] so I am not sure OEMs do what quicker cycles for PCs [16:05] Windows Vista, 2006-2007: Windows 7, 2009: Windows 8, 2012. [16:05] specially thanks to the hardware updates in point releases [16:05] http://en.wikipedia.org/wiki/Os_x [16:05] mpt: Agreed, WIndows every 2 years. [16:05] OS X was every 2 until last year. [16:05] (scroll down for versions - release dates are ~2yrs) [16:05] albeit XP still has a huge chunk [16:06] rolling release would avoid version stagnation if updates are continuous [16:07] People don't care that things don't change, they care that they don't break unpredictably [16:07] Some different meanings of "stability": (1) error rate (2) amount of UI change (3) amount of API change (4) number of SRUs [16:07] snwh: the plans look like just dropping 6-monthly releases, not doing significantly more work on the development release than happens at the moment [16:07] so let's cut support for the 6 month release to 8 months? [16:07] Who will the rolling release be suitable for? [16:08] dmj_nova: developers, and tech enthusiasts [16:08] Why don't we release 13.04, cancel 13.10 and make it rollling for a year, then review and see of we are to implement only LTSes + rolling release forever? [16:08] mpt: do you want to join the session? [16:08] Would you ship rolling? [16:08] tumbleweed, oh yes I know :) [16:09] 2 years is *far* too long to go without a stable system with fresh software [16:09] dmj_nova, "software" is a vague term [16:09] dmj_nova: for some classes of users, not all [16:09] dmj_nova: what do you mean by "fresh" there? [16:09] in this context, at least [16:09] dmj_nova: this comes back to the disagreement about what "stable" means [16:10] is 2 years, really THAT much? People still stick to god know how old XP. [16:10] Would you recommend the rolling release in production environments (workstations, not servers) [16:10] I would not [16:10] no [16:10] no [16:10] no [16:10] I wouldn't recommend the 6-month releases in a production environment either [16:10] I only recommend the LTS [16:10] me4oslav: it can be a huge amount of time, especially with a fast-moving platform [16:10] I only recommand LTS in production enviironments [16:10] I only *deploy* the LTS on family machines that I have to support [16:11] I think the argument is that 12.10 wasn't completely stable in comparison to 12.04. With features you lose stability. That is all a rolling release is suggesting. [16:11] Already, we are having trouble supporting Precise [16:11] dmj_nova: what trouble, why? [16:11] (without backporting gstreamer, webkit, python, udisks2, etc [16:11] dmj_nova: hrm? [16:11] slangasek: the thing cjwatson is saying in the hangout right now [16:11] Already using "rolling releases" on different level in production environment [16:12] I also only recommend LTS, since most of the people I recommend Ubuntu to are used to Windows. They would be irritated by updates more regular than every 2 years. [16:12] CarlRichell: would you ship rolling ? [16:12] we have too many releases to support at once, so it's difficult to spend time both developing the new release and supporting all the existing releases [16:12] the alternative being to stick with old tech until the next LTS (and thus not track upstream at all) [16:12] Its to risky as somene recommending software to someone to recomend that is not lts for nots us, and the user. There computer is more likely to not funtion well with Ubuntu, and cause their Ubuntu experience to not be as good as it might be. [16:12] what about either a) reducing the support for the 6-month releases [16:12] 12.10 was buggy, and a daily iso will be buggy, but both are stable enough to use if you do not need complete stability, which you can only get in an LTS. [16:12] cjwatson: firefox is also self-contained [16:12] pmk1c: Windows ismoving to a yearly cycle [16:12] or b) using a different cadence for different "levels" of system? So maybe glibc is on a 2-year cycle, but Unity is on 6 months? [16:12] dobey: I was asking dmj_nova what trouble they were having in particular, not what trouble *Ubuntu* has had supporting the LTS [16:12] I'm well familiar with the latter :) [16:12] dmj_nova: you're working under the assumption that LTS will *never* get updates [16:13] slangasek: basically it was beginning to take about 3 man-days to backport every month === udsbotu changed the topic of #ubuntu-uds-community-1 to: Track: Community | Should Ubuntu adopt a monthly cadence/rolling release? | Url: http://summit.ubuntu.com/uds-1303/meeting/21683/community-1303-rolling-release/ [16:13] slangasek: ah, I read that as the royal "we" :) [16:13] shipping hardware with a dev release sounds crazy [16:13] dmj_nova: backporting what? [16:13] jdstrand: well. yes it mostly is now that we threw all the extensions out of the archive ... [16:13] for the ubuntu documentation (in my langage) it's also hard to support lot of branches, but I don't know how we will support a rolling release yet, seems tricky too… :] [16:13] (almost all, actually - there are still a few) [16:14] rickspencer3: I hope it's not yet decided whether we do the monthly release thing at all? (as opposed to monthly planning and install images) [16:14] in this case we were backporting gstreamer 1.0 and a couple other things [16:14] dmj_nova: I don't understand what your use case is that you need backports of "gstreamer, webkit, python, udisks2, etc" [16:14] tumbleweed: and yet, OEMS supported by Canonical did precisely that. in the *bad* old days [16:14] Google Chrome [16:14] achiang: not with daily updates [16:14] dmj_nova: the "old" tech is the /stable/ tech - why would backporting python, gstreamer, etc be a priority for you? [16:14] and also having to hold back on using useful new functionality from updated versions of say python [16:14] cjwatson: sure. point I was making that unlike unity, it is more or less a leaf application [16:14] rickspencer3: I think most devs want to run their own "trunk" may are more conservative with everything is that is not part of their strong dependencies [16:14] tumbleweed: i think we're agreeing? shipping hardware with a dev release is crazy [16:14] achiang: :) [16:14] * mitya57|uds agrees with infinity's mail — it's not a good idea to leave "monthly" users without critical updates [16:14] CarlRichell: Server users. :) [16:15] rickspencer3: getting updates from everybody at once may not suit everybody [16:15] achiang: but slightyl less crazy, if it' sa snpshot of a dev release, with security support [16:15] slangasek: there's a big difference between "latest stable" and "old" [16:15] i think "real users" want the LTS. they are generally annoyed by windows downloading updates and rebooting 6 times. [16:15] I'm on LTS because ATI drivers don't work with anything later right now. [16:15] so the rolling release would be our general release, and LTS is for those who need really really stable things? [16:15] they want stuff to work and to use their OS. [16:15] tumbleweed: not from my POV. LTS is the only thing that makes sense to support OEMs on from where i sit [16:15] they may also be adding their own software that has dependencies [16:15] jcastro_, I'm not sure "server users" should be using LTS though. [16:16] The more of an OS's user base is made up of institutions, the slower that user base upgrades on average, so the longer the optimal release cycle. [16:16] achiang: well, it comes with free security support from canonical... [16:16] LTS systems support schools, office buildings, etc. My university runs 12.04, and they cannot afford to change every six months, so they stick with the LTS. [16:16] Universities indeed love LTS releases [16:16] jcastro_, There are better tools available to address the issues of stability + managing updates ( a la Landscape ) that are much more elegant and efficient, right ? [16:16] Because upgrading a PhD student's laptop in the middle of their three/four/whatever-year PhD is REALLY BAD [16:16] I dont want people showing up to my Ubuntu hour complaining "Their Ubuntu is not working" and this is because they installed a dev release. Even the not lts release now is unsafe to the average user who does not want any problems. [16:16] does it make sense to find a middle ground on LTS period: perhaps yearly [16:16] FunnyLookinHat_: most server users I know want/need the 5 years of support of an LTS. [16:16] dmj_nova: I don't really feel that you're articulating why this difference matters in your case. I can't imagine ever taking a decision to backport a newer version of infrastructure like gstreamer and python instead of just using the platform that exists. [16:16] maybe we need a separation between the OS itself, and the apps such as LibreOffice...you want the OS to be stable, but you want to get the latest libreoffice [16:16] TBH, most of my family and friends is using LTSes, too -- much less hassle and potential breakage of what you got used to [16:16] Universities also liked the interim releases, because their user base of enthusiastic undergrads was clamoring for the latest software. [16:16] If ATI, Nvidia and others would keep their drivers updated, I would be good. But, I can't run my desktop without LTS. [16:16] That explains why mobile OSes have new versions more frequently, for example [16:17] (I've heard this specifically from University of Cambridge sysadmins) [16:17] the number, frequency and size of updates honestly make using any sort of ubuntu development release unreasonable. [16:17] mdeslaur: +1 [16:17] There's a whole class of users who are happy with the software they have and don't need anything newer. For them, it's easier if things don't change on them all the time. [16:17] mdeslaur: +1; [16:17] jcastro_, how many people do you know using 8.04 ? I know none. :) A two year upgrade schedule is more than reasonable for many use cases [16:17] mpt: sure but enterprise managed phones are upgraded every 18m. [16:17] mdeslaur: +1 [16:17] The problem with rapid updates is Update Manager which is stupidly slow, maybe use apt demon with indicators. And most importantly Libreoffice and a lot of other debs are just to big. [16:17] * achiang is still using 12.04 quite happily [16:17] slangasek: using the old gstreamer would mean investing significant development effort improving undeveloped old branches [16:17] mpt: or not updated at all. [16:17] over the course of any development cycle, we're probably looking at a 700M delta in downloads every week. [16:17] mdeslaur: I'm much more nervous about behavior changes in applications than I am in the base OS :) [16:17] Sometimes one of these users finds he needs something newer, and then he's happy to trade of for the latest release (or rolling release or whatever) [16:17] FunnyLookinHat_: plenty of people using 12.04 (and there's probably 8.04 around as well) [16:17] It would be like us developing in linux kernel 2.4 [16:17] that is absurd to believe is acceptable for any consumer level user. [16:17] (server or desktop) [16:18] Right, the university use case where behaviour changes are scary is absolutely about applications, not the base OS [16:18] My family uses interim releases ever since 10.04 [16:18] olafura, Update Manager aka Software Updater already does use aptdaemon. Adding indicator pixie dust wouldn't make it any faster. :-) [16:18] achiang, I tried to stay on 12.04 but running armhf lxc containers on Raring is so much better :) [16:18] dmj_nova: as novacut is not a stable application/system, and it's a seriously complex system, i think targeting the LTS at this point is probably just a poor decision for novacut [16:18] We update Firefox in lts, and all people generally do for computing these days i live in a browser. Especially people who do not need anything beyond an lts. [16:18] If your scientific application changes its syntax three months before you submit that's disastrous [16:18] Not quite sure what stability issues everyone is talking about. [16:18] I am using quantal currently, but it's already hard to develop raring from quantal — for me mostly because of python 3.2 vs 3.3 [16:18] dobey: yes [16:18] jcastro_, Yeah - I understand what you're saying... I guess my point is, the reasons that people take slow upgrade paths can be replaced with better solutions... we've stopped trying to reform the process and instead are using the same processes that we did in 1999 [16:18] * balloons pops in [16:18] stgraber: and victorp__ win the best backdrop award of this hangout. [16:18] However, as we are NOT targeting hobbyists [16:18] mpt my problem is that it's tied to the update process [16:19] dmj_nova: once novacut is an actual stable product, you will almost certainly want to only target the LTS [16:19] so we have to consider the cycle once we are ready for professionals [16:19] mpt I use apt-get, most of the time === tomk is now known as Guest45188 [16:19] olafura, if you want updates to be faster, there are a bunch of ideas at [16:19] It's more trouble for me to recommend something thats not a lts. [16:19] dmj_nova: you can't choose to support both a stable platform, and a constantly changing platform, without a cost [16:19] rickspencer3: good summary. +1 [16:19] We really ought to sort out the half-completed debdelta project [16:19] We will likely have issues with recommending Rolling release [16:19] What's the solution for ISVs who need to update software frequently, but also be able to target solid releases ( instead of a rolling / moving target ) ? [16:19] cjwatson: that'd be fantastic [16:19] i think that anyone "non-geek" probably hates updates [16:19] 6 month cycles are different [16:20] dmj_nova: we have the same problem with Ubuntu One. there is a cost to supporting multiple versions of Ubuntu, and we're currently paying that cost to do so [16:20] On server you might no change ubuntu release but you change i.e. ruby via rvm.. or drupal [16:20] I mean they generally don't feel any difference between an LTS and interim version. As the supporter I don't feel any difference either. [16:20] (Since people keep bringing up upgrade bandwidth volume issues) [16:20] on windows, on OS/X, on iphone .... [16:20] we could easily recommend them because they're at least predicable [16:20] they do not want their OS to change. its just an annoyance. [16:20] +1 smoser_ [16:20] smoser_: Or their applications [16:20] maybe hteir applications. [16:20] yes, there is some sort of line there. [16:20] Usually applications changing is even worse [16:20] and that is hard to identify. [16:20] Because they're the things you interact with directly and learn [16:21] (i do go into the app store and look for updates) [16:21] Do we need to move on? [16:21] dmj_nova: then novacut made a poor decision to build on an unstable platform, instead of the stable one. you need to find the balance between what your users need and what you want to build, and make necessary compromise [16:21] cjwatson: but it seems to be accepted on all other oses and devices [16:21] smoser_: to your point...one of the things that I hate about android is how my apps are always updating. [16:21] QUESTION: It seems to me that people fix the easy bugs as they come up but the hard ones wait for the freeze. I think that's human nature. Will the freeze time be extended for this to happen for the LTS releases? Will people have the energy (and patience) to fix difficult bugs that have accumulated over 2 years. [16:21] It's probably easier to manage support with well-defined releases [16:21] I hear more complaints about Firefox changing its UI du jour than just about any other software change [16:21] observation: the six month release schedule provides key milestones (String Freeze) that enable the translation community to organize their work [16:21] So I have never bought the idea of an OS/application divide [16:21] yes, like firefox deciding to embed its pdf viewing [16:21] Here at MIT, we did annual releases of our Ubuntu derivative on the .04 release. We couldn't stick with LTSes because faculty and students wanted new things every year. But at the same time, we cannot possibly support rolling release in a public environment because my users will freak out when the UI changes (it's taken us a _year_ to get people used to Unity from) [16:21] :) [16:21] but not the new bugs... :/ [16:21] Easier to train support people on an OS that is relatively static (or only getting security updates) [16:21] ok. so i'll admit i was wrong. both techy-geeks *and* real users hate updates. [16:22] so lets not suggest to anyone that devleopment release is good for them. [16:22] thats just absurd. [16:22] things like the cloud archive are a better solution than non-LTS releases for servers [16:22] You don't have a choice of not installing updates, since you have to for security updates. [16:22] Sticking with LTS *for development* means that when the next LTS comes around you've got months of work to fix things based on the massive changes [16:22] rickspencer3,you made a new meme [16:22] sounds like we have similar wives [16:22] rickspencer3: +1. your wife is a non-geek. [16:22] rickspencer3, and she's right not to change if she feels confortable with her mac :) [16:22] i have one of those too [16:22] So what do security teams base their updates on? [16:22] (instead of a constant upgrade to non-LTS releases) [16:23] people *need* to install their updates (the security updates). that is probably the single most important thing people can do security wise [16:23] smoser_, I'm a geek and hate change :) [16:23] mfisch: who hasn't :) [16:23] moved buttons seem to really upset wives [16:23] rickspencer3: completely right. many people don't like updates on their system [16:23] While not *everyone* will be comfortable with non-LTS among our customers, [16:23] dmj_nova: yes. like i said, as a developer you have to make a decision, and compromise [16:23] change is never a issue for me [16:23] My sister was using Oneiric when quantal was about out the door, and she was happy. [16:23] jdstrand: yeah, but they don't change behaviour (except for firefox) [16:23] Good point, with a rolling release, if you make sure not to introduce too much of a big change at once people are likelier to adapt without noticing. I think. [16:23] jdstrand: ideally, there should be zero visible change after installing security updates [16:23] The interim releass make it viable for some of our users to track our development more closely [16:23] jdreed: I understand you wouldn't want to support a rolling release; but if you just had to do a release every 2 years instead of 1 year because there was no 13.04 (15.04) to base on, what would be the consequences for your use case? [16:23] which is distinctly not the case for RR or dist-upgrades [16:23] pitti_uds_: I was referring to a comment in the video stream where people have a choice [16:23] Our innovations are the things users curse at when they upgrade [16:23] This feedback loop is *very* important [16:23] so by all means LTS is desireable. thats clear. from what i'm hearing here is that there is very little value in development release. no one (even many of you here in this channel) is actually going to use it! [16:24] jdstrand: ah, yes; I interpreted this as dist-upgrading every 6 months [16:24] I guess the point is that on the day one ships we want the newest software. But how to support security and bugfixes wise is an interesting question. [16:24] jdreed: I understand that users will want newer things every year, but maybe they don't actually need them? [16:24] pitti_uds_: well, of course I agree in principle. the firefox issue is contentious and we don't have to rehash it here. I think we are doing the right thing still, but admit there are problems [16:24] but out lack of a new libreoffice version that _finally_ unbreaks the documents they want to read is _also_ the things users curse at [16:24] there are already eevry day updates in the software manager today [16:24] I've got my non-techy parents in Indiana running the latest LTS. It would be awesome for me not to have to make a special trip in just to upgrade the release when the time comes. Rolling release would be great for us [16:24] slangasek: Then I have some faculty clamoring for a newer version of $foo, because it doesn't support $bar. [16:24] can someone identify a class of users who we think is going to use a monthly release? [16:24] ^this [16:25] pgraner: or more importantly, why at all? :-) (seriously, there hasn't been a satisfying answer yet) [16:25] smoser_, Canonical QA [16:25] We have no problem pushing back on people who want the latest and greatest "just because", but with evolving products, we get people who seriously need new versions [16:25] smoser_: enthusiasts? [16:25] ;) [16:25] Monthlies to provide sync points for people who want less frequent updates? Perhaps optionally to base security updates on, perhaps not? [16:25] -1 for updating monthly [16:25] My concern with monthly snapshots has three parts: [16:25] 1. I haven't seen a compelling use case for it to exist, other than "cool for enthusiasts". [16:25] that combines the hassle of upgrading with the cost of doing releases [16:25] 2. There are high development costs in creating and maintaining monthly snapshots, and user base fragmentation costs in advertising their existence. [16:25] rbasak: do you know anyone that fits into that class ? (other than Canonical QA)? [16:25] and makes both worse [16:25] "monthly" is just a label. it will allow people to quickly describe to others what they are using. [16:25] 3. If there are credible use cases, it is possible (depending on the use case) that we might be able to address them with the updates UI instead. [16:25] "normal" users never upgrade (or even update) because they're much more risk adverse. they're afraid something they love and or depend on will break, which outweighs getting something new that will be marginally cooler [16:25] indeed. monthly "supported" release makes no sense to me [16:25] we can't be sure there are no critical bugs in monthly snapshots [16:26] rbasak: the current thinking is that if there is a monthly snapshot, high priority updates will be pushed to it. others people will get in the next monthly [16:26] do transitions freeze on the monthly date? what if one crosses the border? [16:26] Why not reduce the support timeframe for the 6-monthly releases? [16:26] in fact, we even have critical bugs in stable updates (in chromium, for example) [16:26] i really think those people will very quickly decide "this sucks, too many updates" or "this is too slow". and move to LTS or rolling very quickly. [16:26] if the user is averse to change, why wouldn't the LTS be sufficient? [16:26] if you want user to decide, you need to have parallel versions of packages available for selection [16:26] as a user looking for support, i need to be able to tell system76 what i am using. [16:26] I'm not adverse to change, I want my computer to work. If updating requires rebooting and/or critical path breakages. [16:26] jtaylor: no freezes. [16:26] then I am not interested [16:26] rickspencer3: that's fine -- monthly planning cadence doesn't need to imply monthly releases [16:26] that needs a label, even if arbitrary [16:26] would a monthly release schedule even give developers enough time to change/improve anything? [16:26] dmj_nova: why would you not have that same feedback loop from users of a rolling release? [16:27] no one "outside the project" will use montly releases. [16:27] question : for an install, do we start with a LTS for everyone (like starting with a 12.04.2 and upgrade to have the rolling release up to date on the november 2013 for example) ? or do we have daily/monthly images ? [16:27] smoser_: say for example a development team who work on PHP but develop against 5.4. LTS only has 5.3. Then they can all install the same monthly and know they're developing against the same thing. [16:27] +1 rrnwexec [16:27] surely if there were monthly releases people could still grab one every six months if they wanted to [16:27] slangasek: We couldn't get our users on a rolling release [16:27] theDoctor: python 3.2 -> 3.3 got switched in 12 days. [16:27] surely you're just creating a label for the sake of it then [16:27] Custom update cycles are best suggestion ever [16:27] not for any extended testing [16:27] theDoctor: so yes. [16:27] smoser_: and same deal with deployment [16:27] * txwikinger-uds feels like being back at LFS [16:27] (The mailing list summary doesn't fairly cover monthly snapshots yet, because I haven't got to that part of the thread) [16:27] (and testing and CI) [16:27] People will probably select a monthly snapshot based on what it introduces [16:27] xnox: but people still keep discovering bugs / incompatibilities [16:27] rbasak: i dont know. i tihnk even then they'll just be annoyed at the speed of change. [16:27] if there's no extra testing, freezing, etc… then there's nothing special about calling a release on the last day of the month, the monthly release. there's no guarantee for the user, and if we have to do all that work, it's extra work for us, that doesn't really gain us anything [16:27] yeah but can't we get small updates like that through update manager? is it really worth a monthly release schedule? [16:27] xnox: without a freeze whats the significant of the monthly? [16:27] slangasek: Would you use a rolling release on a production workstation? [16:27] rrnwexec: This is the question I asked last week. How can a user tell me (or how can I tell a vendor) what I'm running on my system, short of "dpkg-query -W" [16:27] So though they may not update monthly, they can find a snapshot that includes whatever they want [16:27] jtaylor: transitions are staged in -proposed and aren't copied over to another pocket until they're complete [16:28] smoser_: they will be, but that's an unavoidable cost of not using the LTS [16:28] We need at least some users who are [16:28] rbasak: why would those people not use the lts. and put their new php package on it. [16:28] another way to identify a "build" is to checksum all the packages... and that would result in a LOT of versions ;) [16:28] smoser_: sorry, I mean web developers developing PHP, not developing PHP itself [16:28] thing is: everyone updates as he wants. I know a lot of people don't caring about updates even when they pop up right in their faces [16:28] Simply put: insulation for RR SNAFUs [16:28] rbasak: so i assert they'll use LTS. [16:28] with 6 monthly releases many are comfortable with that [16:28] dmj_nova: no, I've said I wouldn't. But I don't have any idea who your users are, sorry, I find you've been a bit vague about what exactly it is you're developing [16:28] The security repo would solve that [16:28] smoser_: what about per package rolling release? [16:28] Perhaps they just need a backport. [16:28] "You are safer on monthly snapshot than on the rolling release" is true only if we have tested monthly-to-monthly upgrades. [16:28] my expectation is that there /will be/ users of a rolling release [16:28] slangasek: so we target all rebuilds to proposed, but what stops them from going to out? [16:29] smoser_: where you flag the packages that you want to update on rolling release basis? [16:29] slangasek: We're developing a video editor for professional artists [16:29] dmj_nova: ah, nova == novacut? :) [16:29] smoser_: if you want the latest greatest PHP, then you can have it [16:29] that makes more sense [16:29] slangasek: yep [16:29] My question is, if I choose a monthly snapshot then decide I'd like to be on the cutting ege daily, do I have to then get a whole new iso? [16:29] mpt: yeah [16:29] utlemming, well, where I carry my delta on my own (quite possibly through juju or some other mechanism) [16:29] if we have to do monthly testing at all, whats the point of rolling ? [16:29] Everyone knows that the daily after the milestone has always been so much better! [16:29] mathor, probably just change tour sources? [16:29] mathor, I think you would just change the frequencing that you update [16:29] xnox, yes +1 [16:29] *your [16:29] jonobacon: what if Ubuntu went the route of creating a rolling release VERSION of Ubuntu? Ubuntu-rolling? [16:30] jtaylor: that question is backwards; *all* updates are only allowed out if they don't break the consistency of the release pocket [16:30] I agree.. if we're monthly releasing/testing, we're not rolling [16:30] smoser: that is a server perspective, but it could apply to the desktop too. Like always tracking the latest unity. But then again we do some of that with PPA's already [16:30] cjwatson: I would like to phase all package updates in rolling, such that one gets a dial that one can set to stable. [16:30] If we have to stick with LTS for both development and end users, we'll be having problems every time the LTS cycle comes around [16:30] dmj_nova: honestly, i think every app developer on every platform has this issue, they have a branch for $CURRENTOS and a branch for $NEXTOS [16:30] utlemming, right. some stack that you're interested in. you may want to track that. [16:30] anything else is noise and annoyance. [16:30] achiang: exactly [16:31] smoser: I think that sort of option would be compelling...and something I would use. [16:31] If we develop on rolling, then we have to either spend enormous amounts of work on backports (some of which may break other things on the LTS) or just not get any feedback [16:31] i think we need to think in terms of "labeling" rather than "releasing". if we have daily quality, then we can label at any interval. it's just a label. [16:31] rrnwexec: +1, similar to tags in DVCS [16:31] +1 rrnwexec [16:31] rrnwexec, I would prefer we simply don't label the "rolling" at all.. It sort of confuses the issue [16:31] dmj_nova: we *do* develop on rolling [16:31] or... with the update manager, selecting lts, non-lts, or rolling might be an option [16:32] xnox: Some, certainly [16:32] cjwatson: I'm confident we can do just that [16:32] what is with transitions for packages not having a proper versioning scheme, like python? [16:32] dmj_nova: well, I guess I don't see how this is different from development of any other software where you need to support a range of versions on different platforms [16:32] rickspencer3, its reality. [16:32] your wife will not use it. [16:32] cjwatson: would it make sense to start using the priority field for the cadence bypass? [16:32] @smoser - True story [16:32] That's just not true. We use non-LTS in production as well. [16:32] jtaylor: ? [16:32] very important point rick spencer [16:32] PROPOSAL: don't make any changes for raring, it's very short notice considering we all agreed to doing raring and there's obviously issues still to be worked out [16:32] rickspencer3: "people outside the project will use the LTS" == "rolling is for developers"? [16:33] cjwatson: the beauty of sticking phased-updates logic into update-manager is that if told, apt will resolve additional dependencies needed to pull in a particular phased update. [16:33] slangasek: I'm saying that by not having the interim release, all our users are forced to the LTS [16:33] rickspencer3, those other frequent-landing software organizations are usually making apps, not platforms. (Facebook may be an exception.) [16:33] dmj_nova: if 12.04 and 12.10 have different versions of gstreamer, and you intend to support both, your software needs to be written in a way that works with both... rather than backporting gstreamer [16:33] Riddell: +1 [16:33] which makes it difficult to support both the LTS and the rolling with current development [16:33] Riddell: +1 [16:33] it's not about quality (although we aren't anywhere near LTS quality with raring yet), but about what you want to use [16:33] How many other OSes do rolling releases? [16:33] rickspencer3, if we really do it like that, whats the point in monthly image or monthly updates ? lets drop that stuff if every daily is installable there is no need for montly stuff [16:33] rickspencer3: the problem here is that our apps are too tightly tied to the platform. no other OS does that [16:33] maybe montly upgrades need to be automatic? [16:33] timrc: arch. [16:33] slangasek: that's a huge amount of work, and would entail huge differences in functionality [16:33] cjwatson: +1 [16:34] (the release cadence, that is, not the APIs) [16:34] xnox: it already has phased-updates logic :) [16:34] ogra_, exactly.. make it work daily, only care about keeping everything working day to day [16:34] +1 to Riddel and to ogra_ [16:34] cjwatson: naughty ;-) [16:34] balloons, ++ [16:34] rickspencer3 if we are going to have a rapid update them we need to allocate disc space that can be used if the disc is full. If we try to update when it's almost full then it's a mess. [16:34] I used to do that with Ubuntu. I used 6-monthly releases, but not every one. [16:34] dmj_nova: it's not terribly hard to support both; at least in Python. it's much harder in other languages. [16:34] Rolling releases are very agile.. but we also need the other site.. the automated testing for that [16:34] stgraber: Maybe - I haven't thought much about the priority suggestion yet [16:34] gstreamer is lucky in that it doesn't have issues with installing both side by side [16:34] maybe cut support for raring to 9 months? [16:34] Would making changes "destroy potential quality"? [16:34] CarlRichell: rickspencer3: what if 12.04.3 is better than 12.10? [16:35] dobey: there's all sorts of gstreamer issues with python [16:35] rickspencer3, you should take the people using Ubuntu into consideration... [16:35] cjwatson++ [16:35] especially the people paying money for support and what not [16:35] xnox: It still only has features that are a year old. [16:35] dmj_nova: gir vs. static bindings problems/ [16:35] We have spent three months on this thing. It seems harmful to all the sudden change now? [16:35] ? [16:35] How about transitioning raring directly into the rolling release? ("We will support this until the first RR snapshot happens") [16:35] dobey: yep [16:35] and python2 vs 3 [16:35] or don't release Ubuntu Unity for 13.04? [16:35] isantop: it has 12.10 kernel and X stack. So in 12.04.3 we can push out latest unity. [16:36] we were actually depending on having *both* python 2 and 3 because of gstreamer 0.10 [16:36] dmj_nova: 2 vs 3 should be less of an issue. it's only an issue if you're using static bindings [16:36] It's just such short notice. Problems with waiting are well taken, but seems necessary. [16:36] 13.04 should be released as scheduled, but it doesn't have to be supported [16:36] Isnt "there isnt a 13.04 - surprise!" merely a psychological problem? [16:36] BigWhale: nope should be supportd 18 months at least 13.04 [16:36] Sweetshark: Yes, a big one. [16:36] how about reducing the support duration for 13.04? 6 months? [16:36] PROPOSAL: if we won't release raring, can we release 12.10.1 / 12.10.2 / ... to have something to point people to until 14.04 is released [16:36] xnox: mvo was the naughty one [16:36] dmj_nova: :( [16:37] he did it ages ago [16:37] cjwatson: hehe. [16:37] Would allowing and testing 12.10-to-14 upgrades be noticably harder than 13.04-to-14 upgrades? [16:37] mitya57: we'd have to extend 12.10's life by half a year, which seems a lot cheaper than supporing 13.04 as a whole [16:37] dmj_nova: well, if it's any consolation, we have the same issue in U1, because of apport [16:37] ev: apt has phased updates support thanks to mvo tinkerings ^ =) [16:37] Shortening the maintenance period could be a good alternative to rolling [16:37] mitya57: what would tha tmean? that sounds like an entirely new thing [16:37] ev: grep cjwatson/xnox. [16:37] You get a 13.04 that gets constant updates and that you can get a week earlier even without feeling guilty. [16:37] you could release 13.04 now if you wish.. upgrade to raring (which is continues to roll as the rolling release) [16:38] dmj_nova: also, iirc, novacut is targeting python 3.3, which isn't on precise anyway, so trying to support precise seems like a problem regardless [16:38] doesn't that meet your need to release something? [16:38] why not just support it for less time? [16:38] perhaps a 7-12 month period for interim releases going forward [16:38] rickspencer3: just a thought, but if the main issue is the number of releases you have to support at once, then actually the release frequency doesn't affect that [16:38] dobey: yes, precise was becoming a problem for us [16:38] cjwatson: If you're going that close to canceling it without finishing it, just finish it off [16:38] which would means that RR cannot equal "rolling release". a small price :) [16:38] tumbleweed: currently the download page suggests 12.10, if it won't change until 14.04, people who have just installed Ubuntu will either use old and unsecure package versions [16:38] How will users be expected to keep that straight? [16:38] the thing that affects that is (release support time / release frequency) [16:38] skellat: I'm not sure I agree [16:38] or have to install 1GB of updates [16:38] transition period is simple enough.. if you running 12.10 you move to 13.04/rolling.. if your on lts you stay on lts [16:39] Security and stability updates could be "forced"/suggested daily and other updates could be hold back for a custom amount of time (default 1 month). So there are no snapshot to snapshot updates, that have the drawbacks of support, and no annoying bunch of updates each week [16:39] without raring, we would have zero releases we could recommend to our users [16:39] it gets weird if you release 2 "stable" versions [16:39] balloons: 12.10 itself still has 18mo support [16:39] so maybe you should fix the number of releases you want to support at once, and then play with frequency given that [16:39] Ah, pitti pointed out the error in my equation :-) [16:39] dmj_nova: but i don't think that's related to ubuntu having interim releases or not. it's more that it's just a bad time for writing something large like that in python 3, due to the transition to python 3. [16:39] cjwatson, rickspencer3, we offer opt-in milestone freezes for flavours, cant we offer an opt in release for then as well ? [16:39] dmj_nova: why can't you recommend 12.04.2? [16:39] slangasek, sure.. but we don't want to release 13.04 with the same support levels [16:39] sebsebseb, supported as releases are supported now, it would start rolling [16:39] mpt: you made an error :o are you sure ;) [16:39] *them [16:39] xnox: No, apt does not have phased updates support; update-manager does [16:39] ev: ^- [16:40] yes, we can recommend 12.04.2 in the download page [16:40] balloons: I mean that users on 12.10 aren't required to upgrade to the RR [16:40] cjwatson: ok. so i had it right. [16:40] pmk1c: the problems are the same. it's not a kernel issue. it's major API changes that aren't in 12.04.2 [16:40] The number of releases we're providing updates for is not a reason to introduce a rolling release. It is a reason to drop non-LTSes. (That's why I separated the two in the thread summary) [16:40] slangasek, ohh certainly.. so no one who didn't want to change/move would have to [16:40] pmk1c does 12.04.2 have up-to-date gstreamer, python, webkit-gtk, udisks2, etc? [16:40] ogra_: there's a differece between a milestone release and a final release, the archive gets frozen [16:40] rickspencer3: that makes sense. thanks [16:40] dmj_nova: no, 12.04.2 is not 13.04 :) [16:40] and if it does, how is it different from quantal or raring [16:41] so 12.04.2 does nothing for us that 12.04 does [16:41] mpt, that's actually a nice distinction, but a bit lost here [16:41] that is just the defaults [16:41] _you_ don't uninstall software, but apt might decide to. [16:41] Only if there's a conflict [16:41] it wouldn't have a reason to, though [16:41] CarlRichell: I guess that's where the new UDS model will kick in. [16:41] autoremove? [16:41] rickspencer3: the woudl have dobey's new player installed [16:41] pitti_uds_: but indeed also keep RB [16:41] Each of (a) dropping non-LTSes, (b) rolling release, and (c) monthly snapshots, could be done without either of the other two. [16:41] Also, say there are new features in rolling, we wouldn't be able to support them or benefit from them until potentially 2 years later [16:41] If you choose to use autoremove, then you need to review what it removes, certainly [16:42] underlying dependencies might cause uninstall by apt [16:42] dmj_nova: well, I think that goes right back to the point that I (and many others) would never deploy anything other than an LTS in production anyway, so the fact that the LTS is hard for you to support because it uses older technologies is orthogonal to the question of 6-mo. releases vs. rolling releases [16:42] dobey's new player? [16:42] I think we just need a smaller ubuntu-desktop-core without the default apps [16:42] dobey is not writing a new player [16:42] mpt: To me, rolling release is effectively synonymous with dropping non-LTS [16:42] Perhaps with better testing [16:42] cjwatson, Microsoft Windows is a counterexample. [16:42] ++ [16:42] mpt: ? [16:42] stgraber: i thought it was a plant and a mountain in a distance, but it's a folded desk! you just lost the backdrop award. [16:42] I mean "rolling release as we have discussed in the context of Ubuntu up to now" [16:42] cjwatson, of roughly-two-yearly releases without a rolling release. [16:43] Not some other hypothetical thing [16:43] slangasek: There are many contexts here, and it's not that we want *every* user on the current ubuntu or the current Novacut [16:43] xnox, lololol [16:43] dmj_nova: but for something like novacut, "in production" might mean something else - I might have a dedicated machine/VM for it, running whatever OS is required... whether that's a rolling release, an LTS, whatever [16:43] but with rolling, we couldn't recommend *any* of our users to use rolling [16:43] mpt: OK, sure, we could drop interim releases and also take all our intermediate development private. I don't consider that worth discussing. [16:43] dmj_nova: at some point you just have to stop supporting the older releases [16:43] Why not have a separate check box for optional software installation to not bloat systems? [16:43] If I uninstall the default software, it will be reinstalled with every update???? [16:43] dmj_nova: well, then that simplifies your support matrix too :) [16:43] dmj_nova, so you don't want LTS or rolling? [16:43] dmj_nova: but that's a decision the developer has to make, not that Ubuntu makes for them [16:43] dobey: we do not support precise right now [16:43] There's still an issue of stability with the rolling release - is the current daily essentially what we're calling "good enough" to be called a rolling release? Speaking from experience - it still frequently crashes, etc. [16:44] supporting it for 2 years would not work [16:44] jono_: You really want to call it crap? I mean a month before it was state of the art and the Ubuntu community said it was good. :) [16:44] cjwatson, that's a false dichotomy. There's a difference between the current development release model, and the rolling release model. [16:44] I'm pointing out issues that we are having and will have if the rolling release is chosen [16:44] One idea for Kubuntu to stay aligned to KDE releases is to do Kubuntu releases on LTS+PPA, would this be possible? (currently it goes against tech board policy) [16:44] mpt: Name it [16:44] 14.06! victorp__ is slipping our release schedule [16:44] And it isn't the presence or absence of LTSes. [16:44] cjwatson: can you point me to the session? [16:44] xnox: hahaha, I thought it was a bridge [16:44] with some mist [16:44] and a tree in the distance [16:44] mainerror, lol [16:44] stgraber: so deceptive! [16:44] slangasek: didnt say it was the first one ;) [16:44] Because Ubuntu needs to consider what its release pattern means for the ecosystem [16:44] dmj_nova: the issues you're pointing out will exist regardless of whether there is an interim release every 6 months, or not [16:44] victorp__: Not off the top of my head, sorry, I was just fairly sure it was on the appdev track somewhere [16:45] I don't know, FunnyLookingHat, I've been running 13.04 for two months and haven't had any bug at all? [16:45] Ubuntu is not valuable becuase it's an OS for people to use Unity on [16:45] mathor, but it's completely relative to hardware, drivers, specific installed applications, etc. [16:45] It exacerbates the existing problem. [16:45] So in that case, there's no reason for me to sit on 12.10 for 6 months when things are running smoothly on 13.04 [16:45] Ubuntu is valuable because it is a great platform for application development [16:45] cjwatson, it's the degree to which we are able to recommend it to people who aren't Ubuntu developers. I think it was Colin Watson who said, "If it's for developers only then I would account that a failure". [16:45] mathor, the problem is exactly that - You, as a dev, could be not having any issues, but the rest of us could have apport throwing errors left and right [16:45] and its these applications that make Ubuntu valuable to users [16:45] And because we're regular "users" we're not constantly filing bugs ( as we should be ) but instead are just getting frustrated [16:45] mpt: That's a difference of intent, but there is no difference I'm aware of in the underlying model [16:45] cjwatson, mpt is correct. We could simply switch to only releasing LTS's, without rolling.. We are effectively doing both here. also, we could drop lts's and only do rolling.. they aren't tied persay [16:46] If Ubuntu hurts this development ecosystem, that hurts the platform [16:46] dmj_nova: but is novacut going to migrate to Qt/Qml now, because Ubuntu has decided to do so? [16:46] balloons: I disagree that there is a real difference; I also fail to see how this is more than an academic discussion anyway :) [16:46] cjwatson: I will keep my eyes peeled ;) [16:46] cjwatson, I agree it's academic.. not really a discussion point here [16:46] dobey: We may, particularly qtwebkit but that's another matter [16:47] and doesn't impact this disccusion [16:47] mpt: To clarify, I consider it a failure if our current development release is for developers only [16:47] cjwatson, well it has consequences in the degree to which PPAs are used, or a new -experimental pocket, or both. But I'm here mainly because I'm interested in the things that aren't the underlying model, e.g. how we talk about the release to people who aren't developers. [16:47] mpt: So your attempt to use my words to prove your point fails, I'm afraid :) [16:47] the point about fragmentation of targets of development also applies to translation. they may end up focusing on LTS leaving intermediate releases irregularly translated. [16:47] LTS [16:47] GNOME guys were also going to request changing that policy, IIRC [16:47] Laney: we do have a PPA for 12.04 now and it's not much work [16:48] (Specifically, I've been working for well over a year to make sure that our development releases are usable by reasonably technical users who are not Ubuntu developers) [16:48] LTS should just be truncated to S (for supported vs not-supported) [16:48] stgraber: is this actually a TB policy, or does the issue come back to Canonical trademark policy? I'm not clear on this [16:48] I just do not understand why a user cannot use an LTS in the case that they need their system to have complete stability. [16:48] rickspencer3: perhaps it's worth discussing providing policy to allow for exceptions for the release+PPA situation [16:48] mmm [16:48] cjwatson: translations, documentations, making images fit the size, the "6-monthly" tasks. [16:48] It works if KDE have a strong policy about bumping deps [16:48] slangasek: it's some mix of both IIRC, it's a TB document defining what flavours can do, what remix can do, ... and the actual use of the name is covered by the trademark policy [16:49] mathor, because, at this point, they'd have 2-year-old applications ( for the most part ) and any ISVs might be saying "you're on the old version, too bad" to their support requests [16:49] translations requires string freeze, would that exist in rolling elease? [16:49] Laney: well we either backport the dep or drop it in that case [16:49] A rolling release is for those who would like to be on the cutting edge. Those people will not have the most stable release, but it will be usable [16:49] xnox: None of those differ between "just drop the non-LTS releases and otherwise keep our current development model" and "rolling release" - again, I consider them synonymous [16:49] slangasek: so last time the TB got a proposal for a policy change, we reviewed it, voted on it, then someone had an action to get the trademark policy updated accordingly [16:49] In that case, why not drop support for old interim releases much shorter? [16:49] :) [16:49] My problem with your argument. Is that 12.10 gave me apport errors out of the box on release day [16:49] kyleN, i think string freezes would be closer to a package level than system level [16:49] slangasek: (I believe that was for the Ubuntu business stuff) [16:49] If anything 13.04 was more stable than 12.10 [16:50] slangasek, for the panda images (that shipped a PPA installer) i was asked to either call them remix or drop the PPA installer by the TB [16:50] What is the point of releasing software if it isn't completely stable [16:50] mathor, + [16:50] mathor, That's because they forgot to turn off apport ;) [16:50] dmj_nova: next month there will only be 1 interim release in play, assuming 13.04 is not released [16:50] dholbach: is that you editing the notes? :) [16:50] dobey: And if the window were shortened, then we'd never have more than 2 [16:50] It seems clear that the TB would have to update some of those kinds of policies to keep up with events [16:50] One way or another [16:50] rickspencer3: +1 :) [16:50] yep [16:50] thats why we have proposed repos [16:50] Good luck rickspencer3 [16:50] maybe we need unittest coverage to report that strings have changed [16:51] QUESTION: what do we do with beta and RC releases from upstream? [16:51] what happens to release stability when we start auto-import from debian when debian unfreezes? [16:51] ogra_: that's indeed how things currently work, but I'm happy to rediscuss this and change policies if that's what we need for the LTS+rolling scenario [16:51] e.g. a KDE beta? [16:51] If you're adding new code (or strings) in the release you _have_ to make sure it works. [16:51] CarlRichell: what is interesting to you? is python3.2 vs python3.3 interesting? or only new unity is interesting (leaf packages)? [16:51] in practice, I don't think we have a very good workflow today for getting translations in /before/ the upload [16:51] dmj_nova: I don't think you can support interim releases for only a half year. that would mean forcing the update. But I think 1 year should be enough support [16:51] I jus think if you release 12.10 or 13.04 on a computer, you will not ge the stability of 12.04, and those cases LTS works best for users, and I would think System76, Dell, HP, etc should only ship 12.04 in that particular case [16:51] xnox: python 3.3 is very interesting to Novacut [16:51] kees: I know ScottK has raised concerns about that. I'm not sure I think it's as big an issue as all that, but neither of us really have data [16:52] kees: it breaks, what else? :P [16:52] rickspencer3: we'll have to have some kind of staging areas for transitions, betas, etc. anyway [16:52] dmj_nova: available or default? [16:52] pmk1c: why not support it for 1 year [16:52] ogra_: heh, calling it a remix if it's using a ppa is explicitly contrary to the trademark policy anyway [16:52] However, from past experience dealing with stabilising after auto-syncs, I genuinely think that most of those should have been addressed by proposed-migration [16:52] stgraber, ++, sounds like a valid solution for flavours [16:52] mpt: yes [16:52] so the user must update to 13.04 sometime before 13.10? [16:52] Riddell, fwiw: My plan for LibreOffice is to release ~alpha/~beta to ppa and only bring it to raring on .0/.1/.2 depending on quality/feedback. [16:52] dmj_nova: yeah as I said: one year of support should be considered [16:52] slangasek, it was only an installer icon on the desktop that enables the PPA , not even a PPA in teh build :) [16:53] xnox: we don't care if there are several pythons but we need 3.3 [16:53] ogra_: heh, ok [16:53] dmj_nova: we are planning to extend 12.10 support length to get overlap with 14.04. [16:53] dmj_nova: point taken. [16:53] xnox: We're already dropping 12.10 [16:53] xnox: that's been proposed, FWIW I wouldn't say that it's an agreed plan quite yet [16:53] It seems like we're going in this direction with a lot of unknowns - and if it hits the fan, we haven't necessarily planned for an "eject" button of sorts . [16:53] Sweetshark: also, you still have ppas to evaluate how good/bad the LibreOffice release is beforehand [16:53] xnox: that wouldn't be necessary with a 13.04 release, though [16:53] I just worry that the "rolling release" topic comes up every time debian is frozen. as soon as debian opens the flood gates, it's going to be a lot of constant pain, I worry. I hope it's okay. I'd prefer a rolling release. [16:53] how are rolling releases going to affect PPA building? Will I have a choice to build a package against: daily, monthly, or LTS? [16:53] cjwatson, ^ [16:53] I used debian unstable day-to-day before ubuntu existed. [16:53] slangasek: didn't you have some numbers on this topic (Debian unfreezing and its affect on the monthly release)? [16:53] (though only because at the moment we *don't* want users playing with a couple dangerous changes until 13.04) [16:54] timrc: you can build against LTS or rolling. No other options. [16:54] timrc: That depends. Don't know yet. [16:54] (this is a rare one off situation) [16:54] slangasek: maybe they weren't metrics :) [16:54] How much of main is synced directly from debian? [16:54] xnox: It's not decided yet. Other discussions this week affect that. [16:54] Again, FunnyLookingHat. Just keep users on the LTS [16:54] mdeslaur: chart at the bottom of merges.ubuntu.com/main.html [16:54] kees: yeah, I'd also prefer a model where we sync from "unstable minus 5 days" and check for new Debian RC bugs [16:54] mdeslaur: check merges.ubuntu.com it has graphs =) [16:54] With the idea of the rolling release, I don't feel like a little breakage (should it happen) is unacceptable [16:54] jdstrand: not on that specifically... I had numbers on how many packages on the desktop CD are unmodified from Debian [16:54] kees: as a kind of middle ground between unstable and testing [16:54] oh duh, yes, the graphs :P [16:54] Riddell, Do IT ! [16:54] Generally, we like to support at least the development and the current interrim [16:54] err [16:54] (Bear in mind most of "local" is language packs) [16:54] sorry Riddell [16:54] rickspencer3: just drop 13.04 release and do it :) [16:55] rickspencer3, Do IT ! [16:55] we don't care about interrim releases from over a year ago in any way [16:55] rickspencer3, but forget that monthly insanity ;) [16:55] Can we come to a conclusion on raring? === udsbotu changed the topic of #ubuntu-uds-community-1 to: Currently no events are active in this room - http://summit.ubuntu.com/uds-1303/community-1/ - http://ubottu.com/uds-logs/%23ubuntu-uds-community-1.log [16:55] only ship LTS [16:55] the reason I switched to Ubuntu originally was because of the release cycle. if the goal is just "switch to LTS releases", that's different. [16:55] Laney: indeed. "is there a feature freeze on thursday?" [16:55] So what's the discussion result? [16:55] cjwatson: But we can't ship a two year old release. [16:55] fnar bot [16:55] the plan B should be for system76 to ship LTS, not a rolling release [16:55] ogra_, I dunno, I think the monthly cadence is good for users who like Interim releases today [16:56] isantop: As it stands now [16:56] rickspencer3, they should ues the well working daily [16:56] didrocks: right, this is why I want a armhf builder for the LibreOffice ppa. to have at least one build see some enduser testing before I hit raring. [16:56] cjwatson: why not a simple dual-boot LTS + RR [16:56] Laney: please :/ [16:56] 4 minutes left until plenaries! [16:56] vila: That's not realistic for ordinary users [16:56] Sweetshark: need help with that? I know who to bribe :) [16:56] ubuntu needs to find a solution to continue updating software in LTS [16:56] bah [16:56] seems not [16:56] didrocks: nah, seems mostly ok now. [16:56] Good session guys! [16:56] cjwatson: rolling releases with snapsshots? === wedgwood is now known as wedgwood_away [16:56] dholbach: intersting convo [16:56] Sweetshark: then, you can have pocket copy :) [16:57] mathor: this is part of that solution [16:57] * ogra_ gets a break because rickspencer3 said so (my loung cancer is all your fault !) [16:57] didrocks: you rocks [16:57] Laney: that's up to the tech board n'est pas? [16:57] rickspencer3: excellent session, thanks all [16:57] ogra_, nooooo [16:57] cjwatson: may be. But 1) may be made usable for ordinary users 2) Using RR is for devs right ? [16:57] hehe [16:57] Riddell: well, the support situation is a Canonical decision [16:57] txwikinger-uds: Maybe, but I think those are broken in other ways [16:57] thanks guys [16:57] *claps hands* [16:57] bye [16:57] vila: I would rather make 1) usable and sufficiently up to date that it doesn't make everyone go "oh my god it's so old" [16:57] rickspenscer3: can we get an updated doc on what was decided/changed? [16:57] In the case of our users, it is important not to have updates break things at unpredictable or unfortunate times [16:57] Released two years ago doesn't fundamentally have to mean that it looks stale [16:58] utlemming: no decisions that I saw [16:58] Riddell: Strange, that [16:58] Riddell: I wonder if and when we'll know an outcome [16:58] Laney: so, did we answer any of the open questions that were in the whiteboard? :-) [16:58] cjwatson: it does mean we need to move faster with HWE, though [16:58] Rick did say at the start of the session that he didn't expect any actions from this session and was mostly trying to gather feedback [16:58] So upgrading every 6 months on average isn't a problem so long as they can choose when the potentially breaking changes happen and test them on a non-production environment beforehand [16:59] I wonder if rolling releases only work when you have different streams [16:59] Riddle: that is what I thought...but people react emotionally. I would rather I react -- and people react -- to a document then what they think they saw. [16:59] er, Riddell: ^ [16:59] skellat: this isn't the best forum for decisions, little notice and timezone dependent and it's up to tech board (or bits to canonical) anyway [16:59] cjwatson: indeed... I was hoping we would at least get a consensus wrt what to do Thursday for the FF === wedgwood_away is now known as wedgwood [16:59] But updating every month, could be a big problem, especially if there isn't much choice there about the timing [16:59] that's the only reason I wanted this session... [16:59] say 3 days before a deadline and no way to roll back [17:00] slangasek: if no decision to change has been taken, nothing has changed :) 13.04 should be released ;) [17:00] vila++ [17:00] slangasek: Well, none of mine about what to do in the immediate future [17:00] Riddell: Well, off to the showers for me and and I've got a couple hours until next round [17:00] this week and for raring [17:00] my preference is to carry on as planned and consider all this thereafter [17:00] Laney++ [17:01] slangasek: I'm all for RR myself which is why I would rather delay it for a few months that fails by rushing (but that's just MHO) [17:01] Laney: +1 [17:01] Laney, Riddell: yes, I think there is a consensus to carry on with the release schedule as written for now [17:01] slangasek: I think so, but I would like to know that it will be a useful use of time [17:02] though I don't think that necessarily means we'll still do a 13.04 release [17:02] Laney: indeed, and I don't think we can answer that until e.g. the TB have weighed in [17:02] not that we're going to review exceptions for a month and then have the doors thrown open again [17:02] I think there needs to be another blueprint about this topic [17:03] Developers are still torn on this issue, and so much time was spent arguing for a rolling release, that it was hard to even really discuss abandoning 13.04 [17:03] We started to have a discussion about this and I *think* people were receptive to keeping 13.04, but then we moved on [17:03] I think rolling release is a bad idea [17:03] Rolling is bad for developers [17:03] We already effectively have all the rolling we need thanks to daily quality [17:04] slangasek: ↑ - don't know if there will be an opportunity to finish it off some other time this UDS [17:04] but we can't reach end users as easily without the interrim releases [17:04] And is bad for bussines. Enterprises want stability sensation, and LTS versions are the best way to do that [17:04] * Riddell objects to the term 'interim release' it only got created to make the idea of dropping them seem more favourable [17:04] lol [17:05] yeah [17:05] Riddell: non-lts release? [17:05] Laney: I suspect that may require some Canonical-internal discussions that are out of scope for UDS, as far as what resources Canonical is willing to put into carrying 13.04 forward [17:05] Riddell: how were they name before ? dev releases ? [17:05] same feeling dmj_nova [17:05] vila: just release [17:05] Yes call them dev releases [17:05] slangasek: Hmm, yeah [17:06] I at least get the feeling that it's on the table which is positive for me [17:06] smartboyhw: they're used by much more than devs [17:06] ph [17:06] *oh [17:06] Much of this can come up at the Xubuntu Contingencies discussion tomorrow [17:06] Most users I know do not use the LTS [17:06] rickspencer3: Cant we 'release' 13.04 with an announcement as a rolling release? That is: announce it as being the rolling one and pitch the awesome fact that you even get quicker updates for it? Essentially its a policy change only: instead of SRUing stuff, we are trying to make every distro upload a good one -- not only the stuff we SRU back. [17:06] they do use a stable release [17:06] Riddell: mixed feelings on that, I have a few friends to which I had to say: can't help you with bug X, revert to LTS. They did so and were happy [17:06] but it's generally the regular most recent releast [17:07] Most of those would not use rolling [17:07] Sweetshark: + 1 on the idea, the wording needs to be better though [17:07] My university will only use LTS, and a full-scale deployment of ubuntu on pre-installed machines, for instance, if my university or another university or an office building were to choose to purchase a computer preinstalled, this is what they'd want as well. Just as rolling releases only contribute to enthusiasts, interim releases only contribute to enthusiasts as well. [17:08] One nice buffer with the 6-month cycle is that in most cases (and the others are arguably bad judgement) there is time to discuss and refine new big changes [17:08] Stand up users maybe, but the 2 SME's I know are using LTS with a paid support to canonical [17:08] Sweetshark: i.e. use 13.04 to experiment with RR and decide how to continue from there [17:08] so that only the devs get bombarded with controversial changes [17:09] vila: developers dont lie ;) this is why you keep them away from the mics [17:09] whereas a monthly pulse could result in upsetting changes hitting everyone before there has been an opportunity for refinement [17:09] Sweetshark: hehe, yeah, they are notoriously bad to communicate with users ;) [17:09] and good community feedback [17:09] the Ubuntu On Air, and what they are trying to do with that sort of addresses this issue [17:10] if it becomes a case where talks such as this are happening every day or once a week, there is no point for a UDS, it leads to a rapidly-changing ecosystem [17:10] I think that the 6 month rhythm is very good for desktops [17:11] * smartboyhw agreea [17:12] so if anything, the landscape has become more transparent than it was === sg is now known as Guest66646 [17:14] So a rolling release paired with constant community feedback leads to development at a much quicker pace [17:14] Rick Spencer: Does anybody include professional users? [17:15] rickspencer3: ^^ [17:15] aw man I'm too late. [17:15] dmj_nova, not sure what you are referring to [17:16] rickspencer3: as in people who depend on their machine for work project who are not developers [17:16] dmj_nova, I guess I would say if a user prefers to update eveyr 6 months today ... [17:16] the rolling release should be good for them [17:16] if they prefer the LTS, they can/should stick with that [17:17] mathor_, yes, we still have the worst aspect of UDS, i.e. "let's discuss everything all at once" [17:17] rickspencer3: One big difference is how can the user choose when the big changes happen [17:17] or would you support say the last three monthlies? [17:17] dmj_nova, as we discussed ... [17:17] which doesn't sound like it saves you much [17:17] there are 2 things [17:17] 1. Rolling release [17:17] 2. Update cadence [17:18] rolling release means that we only support LTS and the latest and greatest [17:18] I proposed that we also have a monthly update cadence, so users can choose to update monthly [17:18] mathor_, but there's nothing to prevent the lead of any particular feature saying "that's daft, I'm going to organize a hangout on topic X, and let interested people register what times would suit them" [17:18] but they *have* to update to get support [17:19] dmj_nova, ^ [17:19] rickspencer3: the tricky thing for many of our users is not that they need "stable" in the LTS sense, so much as updating several times mid-project [17:19] and what is wrong with that, mpt? [17:19] Now shortening the support cycle substantially for non-lts is certainly fine === wedgwood is now known as wedgwood_away [17:20] mathor_, nothing! I'm proposing that as an escape route. :-) [17:20] rickspencer3: wrt sunk cost, it sounds like many players would have an issue going forward with no raring release [17:20] dmj_nova, yes [17:21] there's been a bit of a spread in the reaction [17:21] Ubuntu could essentially get to a point where UDS is continuos with the all-web concept [17:22] Why not make the rolling a 6 month rolling [17:22] instead of the monthly [17:22] dmj_nova, that's a very fair proposal [17:22] and only have enough overlap to ensure people have a bit of time [17:22] we keep doing the interim releases, but we don't support them past the next one [17:22] say 1-6 months [17:22] dmj_nova, well, I think we would not do that [17:22] or it's not rolling [17:23] so each release gets 7,9, or 12 months , whichever time period to choose [17:23] to get support you have to update, I would think [17:23] or stay with the LTS, sounds pretty fair to me [17:23] rickspencer3: should be a tiny bit of overlap I think [17:23] could be [17:24] to give a bit of leaway for updating [17:24] well ... I don't think we would want to do an SRU or security updates [17:24] with our current options, users are forced to upgrade every 6 months, which really means re-installing their OS. Upgrading from 12.10 to 13.04 is not an easy process [17:24] so you have a bit of time to get eggs in order or get through the deadline next week before rocking the boat [17:24] dmj_nova, I think if updating rocks your boat, you are an LTS user [17:24] rickspencer3 how about 3-month rolling release? [17:25] smartboyhw, all good ideas [17:25] I see lots of people wanting lots of different frequencies :-) [17:25] rickspencer3: I'm talking very short term update qualms [17:26] well, then they could just not update [17:26] Almost like it's a job for a slider in a dialog somewhere [17:26] As mentioned, if the 3 month release needs more testing or other effort than the daily images, there's something wrong... [17:26] so you don't get surprises on a bad day, even when updating is not a general problem [17:26] mpt, it's not that easy [17:26] rickspencer3: what about security? [17:26] dmj_nova, that's my point [17:27] we don't want to duplicate support across relesaes [17:27] rickspencer3: I'm trying to suggest a happy medium [17:27] This is why LTS works. If you need stability and you do not like change, it just works. You cannot have your cake and eat it too, so you cannot have the latest and greatest and have the amount of stability to warrant shipping software on a system. LTS is the only good option for enterprise customers, professional users, institutional users, and OEMs in general [17:27] this is more of a Fireside Chat w/ jcastro than a plenary :) [17:27] reduce duplicate support a lot [17:27] dmj_nova, understood [17:27] rickspencer3, I know, we couldn't just update to latest if we want it to be more stable than the development version. But maybe we could update to latest-known-good. [17:27] erm [17:27] timrc: ask questions then [17:27] :) [17:27] but provide just enough to make life easier on certain kinds of pro users [17:28] mpt, it's a matter of providing security support [17:28] which is why I suggested the small overlap [17:28] like say a 7 month support cycle for a 6 month release [17:28] yep [17:28] rickspencer3, AIUI, the issue is that security updates will sometimes depend on non-security packages. First is there any data on how often that happens, and second would it matter? [17:29] and that means that for part of a year you do have an additional release to support [17:29] mpt, yeah, if we force an update of a subset of newer packages for a security update ... that is a terrible experience and very difficult to support [17:29] but your efforts are otherwise much reduced from current [17:29] while not rocking the boat for developers like us [17:29] dmj_nova, yeah, I understand what you are saying [17:29] rickspencer3, terrible in what way? [17:30] or vendors like system76 [17:30] mpt, you take a "security update" and it updates some of your apps? [17:31] then you may was well be on a daily upgrade cadence [17:31] rickspencer3, ah, so you'd have unexpected UI changes as a consequence of a security update. Understood. [17:31] I just do not understand the viepoint of a vendor releasing a 6 month release on their hardware. Why would System76 want to release ubuntu on a computer if it's obsolete in 6 months? [17:32] LTS makes the most sense [17:32] mathor_: 6 month releases have much better hardware support [17:32] new drivers, graphics stack etc [17:32] mathor_, from what I understood in the talk, it's because the improvement in big things like Unity and HW compatibility is worth the cost of users having to upgrade sooner. [17:33] That surprised me, tbh. [17:33] mpt is right [17:33] Yes, new drivers and improvement in the release drive our decision to ship non-LTS releases [17:33] mpt: so imagine using a release as old as Natty [17:33] But, essentially, what you are releasing is not the same as what they will have on their computers in, say, 9 months time [17:33] and then compare to Quantal or raring [17:34] mathor: same goes for LTS [17:34] so, at that point, they are running different software than was shipped with them [17:34] they get software so they can abandon it? [17:34] dmj_nova, I think that was a *bit* of an unfair example, because part of the reason for landing a ropey Unity in 11.04 was to give it more time to bake for the LTS. :-) [17:34] That is the nature of software. [17:34] It progresses. The hardware is static when manufactured [17:35] mpt: a *bit* but generally changes will be quite large and beneficial to users over a 2 year period [17:35] it is the nature of software, but goes against the argument that vendors needs a very current yet stable system. === FunnyLookinHat_ is now known as FunnyLookinHat [17:36] mpt: perhaps compare 10.04 vs 12.04? [17:36] mathor_, the current release schedule is very current yet stable - not sure I follow your argument ? [17:37] If we can't ship anything but LTS, that's not good [17:38] dmj_nova, Agreed - it leaves a gaping "middle ground" hole [17:39] the issue with hardware shipping is a big BIG deal for novacut [17:39] I'm not going to purchase a new Ubuntu laptop from System76 every 6 months, but I might every 2 years. If, as a user, I chose to get constant updates, then what I am running will be different than what I was shipped. If I want my system to remain secure because I am a professional user, then I will use the LTS and continue with the software I was shipped and remain happy. In fact, the LTS backports a lot of softw [17:39] I think we need to realize that, if there is no good solution for OEMs, then coincidentally, there is actually no good solution for a large chunk of Ubuntu users. [17:39] will suffer === wedgwood_away is now known as wedgwood [17:39] We have to support at least the newest release that users can buy hardware for [17:40] right now, that is the 6 month releases [17:41] rickspencer3, ping? Will users be able to upgrade from 12.10 to 14.04 or from 13.04 to 14.04? We need to make sure that we aren't leaving customers on an island [17:41] hi FunnyLookinHat [17:41] We really don't want to have to cripple our software for years at a time or alternatively introduce radical changes back into the LTS [17:41] rickspencer3, howdy ( s76 guy here ) [17:41] if we chose to go to a rolling release today (*if*), we'd have to extend the support period for 12.10 [17:42] What does that mean specifically? With the network upgrade option to 14.04? As I understand it, 18 months of 12.10 support leads right to 14.04 [17:42] This question you ask was the original point that I brought up. If you buy a computer with 12.10 there is no easy way to upgrade to 13.04 in a way that leads to a stable system. So every six months, essentially, the user is left on an island [17:43] fwiw, Dell has switched to preloading only LTS releases [17:44] danjared, That's because they are building their own kernel and it would be literally impossible for them to support anything more [17:44] FunnyLookinHat: we're not building our own kernel [17:44] danjared, I was misinformed I guess - last I checked you had to pull in some patches for your sputnick project [17:44] My mistake :) [17:45] FunnyLookinHat: are you talking about the PPA for Sputnik? [17:45] danjared, I'm talking about the linux-image that is being generated for the XPS 13 Laptops [17:46] FunnyLookinHat: We have a PPA for Sputnik, which includes a couple bug fixes and a driver Cypress generously wrote for us, but all those changes are being upstreamed. The PPA kernel is a temporary measure. [17:46] FunnyLookinHat: This is not done for the many, many other systems we ship with Ubuntu (though not currently in the US) [17:47] danjared, Ah ok - that is different than I had read. Thanks for clarifying :) [17:47] Do you think shipping 12.04 for the next 14 months is a good product decision for Dell ? [17:48] Shipping every release, LTS or not, is possibly feasible for a small company like System76. It's a miserable migraine for a company like Dell. [17:48] Understood. Good point. [17:48] FunnyLookinHat: not speaking for the company as a whole, we're happy about the HWE stack for 12.04 [17:48] (by we I mean the engineers) [17:49] It's easier to revalidate a small part of the release than an entirely new release [17:50] My feeling is that 12.04 is going to be a horrible solution for your customers within two months... and even moreso in 10 ( right before 14.04 ) - you're essentially stuck shipping two year old software. [17:50] FunnyLookinHat, danjared: fyi, at this point the kernel enhancements for Sputnik have all been upstreamed already (the Cypress driver and the backlight control fixes): Those fixes all appear in mainline Linux 3.9 kernel [17:51] To put it in perspective, that would be like shipping 11.04 today - which, if you even run it for a day, will feel like 1997 [17:51] kamal: right, I forgot to mention that they're all upstream now though not in 3.5 AFAIK. thank you so much for all your work! [17:53] FunnyLookinHat: I would surmise that our customers are more concerned with receiving a system that will still work after they first update their system than having something closer to bleeding edge. Keep in mind that, for us, our customers include companies. [17:53] danjared: my pleasure -- happy to have the slick XPS machines working so well in Ubuntu (and other Linux) [17:55] kamal: my pleasure too. my personal system is an XPS 13 :) [17:55] kamal: me too :-) === kentb is now known as kentb-afk === udsbotu changed the topic of #ubuntu-uds-community-1 to: Track: Community | Ubuntu Advocacy Kit - next steps | Url: http://summit.ubuntu.com/uds-1303/meeting/21642/community-1303-uak-next-steps/ === falk is now known as Guest31800 [18:07] Alright, feel free to invite me to the hangout whenever works. [18:07] lets get this adk going === Guest31800 is now known as falkowich [18:10] hello my friends [18:10] rrnwexec, I marked myself as essential to the session, Can you invite me? [18:10] hey dholbach [18:10] philipballew, let me see if I can get that done [18:10] dholbach, alright [18:10] I think we still have 5 [18:10] ah, no hurry [18:11] hello all :) [18:11] philipballew, if you reload the session page, you should see a "join hangout" link or some such [18:11] I see it now [18:15] Whoa! I didn't even know that an Ubuntu Evangelists team existed, that's mainly what I do. :) [18:16] http://summit.ubuntu.com/uds-1303/meeting/21642/community-1303-uak-next-steps/ [18:18] That's because the most visible example was the UDS which is big, awesome and stuff. [18:19] press releases :) [18:22] It sounds like the messaging around a 1-month cadence is very important to get right. [18:22] wendar: good idea :-) [18:23] (wrong channel) [18:24] I can't attend the hangout from my work. I'll read you from here [18:27] surely sorting out where its resting plce is the last bit to think about [18:27] putting the content in to the kit is the main thing [18:27] yeah, good point czajkowski [18:27] eh no whack [18:28] :D [18:28] :) [18:28] philipballew: I've been known to make the odd one from time to time! [18:28] * philipballew laughs and smiles [18:31] is there a specific editor to write this restucted text on ? [18:31] czajkowski, every normal text editor you want. [18:32] nods [18:32] czajkowski: I've seen something that displays a preview so that you know how it will look rendered [18:32] There's quite a big cross-over between RsT and Markdown. We've switched to RsT in CouchDB as well. [18:33] jonobacon: well the big thing is a person may think they write good docmentation and the person reviewing may not think it's great, and then it can also be a case of it reviewing of text can become rather long winded [18:33] Cracknel: thats what I would have thought of [18:33] czajkowski: found it: http://www.webupd8.org/2012/03/retext-30-released-text-editor-for.html [18:34] QUESTION: Are you planning on creating a unified website which could be used by LoCo Teams to run on their own server and customize it to their needs? It would be nice to have a unified Loco Website expirience which automatically shows news on all Loco sites etc. [18:34] recommendation: not long winded sentences :) concise and to the point sentences are far better. [18:35] aye acctive writing is good point [18:35] czajkowski, if you are a Sublime Text 2 (or 3) user there's also an RST preview plug-in as well as RST snippets. [18:35] jonobacon: would putting this on the ltp as the resouce there, again adding another incentive to go there, currently we haev the wiki, AU, and then spreadubuntu, and it'd be nice to have it in one location. [18:37] czajkowski: would you like to join the hangout? [18:37] sure [18:37] cool [18:37] dholbach: can you invite czajkowski ? [18:37] sure [18:40] Doesn't sound so happy at the moment. :/ [18:41] I wish there was a "Reference Sheet" ... one double-sided page (Letter or A4) designed for a three-ring binder. [18:42] It's very difficult for someone new to Ubuntu to find where the "useful" information is ... it's all scattered. [18:43] I'd been using Ubuntu for years before I found out about the Ubuntu Manual Project [18:43] I didn't know about Ubuntu Friendly [18:43] yeah, a link to "blessed resources" might be good :) [18:46] the actual licence : « Version: 0.1. © Copyright 2012, Ubuntu LoCo Community » [18:47] not open === X3MDroid is now known as X3MBoy [18:47] http://bazaar.launchpad.net/~uak-admins/uak/trunk/view/head:/COPYING [18:47] hum... you can read the license of everything on /usr/share/doc/uak-en/copyright [18:47] (on the bottom of the generated html) [18:47] YoBoY, how is that a license? [18:47] YoBoY, that's maybe a copyright holder, but not a license [18:48] got a link for the google hangout? [18:48] the majority of the stuff in there is CC-BY-SA-3.0 - some logos and css bits are under other licenses as they were borrowed from somewhere else [18:48] the Uak is already Creative Commons - Attribution Share Alike [18:48] https://launchpad.net/uak [18:49] dholbach: Will you be at the Ubuntu LoCo-Metting in Berlin on 13th of March? [18:49] dholbach, so it's just a bug on the generated html ^^ [18:49] YoBoY, yep, that might be - will file a bug [18:49] dholbach, I think the disclaimer and the license got mixed together there. [18:49] If you scroll way down there is a CC notice. [18:49] Sid_Payton, is that the wednesday meeting in c-base? [18:49] dholbach: Yes [18:50] Sid_Payton, I might - not 100% sure yet - I might be travelling - but it'd be good to hang out with you in c-base again :) [18:51] * mainerror nods [18:52] dholbach: would definitly be nice. would like to talk about the advocacy kit a bit more in detail and my involment in Ubuntu to hekp [18:53] awesome :) [18:53] is it possible for me to join the google hangout ? [18:53] Sid_Payton, if you drop me an email I'll let you know about my plans next wednesday [18:53] dholbach: which adress? you got so many [18:53] Sid_Payton, dholbach at ubuntu.c [18:54] dholbach: will do [18:54] lot of things in the actual documentation in the uak are (or should be) in the loco team portal directly [18:55] YoBoY: yup [18:55] do we plan to add them soon ? [18:55] YoBoY, yes, I noted down an action item [18:55] ok, :p [18:55] I'll reach out to the folks, they can probably reuse some bits from what we do with the packaging guide [18:55] http://developer.ubuntu.com/packaging/ [18:55] sorry, I need more screens and eyes [19:00] thanks everyone [19:00] Should it link in ubuntu-tour and ubuntu-advert even though those teams are kinda quiet === udsbotu changed the topic of #ubuntu-uds-community-1 to: Track: Community | LoCo Community - what's next? | Url: http://summit.ubuntu.com/uds-1303/meeting/21640/community-1303-loco-community/ [19:00] bye === kentb-afk is now known as kentb [19:05] rrnwexec, https://plus.google.com/hangouts/_/9ed4c7672878100c61fc46b147522b5e19b4d0a8?authuser=0&hl=en [19:05] jono_, If there is room, id like to join this hangout [19:06] did not see the link to join pop up in summit and i marker essential [19:06] And I've merely got static saying we're not started yet [19:07] skellat, were gonna start in just a few [19:07] can you folks see us? [19:07] philipballew: you are invited [19:07] Yes [19:07] yes [19:07] You're up Jono [19:07] Yes [19:11] are we referencing Loco member community growth specifically, or growth in the community as a whole? [19:12] BobtheBober: primarily LoCo team member growth [19:12] Interstate LoCo cooperation is not necessarily an issue. I have a hard enough time working to string people together across The State of Ohio since we are so spread apart. [19:13] * YoBoY waves :D [19:14] (and we also have "failures" in france for our events :p no global jam for example) [19:14] Our last virtual statewide event for Ubuntu Ohio: https://wiki.ubuntu.com/OhioTeam/IRC20130128 [19:15] isn't every ubuntu user a type of "loco" member, as they are using, and ultimately a part of the process to help improve Ubuntu as a whole? [19:15] its more so a question of how do we engage the users to help.. [19:15] what about loco type teams that are not purely geographically based? user based communities around interest rather than just geography? [19:18] organization [19:19] it's hard to be enthousiast when we have so many good and bad surprises provided by canonical and have to explain and defend them with people [19:19] Probably quite a bit is change management. Within the past seven days there has been an awfully large amount of changes to digest in one fell swoop. A glide path to transition into these changes would be nice. [19:20] Convergence is great but we needed to ease into it a bit more. [19:23] A community is made up of many people and turns like an overloaded warship which is to say it does not make turns on a moment's notice. Ohio got caught by this with people unable to make schedule changes to be able to participate. [19:24] an example, the shopping lens was a great privacy problem, we work in france with lot of groups defending the internet neutrality, the privacy, etc… and it was hard to explain this new direction [19:25] Canonical cant express every change in advance due to leaked information. [19:25] (and we still can't explain where we are going with that) [19:25] it wouldn't be news [19:26] I perceive a growing lack of transparency, and that's a fundamental paradigm shift. [19:26] EXACTLY - re posting to locos [19:26] maybe there is a bigger question there. If canonical cant pre-inform supporters and Loco members, how can loco members and supporters support those changes that they dont know about. [19:27] BobtheBober: We can't hit everybody with Non-Disclosure Agreements [19:27] if we could give the locos "sales collateral" it would be great [19:27] transparency it's not a problem for me, lake of information when the news is released is the problem. We only have blogs, twits,… no roadmap on where canonical is going with the feature announced [19:28] understood, but loco members can support or help build products they have no option to assist with. [19:28] where are those links ? [19:29] @yoboy in fact after chatting with David Callé I understood that there were more privacy issues with third apps before the shopping lens [19:30] winael, yes, but not on the home dash [19:31] Horse burgers! [19:31] fragmentation [19:31] like android ;) [19:32] for ex lens utility sent privacy info to weather website or other third services without any proxy server [19:33] and was access directly in the home lens [19:33] we should evaluate our incentives for users to join the loco community and contribute. Maybe why they are interested [19:33] Paradigm shifting without a clutch [19:33] I blogged about that on planet [19:33] What if, two years from now, phones and tablets have ended up as relevant as netbooks are today? Desktops are still relevant, particularly for content creators. [19:35] fair summary [19:35] Question: How do you know that a LoCo has fallen apart? [19:36] incentive [19:39] AGAIN...change management was needed [19:39] +1 the uds change is difficult to explain, my point of view is we are loosing lot of fun [19:40] YoBoY: but the upside is more community members can attend. I've never been to a physical UDS, and wouldn't be there this time if it was still physical. [19:40] laura +1 [19:40] what is the reasons, the incentives for someone looking to join the loco team. if we understand this, maybe we can grow those reasons, or re-evaluate what isn't working [19:41] commandoline_, not really… it was already online for people who can't attend physically, we just added video hangouts, and a better summit website for that [19:41] czajkowski: +1 once a month [19:42] Maybe we could revitalize the Team Reports at a glance facility on the Wiki [19:43] a meeting, that the loco leaders must attend? [19:43] BobtheBober: never going to happen due to timezones [19:43] and work commitments [19:43] Reports + meeting is a good idea [19:45] question? the point of moving UDS online was that people wanting to attend couldnt due to the physical world. It was a benefit to move it online so that everyone can attend. Why dont we take this same approach with loco teams? [19:45] czajkowski: it may be necessary to have two different times [19:46] YoBoY: and that's a real improvement. I've followed an audio stream at an earlier UDS, but it's completely different. [19:46] I think the biggest reason to join a loco team is just to know you're not alone in your geographic regioin [19:46] jono_: you're just not a joiner. ;) [19:46] good point snap-l [19:46] ha [19:47] commandoline_, seeing the people is great yes, but I assure you, it's not the same with a real face to face ;) last uds we had hangouts too, it was great too, not like the ones now [19:47] but this discussion on the uds online, is for another day ;) [19:48] YoBoY: agreed :) [19:48] BobtheBober: Doing LoCo events virtually is interesting. We've been experimenting with it in Ubuntu Ohio and had some push back against using Google Hangouts. We haven't had an alternative to it located yet but use IRC & the mailing list heavily. [19:48] CYLONS! [19:48] *cough* team reports. [19:50] ++ [19:50] jono_: ++ [19:51] How does this affect reapprovals, then? [19:52] Well, there goes a few months work for Ubuntu Ohio relative to reapproval work then [19:52] you couldnt strip all approvals, you would have to grant some form of unity [19:52] jono_: amen [19:53] what if we ditch the approvals but keep the notion of "unnapproved" [19:53] what's the point to be "observed" by the loco council if there is no more a small distinction ? [19:53] I cant wait to join that unnapproved loco team. life on the WiLd side !! [19:53] ^lol! [19:53] unnaproved is worse than approved ^^ [19:53] I've had several thoughts of "what would we lose if we bacame unapproved" [19:54] That way there aren't approved teams, but dysfunctional teams can be disapproved [19:54] nothing [19:54] and honestly, outside of prestige, not a lot [19:54] just live cds… [19:54] * commandoline_ doesn't see the difference between the loco team I'm in before and after the approval, either. [19:54] What if the approved, unapproved was completely removed and the Loco council was in place to recognize LoCo groups and the more work each group is doing, the more they get recognized [19:54] recognition is awesome. We should do more of that [19:55] but the approved / unapproved process is overhead [19:55] If we're doing this to advance the Ubuntu project, why does approval matter? Labels do more harm than good [19:55] redefine loco teams and members, rebrand. Loco isnt what it used to be. [19:55] Lcoo is still an important word, though [19:55] system76chick, it's why we are looking for another term [19:56] but approved / unapproved lost meaning [19:57] Nomination/ Recognition remove the negativity all together. The last thing a Loco group wants to hear is that the work they did wasn't good enough [19:57] system76chick: ++ [19:57] i dont personally get the importance of distinguishing loco teams by location [19:57] +1 it would be great to see teams form around communities of interest and not just geographies [19:57] here here weddle [19:58] e.g.... people using ubuntu in corporate contexts [19:58] How about a point system for Loco Groups and the loco council was in charge of assigning the points? "U Points" [19:58] people using ubuntu to do XYZ [19:58] loco > projects, ideas, work [19:58] in my country, if someone want's to create a local group, we help to create it, because it's important for people to fill they have a real group, where they can meet others. Each country is different for that [19:58] @Bob totally agree that location is less relevant... we are on a virtual UDS here [19:58] bullet points on pad [19:59] If I never have to fill out a team report ever again, I'll be a happy man. :) [19:59] could totally have virtual locos [19:59] did you guys get to discuss the LoCo name thing? [20:00] <-- this guy agrees with Jono [20:00] Takes away from the focus of what Loco teams are here for === udsbotu changed the topic of #ubuntu-uds-community-1 to: Currently no events are active in this room - http://summit.ubuntu.com/uds-1303/community-1/ - http://ubottu.com/uds-logs/%23ubuntu-uds-community-1.log [20:01] touch on weddles point [20:01] local is not what local is anymore [20:01] good thoughts system76chick [20:01] here we are on a virtual UDS, talking about Loco [20:01] Thank you [20:01] lol [20:01] Thanks everyone [20:01] good sessio guys! [20:01] *session [20:01] thanks everyone [20:02] * YoBoY is starving ^^" [20:02] thanks all [20:02] utc time [20:02] :) [20:02] thanks [20:02] thanks for the good input! [20:02] thanks [20:02] are the google hangouts recoreded [20:02] Thank you, everyone! [20:02] and watchable later? [20:02] BobtheBober: they are [20:02] yes [20:02] any links? [20:02] BobtheBober, yes [20:02] just refresh the page in a couple of minutes [20:02] alright sweet [20:03] no other events for today right? [20:03] nope [20:03] Short and sweet [20:03] Time for lunch! Have a great day everyone! [20:03] I'm going to make some crepes for my dinner :) [20:04] except for 12.10 [20:05] YoBoY: \o/ [20:05] quesh, avoue que t'en veux :D [20:05] YoBoY: oui [20:59] hi [21:01] hi melodie [21:47] he YoBoY ! [21:48] this is a web chan so we are not notified, right ? :) [21:48] don't really know how the web irc client works [21:48] I use xchat === kentb is now known as kentb-out === wedgwood is now known as wedgwood_away