No edit summary |
Zoltanh721 (talk | contribs) |
||
Line 51: | Line 51: | ||
'''Marcus Moeller - FAmSCo'''<br/> | '''Marcus Moeller - FAmSCo'''<br/> | ||
It would be great to have some kind of staged updates. Other distributions (e.g Linux Mint) have systems like that. For those who like to be on top, it should still be possible to install all updates. | It would be great to have some kind of staged updates. Other distributions (e.g Linux Mint) have systems like that. For those who like to be on top, it should still be possible to install all updates. | ||
'''Zoltan Hoppar - FAmSCo'''<br/> | |||
It would be huge to utilise an 'service pack/rollback/offline update/rolling release' support system. We have all the apps to realise this and the base tech needs, but haven't introduced yet. This could ensure the stable updates, and also test runs, and feedbacks. | |||
'''Stephen (sgallagh) Gallagher - FESCo'''<br/> | '''Stephen (sgallagh) Gallagher - FESCo'''<br/> |
Revision as of 06:59, 11 November 2010
Main questions
What is your take on bundled libraries?
Sandro Mathys - BOARD
In my opinion the current policy is the correct approach and should be adhered to.
Jaroslav Reznik - BOARD
I support our effort/policies to not bundle libraries (and currently fighting with surgery of such a library ;-). The real question is - what can we do if we're not able to unbundle libraries, especially for the first class packages like Firefox and upstream is not responding to our needs. Exceptions should be granted temporary, case-by-case (no free ride for any package) and raised upstream. In case of no reply (one release time?)- we should act even it would mean breaking some features/packages. Top case is Firefox - as I understand it correctly, it's not like they refuse unbundling libraries (some are unbundled already) but some people refuse to do - upstream is probably lacking any policy. I think one, two releases timeframe is ok to make it better.
Marcela Mašláňová - FESCO
There are needed in some specific cases. These packages should be reviewed by FESCO members or FPC.
Joerg (kital) Simon - BOARD
It is not always about the easy way to go, i think Packaging:No_Bundled_Libraries is the right approach!
Stephen (sgallagh) Gallagher - FESCo
Bundled libraries should not be allowed in Fedora except in highly
exceptional circumstances. I do however think that we should be making a
greater effort to promote repos.fedorapeople.org as an alternative
central authority for storing those packages that are unacceptable for
inclusion in Fedora itself but remain important enough for end-users
that they should still be available. It doesn't make sense to require
Fedora end-users to recompile their own packages just because we don't
provide them officially. This has a tendency to drive users away from
Fedora and into distributions that they see as much less restrictive.
Providing these unofficial repositories within a domain mentioning
"Fedora" would also retain mindshare (rather than just telling people to
set up their own competing repositories on unrelated domains).
David Nalley - BOARD
This is a FESCo and FPC issue and not a Board issue.
Sandro Mathys - BOARD
While I understand the need for stability, I think Fedora is well-known and popular for being bleeding-edge. Many developers and users feel that Fedora is becoming too stable recently and not the place to develop new things anymore. The stable updates vision recommends such users to use rawhide instead but I think that's very often a tad too broken for most people so we'd need something in between, similar to the branches of future releases but implemented as a rolling release like rawhide.
Jaroslav Reznik - BOARD
I support any effort that makes our systems more "stable". But I don't think current implementation covers everything what's needed as I think it's much more deeper and wider problem - just banning updates does not work (even it can look like from some povs). It goes up to our release process (what is the use for usually not usable Rawhide and two!!! hyper conservative releases out the wild for example?), to our tools - auto qa, auto rebuilds notifications to documentation (all changes has to be properly communicated!). I know it's a big (huuuuuge) task but it's not possible to decouple it -> I don't like current implementation. We are just loosing our flexibility.
Marcela Mašláňová - FESCO
I enjoyed Fedora as developer distribution, so I'm definitely for more updates. If we decided for more stable releases, then we should solve problems of bodhi and update process. We don't have enough testers for all supported releases. Some packages are not tested at all and users are not willing to give feedback in bodhi. Maybe shorter period in testing for non critical updates would be better. We should look for more proventesters, because at the moment they are testing mainly F-14 and F-13 updates are waiting. Maybe F-13 updates should need less karma from proventester or receive karma from F-14 (even this is not the best solution). For proventester there should be some list of important updates waiting for karma (security updates first).
Joerg (kital) Simon - BOARD
It is a good approach, because no other Linux Distribution stands for new upstreams in a stable "form" like we do.
For my non-critical package-updates which i do from time to time, the karma system for updates works well and i can do my work on Fedora Security Lab activities, i also think FESCo is asked here as well to work with the board.
Marcus Moeller - FAmSCo
It would be great to have some kind of staged updates. Other distributions (e.g Linux Mint) have systems like that. For those who like to be on top, it should still be possible to install all updates.
Zoltan Hoppar - FAmSCo
It would be huge to utilise an 'service pack/rollback/offline update/rolling release' support system. We have all the apps to realise this and the base tech needs, but haven't introduced yet. This could ensure the stable updates, and also test runs, and feedbacks.
Stephen (sgallagh) Gallagher - FESCo
I'll mention the highlights here, but I've discussed this at great
length on the advisory board mailing list here:
http://lists.fedoraproject.org/pipermail/advisory-board/2010-August/009043.html
My general feeling is that Fedora needs to be divided into three phases: Stable, Next and Rawhide. In the past, when a stable Fedora is released, we've branched Fedora N+1 by snapshotting Rawhide and continuing from there. I think we need to change this. I think that Fedora N+1 should always start from Fedora N, and changes from Rawhide should be merged in only when the maintainers feel the time is right.
The goal here would be for Fedora Branched to be a semi-stable feature preview, while Rawhide should remain a volatile tech-preview.
Fedora Branched should be where (for example) the next version of KDE should be dropped. Those who can't wait for an official Fedora stable release should be encouraged to move to Branched. This is in contrast to encouraging them to move to Rawhide, which is generally too dangerous for average testing.
I've also proposed that enhancement updates made to stable releases should require proventester karma similar to critical path updates. This should be done to minimize the effect of "I installed Fedora on release day and it behaved one way, but after all of the updates it doesn't work the same way at all".
David Nalley - BOARD
I think that implementation is a FESCo issue, and not a board issue.
Is there any new initiative you would like to see happen during your term to improve packaging?
Jaroslav Reznik - BOARD
I'd like to see the same packaging process for pre-review time and post-import time (same process). Now with dist-git I think it's much more feasible. Something like OBS would be nice to have too but I know, manpower...
Marcela Mašláňová - FESCO
I support implementation of 'packaging groups' for SIGs. Instead of giving 'everyone' provenpackager, it could be created group for e.g. KDEdeveloper, Perldeveloper... People in this group could fix any package in group marked as part of particular group. Also support of pseudo-user (how anaconda team use it) could be better, but this won't be needed if groups would be supported.
Joerg (kital) Simon - BOARD
No, even me as a not so hardcore technical person can do the work i need to do. I am always open to hear improvements from the community and will address them - but i also think there are a lot more qualified people for packaging than me.
Stephen (sgallagh) Gallagher - FESCo
I think we need to find a way to improve mentoring for new packagers.
Currently, it can be difficult to find a sponsor, and some new package
requests go for months or longer without review. Something that's been
discussed lately was a sort of "classifieds" board that we could set up
to let Fedora contributors request help or offer services to projects. I
think this would be a good way to encourage mentoring in the package
process.
I also think that sponsorship after-the-fact is the wrong way to bring new packagers into the system. I think that a new contributor should write on the bounty board something like "I want to package libmcguffin, I need a mentor" and then the sponsor should help walk this contributor through his or her first new package. Right now, packagers are essentially throwing their contributions over the wall, hoping that someone will bother to review it. I think involving a sponsor right from the start (and having them guide the new contributor) would provide a more welcoming environment for packagers.
David Nalley - BOARD
I think that packaging initiatives are a FPC and FESCo issue.
Do you think it's important to have more people vote in Fedora elections? If so, how would you encourage that?
Sandro Mathys - BOARD
Yes, definitely. I would try to increase the awareness of just how important it is to vote in elections but I think in the end the real problem is that people just don't vote even if they now why they should. Probably only some kind of directly connected reward would help to chance that but unfortunately I can't come up with a good idea on how that could look like.
Jaroslav Reznik - BOARD
The people who cares should vote and votes usually. It's a general problem with all elections - not only in Fedora world. I believe in elections and voting and people should vote even that means my ideas are overvoted. How to encourage? How to bring more people involved? I like the idea of Fedora notification system (for interested people only, not for all users on by default).
Marcela Mašláňová - FESCO
Yes. I missed few votings. There should be sent email about end of nomination period with list of candidates and email about start of voting. Also IRC topics could mentioned it.
Joerg (kital) Simon - BOARD
Yes! Elections are a right to choose and nothing that should be enforced - elections are the way to change things and not questioning all and everything all the time without having a solution and stealing time from working people. This should be pointed out more! If people are elected, we gave them our trust and therefore we should trust them in their doing. I do!
Stephen (sgallagh) Gallagher - FESCo
I think it's important to have Fedora's contributors be a part of the
Fedora process. I think that voters need to have a sense that the people
they are choosing for office really represent them. Unfortunately, when
so few people vote, the selection of leaders tends to become the
election of those best supported by the vocal minority.
I'm not certain what can be realistically done to increase voter turnout. As Larry Wall famously noted, the chief Virtues of the programmer are "laziness, impatience and hubris". In order to have more people vote, there needs to be some sort of positive reinforcement involved. People need to have a real sense that their vote had an immediate impact pertaining to their interests.
Perhaps one approach would be to time the voting period around a major sticking point on Fedora policy (the bundled libraries debate would be a good example). People should be encouraged to vote for the candidate(s) who best represents their interests in that debate. In this way, people have a more obvious vested interest in the outcome of the election.
David Nalley - BOARD
I think it is important, but as in government elections people vote when they care or are passionate about an issue or a candidate. The fact that there is so much apathy does give me pause. I fear it means that there is apathy about the bodies, who gets elected to them, and what those bodies make happen.
A number of communities require participation in elections to retain membership. Perhaps that is a good way to proceed, but I hate to erect yet another barrier to participate, especially when it may little to do with what the contributor does in Fedora. I think my perspective is that it is the candidates responsibility to excite (or anger) people and give them a reason to participate in the election (either to vote for the candidate, or to vote 'against' the candidate). If I can't excite you enough to show up and vote, or anger you enough to vote for others, then perhaps I don't deserve to be elected.
Do you run Fedora on the computer you use most often? If so what Desktop? If not, why not?
Sandro Mathys - BOARD
Yes, I do - actually Fedora 14 with the lovely KDE desktop is running on all of them: my office workstation, my home workstation (there's also a win7 for fancy gaming purposes on that one, though), my laptop and my netbook.
Marcus Moeller - FAmSCo
Yes, I run the latest version of Fedora (managed with Spacewalk) on my office PC and mixed Versions of Fedora at home.
Jaroslav Reznik - BOARD
All my systems runs Fedora! No windows, no gates! Except my cell phone that runs Maemo (I should try Fedora...). As a desktop I use Plasma Desktop 4.5 by KDE.
Marcela Mašláňová - FESCO
Currently, I have Fedora rawhide on my testing desktop (without any DE), F-14 on laptop, F-13 on netbook and home desktop (Plasma, former KDE, on all of them).
Joerg (kital) Simon -BOARD
Yup - gnome and lxde depends on the tasks - and i use it as a platform for my work on Security Testing - this is the reason why i maintain the Fedora Security Lab. Ah, and i love guake.
Stephen (sgallagh) Gallagher - FESCo
I run Fedora 14 with the GNOME desktop on my laptop. I have also used
KDE quite extensively in the past, but for the last few years I've used
GNOME. The transition from KDE 3 to KDE 4 was a rocky one, and I
switched over to GNOME until KDE stabilized further. I intend to find
the time to give KDE another chance in the relatively near future.
David Nalley - BOARD
Yes. All of my personal machines (and $dayjob-assigned machines) run Fedora, and I am currently using KDE on my laptop.
Do you think Fedora should focus on the Desktop? Or someplace else? If so, where?
Sandro Mathys - BOARD
Focusing sounds much like we should stop doing what we're good in - to allow everyone to work on everything. I sure think Fedora should focus more on the desktop than on servers, but as long as there's someone thinking Fedora should become the #1 server distribution why stop them working on that goal? Better have some features that only a few people ask for than missing features a lot of people would need.
Marcus Moeller - FAmSCo
Fedora is a good choice for a lot of tasks, e.g. desktops but even servers. If you are in need of running the latest version of http, php or mysql why not stick with Fedora?
Jaroslav Reznik - BOARD
I'm desktop guy but I don't see desktop and non-desktop cases as mutual exclusive. Even there's no one desktop use case anymore - we have that full-featured desktop environments for your workstation, lightweight ones for netbooks, touch-enabled for tablets, even mobile ones. Same for non-desktop - servers running inside virtual machines, minimal footprint embedded systems etc. What we need is layered platform - the minimal one, even without X-server, the UX one - with X-server, Gtk, Qt, that all spins should implement and build user experiences on top of this base operating system platform. PS: I'm not saying there will be no conflicts!
Marcela Mašláňová - FESCO
Fedora is already oriented on Desktop. There are lot of people using it on server and we should put more work there. I support Server SIG. Packages should have correct dependency because more and more desktop stuff is installed on server and it's not easy to remove it.
Joerg (kital) Simon - BOARD
A more polished Desktop would bring more users, but i am not so sure also new contributors in Areas where we really need them - if we follow that community Pyramid this could be a way to grow. But as long as we have a good symbiosis with our sponsors and can have benefit on doing things together on a more functional level i love Fedora exactly because i can do things in all areas. So i think focus the desktop only would be a fail.
Stephen (sgallagh) Gallagher - FESCo
I don't think "Fedora" can be expected to focus on one particular aspect
of the operating system. I don't think that's the strength of our model.
The strength of Fedora lies in having any contributor be able to say "I
don't think this works quite the way it should" and have that person
decide to fix it.
I think the real focus on Fedora as a project should be to facilitate improvement wherever the contributors want it. We should be focusing on making Fedora an innovation platform. Sometimes I think we focus too much on "Fedora the Operating System" and not as much on "Fedora the Community".
David Nalley - BOARD
I don't think that Fedora should focus on the Desktop. I think that the various DE-SIGs should continue to work towards excellence, and that the 'reference F/LOSS desktop implementation' should be based on Fedora.
My largest concerns with focusing on the Desktop is that Fedora is two-fold. The first is that I don't think we can 'win' in the desktop arena. Desktop users generally don't care about freedom, which is sacred to us. Instead they care about flash (and rick-rollin'), patent-encumbred software, and proprietary - closed source drivers. In short, that's a loss for us.
Secondly, Fedora has long been a place ripe for innovation. I've often described (and probably heard from someone else) that Fedora is the PARC (Palo Alto Research Center) of F/LOSS. Fedora should be welcoming to innovation. Adjusting our mission away from that is largely redefining what Fedora is - I don't think that should be taken lightly.
What do you think of the Community Working Group that was just established?
Sandro Mathys - BOARD
I like the CWGs purpose very much and I offered to serve in it myself. Nevertheless, I'm not exactly sure whether the tasks for the CWG are not actually tasks of the board, the steering committees and the mentors. But the people who decided to create the CWG thought about that already for sure. Maybe a clearer delimitation should be made but I think what the CWG really does and is responsible for will only become clear once it's been actively working for a while.
Jaroslav Reznik - BOARD
In two and half year in Fedora community, I've never seen any problem requiring CWG intervention (even some cases were on the edge). If someone is willing to take care and run it - I'm not against but I prefer having "welcoming" community where it's not needed.
Marcela Mašláňová - FESCO
No opinion. I suppose friendly community is done by people in group. More restrictions or rules don't make us friendlier.
Joerg (kital) Simon - BOARD
I am not sure if we need that group. We have the people in charge in the Steering Commitees and Board. Through my work in FAmSCo i know we always improved the Conduct and Membership experience - the same i do expect from the Board. I think a community should be self regulating in some areas.
Stephen (sgallagh) Gallagher - FESCo
I bemoan the situation that has led to its creation, but I don't know
that it will lead to an improvement. Fedora is a living entity, and like
all complex systems build from independent parts, sometimes things don't
fit together smoothly. This is a simple fact, and no matter how much
lubricant the CWG applies to the gears, sometimes the teeth just don't
fit together.
I'm not saying we need to remove these gears (my analogy is failing a bit), just that we might want to work towards directing their energies elsewhere.
I don't know that adding additional bureaucracy is the way to defuse the situation, but I'm willing to give it a try and see how it plays out.
David Nalley - BOARD
I think it's sad that it was needed. At the same time, when I look at the list of 'tasks/responsibilities' (easy contact/writing a code of conduct/coordinating groups within Fedora/resolving conflict) for this group I have to wonder why this isn't the Board's responsibility. I also worry that we've set this group up, given them the mission of "maintain a friendly and welcoming community" and wonder if they have the authority to truly accomplish it. There have been previously empowered groups charged to make things friendly (with a more limited scope), and that appears to have essentially devolved.
That said, these people are in place, the board should let them get things done.
What if anything would you do about the number of more seasoned contibutors that are reducing involvement in the project?
Sandro Mathys - BOARD
Getting Fedora back into a state where things (can) get done and where bleeding edge is lived.
Jaroslav Reznik - BOARD
I think this is not a problem for us, if these contributitors do not left open source world (-> they left us too). There are many more open source projects than before and it's great.
Marcela Mašláňová - FESCO
This can't be solved by FESCO.
Joerg (kital) Simon - BOARD
I think that the concept of meritocracy is not really enforced enough in Fedora and seasoned contributors are tired of fighting the same fight every year after year if new contributors show up with ideas which were abandoned years before. We(including myself) waste so much time with explain the same things over and over again so the really important stuff - technically, budgetary get not the attention which it should have. Yes i think to draw a line earlier can help to lower the frustration rate. And i think because we solved that almost for the Ambassador Group with our mentoring programm it could help to use good educated and skilled Ambassadors to help here as well.
Stephen (sgallagh) Gallagher - FESCo
I think we get back here a bit to my discussions about mentoring and a
bounty board. There are many reasons why individuals reduce their
involvement:
Some are moving to other platforms. We can combat this somewhat by improving our focus on being the best facilitator for the future.
Some just face real-life issues: family, jobs, etc. In these cases, I think we should be providing an infrastructure to help transition these contributors out. Specifically, I'm thinking of making it easier for these contributors to find apprentices to take over their work.
David Nalley - BOARD
I think some turn over, and even some burn out is natural. Turn over is in fact a sign of a healthy community. I am a bit concerned at the scale of burnout. I don't know that it's a perfect solution, but I wonder if Fedora has stopped being a place where people get things done, and turned into a place where are often paralyzed by analysis (and bikeshedding), and have a need for permission for each step, as if we are afraid of failure. I think we should move boldly, get things done, if it goes wrong, admit it, learn from, and move on. Having a hundred (or more) count thread on $somedecision, seems to be a symptom of the problem and part of the burnout process.
In one sentence, can you summarize what the Board does and in another what FESco does and in a third what FAmSCo does ?
Sandro Mathys - BOARD
The board works on the vision of Fedora and the strategy to get there (and eventually to implement the non-technical points in the strategy). FESCo works on how to implement the strategy, mostly from a technical point of view. FAmSCo works on the strategy, and supports the implementation of it, regarding the ambassadors' main task which is to promote Fedora locally.
Jaroslav Reznik - BOARD
The Board has an idea, vision etc., FESCo and other teams implements it.
Marcela Mašláňová - FESCO
Board does the idea, where we should go, FESCO solves actual developer problems. I suppose FAmSCo works on popularization of Fedora project.
Joerg (kital) Simon - BOARD
All have to enable and protect the freedom to contribute in their certain area!
Stephen (sgallagh) Gallagher - FESCo
The Fedora Board is responsible for deciding the vision and mission of
the Fedora Project at a high level; it defines what Fedora IS.
FESCo is responsible for making all of the engineering decisions about the Fedora Operating System; it defines how Fedora WORKS.
FAmSCo is responsible for presenting the Fedora Project's public image to the world by arranging events and other marketing materials; it defines how Fedora is SEEN.
David Nalley - BOARD
The board is responsible for the Fedora trademarks, and enabling Fedora's contributors to get things done.
FESCo is responsible for the technical leadership of the project, including specifically the technical leadership of the distribution.
FAmSCo is responsible for empowering Ambassadors and overseeing Fedora's discretionary budget.