13:05:22 <jreznik> #startmeeting FESCo Town Hall - Summer 2014
13:05:22 <zodbot> Meeting started Fri Jul 11 13:05:22 2014 UTC.  The chair is jreznik. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:05:22 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
13:05:43 * jreznik will look for folks joining only #fedora-townhall
13:06:09 <jreznik> #meetingname fesco-townhall-summer-2014
13:06:09 <zodbot> The meeting name has been set to 'fesco-townhall-summer-2014'
13:07:18 <jreznik> sorry for inconvinience using this channel, we also have to use a bit different protocol
13:09:27 <jreznik> so let's to it the way, slot for questions and slots for answers, please, don't post more questions when the answers slot is used
13:10:18 <jreznik> #topic questions slot #1
13:10:36 <jreznik> feel free to ask your questions now
13:11:11 <jreznik> #chair amigadave jwb kalev thozza lnykryn
13:11:11 <zodbot> Current chairs: amigadave jreznik jwb kalev lnykryn thozza
13:11:41 <kalev> q: can everyone introduce themselves in a few words?
13:12:34 <jreznik> #info answers from candidates questionnaire are available on the wiki
13:12:43 <jreznik> #link https://fedoraproject.org/wiki/Elections/Questionnaire
13:12:53 <jwb> i'm Josh Boyer.  I work on the fedora kernel and i've been around fedora for a really long time.
13:13:01 <jreznik> wait a moment pls
13:13:19 <jreznik> #topic answers slot - can everyone introduce themselves in a few words?
13:13:28 <lnykryn> Hi, I am Lukáš Nykrýn. I work in Red Hat in Plumbers group, primarily on packages related to init.
13:13:33 <jwb> i'm Josh Boyer.  I work on the fedora kernel and i've been around fedora for a really long time.
13:13:38 <amigadave> i'm David King, working for Red Hat in the desktop team, contributing to Fedora as a GNOME and MinGW packager (and GNOME upstream)
13:14:07 <kalev> Hi, I am Kalev Lember. I work on Fedora Workstation and related packages, and contribute to GNOME upstream.
13:14:21 <thozza> I'm Tomas Hozza, I work for Red Hat in developers experience - packaging team. Looking after most DNS servers. Recently working on DNSSEC, mostly from the client side
13:16:12 <jreznik> thanks guys - for the next question I'm going to rotate your voice and ask you to speak
13:17:05 <jreznik> #topic questions slot #2
13:17:39 <jreznik> so it's time for another question
13:18:10 <kalev> can I ask another one?
13:18:13 <langdon> jreznik: are you waiting for someone to suggest one? or do you have one?
13:18:26 <kalev> q: How do you feel about the long F21 cycle we just had and how often do you think Fedora should release?
13:18:28 <jreznik> langdon: it's townhall for you - you can ask
13:18:51 <jreznik> so it's suggestion based
13:19:01 <jreznik> the questionaire is already answered on wiki
13:19:02 * langdon nods
13:20:17 <jreznik> #topic answers slot - How do you feel about the long F21 cycle we just had and how often do you think Fedora should release?
13:20:29 <jreznik> jwb: you start now, then lnykryn, amigadave, kalev, thozza
13:21:23 <jwb> i think we're still in the middle of the f21 cycle, so the question is phrased slighly odd.  i'd say f21 is a one-time delay to get fedora.next bootstrapped.  if it's successful, we should most likely get back to a shorter release cycle
13:22:21 <lnykryn> Well it depends, it gives developers time to work on more complicated features, but I think that most work is always done few weeks before deadline.
13:22:23 <jreznik> btw. please use eol, so I can move to other candidate and don't stop you talking
13:23:17 <lnykryn> sorry and eol from my side
13:23:46 <amigadave> i hope we can return to 6-month cycles that sync up with the release cycles of other upstream projects, such as GNOME (as we missed 3.12 with F21)
13:23:50 <amigadave> i don't necessarily think that a long cycle is bad, but changing the cycle by such a large amount breaks user expectations
13:23:53 <amigadave> eol
13:24:08 <kalev> I personally prefer predictable schedules. Our users have come to expect and value regular releases. I think we should try and stick with strictly 6-month release cycle.
13:24:16 <kalev> There's always a next release in 6 months if some feature doesn't make it in time.
13:25:06 <kalev> Or if a longer (12 month) schedule makes sense for one of the products, they don't have to follow the 6 month schedule that the whole of Fedora has.
13:25:27 <kalev> But for Workstation I'd say we should stick with a strict time based schedule.
13:25:30 <kalev> eof
13:25:33 <thozza> I think that with upcoming Fedora products, the release cycle should reflect the targeted audience. So for server it should be definitely longer than half a year, for the rest half a year seems fine. I think that having such "short" release cycle enables us to deliver the latest and greatest software.
13:25:34 <thozza> eol
13:26:18 <jreznik> ok, thank you! we now have #fedora-townhall setup - do you prefer to stay here or move there?
13:26:36 <jwb> stay
13:26:42 <kalev> let's stay here, we don't have so many people that we can't handle it here
13:26:48 <amigadave> yep, we are here now, so let's stay
13:26:52 * thozza agree
13:26:55 <lnykryn> +1
13:27:13 <jreznik> btw. thanks FranciscoD for help!
13:27:24 <jreznik> #topic questions slot #3
13:27:27 <cbx33> q: Do you think that Fedora is geared towards your "average" computer user, as I'm sure Ubuntu would say they are, or are we more geared towards developers? And with that in mind, should there be a shift in how we market Fedora to new users?
13:27:30 <FranciscoD> No worries. Sorry I didn't get here earlier.
13:28:24 <jreznik> #topic answers slot - Do you think that Fedora is geared towards your "average" computer user, as I'm sure Ubuntu would say they are, or are we more geared towards developers? And with that in mind, should there be a shift in how we market Fedora to new users?
13:28:44 <jreznik> now lnykryn starts, follewed by amigadave, kalev, thozza, jwb
13:30:09 <jreznik> lnykryn: pls, your turn
13:30:17 <lnykryn> I think fedora is aimed to developers and should stick with it. We can be best only if we can attract developers and expectation how a distribution should look differs in eyes of developers and regular users.
13:30:20 <lnykryn> eof
13:30:47 <amigadave> i think that Fedora is more geared towards power users and developers, because of the drive to have new features in Fedora first, and the short lifecycle
13:30:51 <amigadave> i hope that products mean that users get a release that fits them better, which implies that we need to market products to those users appropriately
13:30:56 <amigadave> eol
13:31:06 <kalev> We need a critical mass of regular users who would spread the word of how good Fedora is; that is what's also going to attract developers.
13:31:11 <kalev> And once we have that, developers are likely to stick around too.
13:31:21 <kalev> Of course, we as the developers of Fedora only have limited manpower, so we have to make some choices where we concentrate our efforts.
13:31:28 <kalev> With Workstation, we are currently trying to make the experience better for developers, but that doesn't mean dropping support for regular end users -- quite the contrary.
13:31:31 <kalev> eof
13:31:37 <thozza> I think it is geared more towards "little bit advanced average" user, or at least user willing to solve their issue when they arise. With e.g. Ubuntu, the average user might like the non-free software and codecs distributed there, that are not available in Fedora itself. But this makes Fedora really great and little bit harder to use for average user. But devels should be prepared for solving issue, o Fedora is definitely sui
13:31:39 <thozza> eol
13:32:05 <kalev> thozza: your sentence was cut off at "Fedora is definitely sui"
13:32:32 <thozza> Fedora is definitely suitable for developer, from my point of view.
13:32:37 <jwb> i think fedora wasn't geared towards any particular audience previously, and we suffered from it.  fedora.next is an effort to change the focus on who we're trying to reach.  that would include marketing.  eof
13:33:12 <jreznik> #topic questions slot #4
13:34:06 <sgallagh> q: What level of importance do you feel we should put on satisfying "traditional" users of Fedora vs. putting effort on the Products and enabling Spins?
13:34:41 <jreznik> #topic answers slot - What level of importance do you feel we should put on satisfying "traditional" users of Fedora vs. putting effort on the Products and enabling Spins?
13:35:02 <jreznik> amigadave, it's up to you, then kalev, thozza, jwb, lnykryn
13:35:42 <amigadave> hopefully, a Workstation product corresponds relatively closely to what a traditional Fedora user would expect
13:35:48 <amigadave> i see spins as supplementary to that, but not the primary focus of the Working Groups (although something we want to keep)
13:35:51 <amigadave> eol
13:37:15 <kalev> I don't understand the question, how can you say "traditional" users of Fedora vs the Products?
13:37:33 <kalev> I would say the 3 new products, especially Workstation are going to cover the traditional user space.
13:38:14 <kalev> So there's no "versus" here but instead we _are_ satisfying traditional users through the new products.
13:38:15 <sgallagh> The traditional approach has been more "box of lego bricks" where the Products and Spins are more like pre-built model kits.
13:38:58 <kalev> In that case, Fedora has a large community surrounding it
13:39:14 <sgallagh> (Sorry, it's a difficult question to phrase well)
13:39:28 <kalev> I'll suspect that people who value a more focused, "productized" if you will, effort are going to join one of the products teams.
13:39:50 <kalev> Other people are are free to help make the "box of lego bricks" better.
13:40:19 <kalev> But I do like the idea that Fedora is now trying to be more focused, since it gives a sense of direction and purpose to both developers and the wider community.
13:40:22 <kalev> eof
13:40:46 <thozza> I think that the effort should equal the approx. size of the users base (e.g. for spins). As for the products, I think they deserve equal priority (at least in the beginnings). If I understand it correctly with the "traditional Fedora" being basically any of the products, then it should and will get the amount of love Products are getting...
13:40:48 <thozza> eol
13:40:51 <lnykryn> I believe we should put an effort to products. Than we can bring better desktop for desktop users and better server for administrators.
13:40:53 <lnykryn> eol
13:41:01 <jwb> the question is still somewhat ambiguous.  if you mean install media, it needs interested people keeping the "traditional" iso alive.  if you mean the posibilities of mix and match package installation after the fact, it should still work fine with whichever Product a user chooses.  Spins will also still exist.  eof.
13:42:05 <jreznik> #topic questions slot #5
13:42:06 <cbx33> q: Are the any areas of Fedora that you personally feel are currently really lacking, this could be marketing, support for hardware, support, community, qa...etc, and how do you see us tackling that issue in the future?
13:43:52 <jreznik> #topic answers slot - Are the any areas of Fedora that you personally feel are currently really lacking, this could be marketing, support for hardware, support, community, qa...etc, and how do you see us tackling that issue in the future?
13:44:01 <jreznik> kalev, thozza, jwb, lnykryn, amigadave
13:44:25 <kalev> I would really like to have support from hardware vendors, so that we could say e.g. that this particular laptop model is certified to run Fedora.
13:44:48 <kalev> Or the other way around, that hardware vendors would say "Hey Mr. Customer, we can ship this hardware to you without software and give you a Fedora CD. This hardware is Fedora-compatible."
13:45:10 <kalev> And then when we rebase kernels and mesa and other hardware support, we'd make sure the supported hardware list actually works.
13:45:20 <kalev> Of course, I have no idea how to do it in practice, it's just a dream :-)
13:45:23 <kalev> eof
13:45:46 <thozza> I think that some processes done by maintainer could be automated.
13:46:15 <thozza> We can solve this by developing some more tooling.
13:47:02 <thozza> Also better quality assurance and come kind of continuous integration will be beneficial. I really don't like when something gets into stable and break things
13:47:15 <thozza> eol
13:47:17 <jwb> i think cross team collaboration is lacking.  fesco doesn't talk to marketing.  the board doesn't talk to fesco.  nobody talks to ambassadors.  i think this needs to be improved significantly so everyone is on the same page.  however, i think that's mostly a task for the Board to tackle.  eof.
13:47:47 <lnykryn> Yep I also think that we lack cross team communication.
13:48:02 <lnykryn> that should be a priority in the future. eol
13:48:11 <amigadave> currently (F20, pre Fedora.next) i think that the lack of focused products hurts Fedora, as it's difficult for a user to know what they want when confronted with all the spins on the download page, or a huge list of packages in the software installer
13:48:15 <amigadave> there's been some good work done in Workstation on the software installer, and i think that differentiable products will help users choose, and give developers better focus
13:48:19 <amigadave> the package review queue is perpetually really long, so it would be good to clear through a lot of that, but i don't think it will happen without more dedicated developers
13:48:23 <amigadave> again, hopefully enthusiasm from the product split will help work through areas where manpower is lacking
13:48:26 <amigadave> eol
13:49:06 <jreznik> #topic questions slot #6
13:49:08 <rishi> Related to what thozza just said, what is your opinion on the frequency with which updates are pushed to stable releases? In my experience this leads to poorly tested updates due to the entire package set changing constantly. If you agree, then how do you think we should fix this?
13:49:47 <jreznik> btw. as this single channel surpress discussion, there's fedora-townhall-discuss now
13:49:54 <jreznik> #fedora-townhall-discuss
13:50:07 <jreznik> rishi: is it question or comment?
13:50:35 <rishi> Question.
13:50:55 <jreznik> thanks rishi
13:51:05 <jreznik> #topic answers slot - Related to what thozza just said, what is your opinion on the frequency with which updates are pushed to stable releases? In my experience this leads to poorly tested updates due to the entire package set changing constantly. If you agree, then how do you think we should fix this?
13:51:25 <jreznik> and thozza starts - good timing! jwb, lnykryn, amigadave, kalev
13:52:24 <thozza> I think that the people dedicated to testing have limited resources. We could have some automation testing that would do some interoperability and functional testing
13:52:45 <thozza> I know this can be really hard for some components, maybe even impossible
13:53:19 <thozza> Also increasing the karma threshold for some core components would be useful...
13:53:21 <thozza> eol
13:53:32 <jwb> this has been a problem for a long time.  there's no great solution here because it mostly involves all the maintainers agreeing on an approach, and nobody does.  automated testing can help, but that doesn't change the frequency problem.  however, we could explore additional update channels that are tiered in terms of frequency/stability, and also change the way we present updates to users.  eof
13:53:35 <lnykryn> I think the problem is in small QE community around fedora, not with a frequency of updates.
13:53:36 <lnykryn> eof
13:53:54 <amigadave> there is a stable updates policy, which some packagers comply with to different degrees than others, but i think it's difficult to automate the packager out of the picture
13:53:58 <amigadave> i think that better automation will help there, but that needs some continuous integration platform (such as GNOME Continuous, an upstream project based on OSTree) and good tests
13:54:03 <amigadave> in other words, lots of work on several layers of the stack, and something that all Working Groups should be interested in :-)
13:54:06 <amigadave> eol
13:55:08 <jreznik> kalev: your turn
13:56:40 <kalev> sorry
13:56:53 <kalev> Yes, I think it's a problem. Fedora pushes out a large number of updates each day and users are between a rock and a hard place:
13:57:02 <kalev> apply updates every day and face the churn of things changing every day and sometimes regressing?
13:57:11 <kalev> or, alternatively, not apply any updates and risk having known vulnerable software?
13:57:28 <kalev> I don't see this improving unless there's a dedicated person working on gating the updates that are allowed to stable releases.
13:57:48 <kalev> And it would not be a fun job for a volunteer; I would say it has to be a full time paid position from Red Hat if we are going to want to gate updates like this.
13:58:20 <kalev> In my ideal world, a user would install a Fedora release and it would stay stable for it's supported lifetime
13:58:58 <kalev> and when they want to switch to a newer desktop environment, they'd update to a newer Fedora release
13:59:23 <kalev> of course, we currently suck majorly with our distro upgrade tools, especially because we lack a graphical distro updater
13:59:27 <kalev> so this are currently not simple
13:59:56 <kalev> Also, while users generally don't want to have changing desktop environments, they do want updates to end user applications -- e.g. Firefox.
14:00:01 <kalev> So it's not an easy balance to make.
14:00:23 <kalev> But in conclusiony, I'd say we have a problem with the sheer volume of updates and we should try to tackle it somehow.
14:00:26 <kalev> eof
14:01:01 <jreznik> #topic questions slot #7
14:01:19 <jreznik> one hour passed, let's do one, two more questions as we started a bit late
14:01:40 <rishi> Thanks, kalev
14:01:47 <langdon> question: do you see the fedora.next products adding new code? or just distributing upstream projects in an integrated way?
14:02:08 <sgallagh> q: Is there a place for "devices" in Fedora's future? (Such as Ubuntu's varied attempts at tablets and TVs)
14:02:28 <jreznik> #topic answers slot - do you see the fedora.next products adding new code? or just distributing upstream projects in an integrated way?
14:02:43 <jreznik> now jwb starts, then lnykryn, amigadave, kalev, thozza
14:03:47 <jwb> i think the first iteration is going to be mostly delivering upstream code.  as we continue on, i would like to see a balance of new code being written in fedora to meet various needs, and upstream communities around that code being started. eof.
14:04:06 <lnykryn> I think that products are about packaging, not about upstream. But I hope that for example in "cloud and stack area" it will lead to changes in upstream projects.
14:04:43 <lnykryn> because for example some projects are not behaving well in containers. eof
14:05:02 <amigadave> for Workstation, a lot of our upstream contributions are focused on the task list that has been created by the Working Group, so it's both new code that we are writing as well as packaging up an upstream projext
14:05:07 <amigadave> i see fedora.next is general as the integration of upstream projects, and adding code where necessary to achieve the desired level of integration
14:05:15 <amigadave> sometimes that might be a whole new project, but maybe it is just adding a customization to an existing project
14:05:19 <amigadave> i hope that the contributions from Fedora to upstream projects can help drive them forward too
14:05:22 <amigadave> eol
14:06:00 <kalev> I would like Fedora.next products to drive development and work within upstream communities to achieve our goals.
14:06:42 <kalev> Integration is definitely important, but we should also work on features to get where it makes sense for Fedora.
14:06:52 <kalev> This is something I think Ubuntu has failed -- they do have nice focused products, but they are disconnected from upstream communities.
14:07:16 <kalev> Instead, I would want Fedora to stay close to upstream and contribute back whatever new development we do.
14:07:28 <kalev> eof
14:07:36 <thozza> I don't think it is "just distributing" upstream projects. I think products are integrating projects and focusing on specific sub-set of components that are needed for the specific use-case. I believe that the intention is for the products to be suitable for a specific use-case out-of-the-box The integration work comes with some new code, but I think it is more of a "side effect" of the integration process.
14:07:41 <thozza> eol
14:07:51 <langdon> can i ask a follow up?
14:09:15 <jreznik> langdon: sure
14:09:26 <langdon> feel free to skip this to get to sgallagh's question.. You all mentioned (i think) contributing back upstream, but what if the f.next product changes are not inline with the upstream project's goals?
14:09:36 * langdon sometimes is a very slow typer
14:09:51 <jwb> jreznik, same answering order?
14:09:56 <jreznik> as it's follow up, same order
14:10:14 <jreznik> go ahead jwb
14:11:04 <jwb> in the case where the changes aren't inline with upstream, we either need to find a different suitable upstream or carry changes in the fedora package.  we shouldn't be afraid of deviation where it is necessary and beneficial.  eof.
14:11:12 <lnykryn> Than we should find a different solution, I don't think we should go against upstream.
14:11:16 <lnykryn> eof
14:11:26 <amigadave> we always try to contribute upstream first, but if upstream doesn't accept the patches then we are forced to carry the patch downstream
14:11:33 <amigadave> this happened with GNOME Terminal transparency support, and can work well for specific features or integration points, but hopefully it is the exception rather than the rule
14:11:37 <amigadave> eol
14:11:46 <kalev> Sometimes things just don't work out. I guess this should be decided on case-by-case basis, but we'll always have the opportunity to do local patches and fork if it's really needed. It's free software, after all.
14:11:52 <kalev> eof
14:12:02 <thozza> I'm for finding a way to solve problems with upstream.
14:12:26 <thozza> downstream changes always comes with painful amount of work
14:13:01 <thozza> It may be doable for one component if really necessary, but for sure not more
14:13:14 <thozza> So upstream first ;)
14:13:14 <thozza> eol
14:14:11 <jreznik> #topic answers slot - Is there a place for "devices" in Fedora's future? (Such as Ubuntu's varied attempts at tablets and TVs)
14:14:31 <jreznik> last one, nice one!
14:14:36 <jreznik> lnykryn, amigadave, kalev, thozza, jwb
14:14:46 <lnykryn> Not now. I think that ubuntu did that just for marketing reasons. First we need a stable compact base system and there are still a lot of bits missing. eof
14:15:11 <amigadave> i think that work on devices would need more manpower, money and a (or several) hardware partner
14:15:15 <amigadave> those things are not impossible, but several Free Software projects have had false starts and failures with similar attempts
14:15:20 <amigadave> targeting an existing popular device (such as the Raspberry Pi) can work well, but to drive development of devices needs a lot more commitment
14:15:25 <amigadave> that probably needs backing, from either Red Hat or another corporate source, unless we can raise enough money from personal donations
14:15:31 <amigadave> there can be a place, but it needs a lot of dedicated individuals for it to work out
14:15:34 <amigadave> eol
14:16:13 <jreznik> kalev: are you still here?
14:16:14 <kalev> Sure, anything's possible if there are dedicated people driving things.
14:16:17 <kalev> yep
14:16:33 <kalev> I would actually _love_ to have a third party vendor use Fedora as a base for their product and contribute fixes back to us where it makes sense.
14:17:20 <kalev> Let's leave doors open for people who want to do this, but I don't see e.g. Workstation driving TV support at this point.
14:18:12 <kalev> Like I said in the questionaire, if we're ever to get a 4th main Product, I think it makes sense for it to be for a completely different market segment, when compared to the existing products.
14:18:21 <kalev> e.g. Fedora Mobile.
14:18:22 <kalev> eof
14:18:57 <thozza> I don't think that Fedora community should focus on introducing new devices with Fedora. But personally I would find it cool if Fedora could be used on some tablet. If there is group of people willing to work on this, they should go for it and get supported if necessary.
14:19:12 <thozza> eol
14:20:16 <jwb> i don't think the Fedora project should be driving "device" efforts anytime soon.  If a vendor wants to use Fedora on a device i would be more than happy for that to happen.  eof
14:21:21 <jreznik> #topic open floor
14:21:45 <jreznik> thank you all candidates and folks here for very nice questions and answers!
14:22:06 <jreznik> and again sorry for me not being prepared in time :(
14:22:29 <kalev> Thanks for the questions everyone, I had fun here today!
14:22:50 <amigadave> thanks all, good to meet and participate in the discussion :-)
14:23:11 <thozza> thanks all, it was an interesting discussion
14:23:15 <kalev> and thanks jreznik for organizing it.
14:23:22 <cbx33> Yeh thank you for the answers
14:23:23 <jreznik> and one more reminder - more questions and answers are at https://fedoraproject.org/wiki/Elections/Questionnaire
14:24:15 * jreznik is going to end this meeting in a few seconds and will send meeting long to the right places
14:24:26 <jreznik> fuse set to 3...
14:25:26 <jreznik> 2...
14:25:54 <jreznik> 1...
14:26:21 <jreznik> #endmeeting