17:35:34 #startmeeting FESCO (2011-01-19) 17:35:34 Meeting started Wed Jan 19 17:35:34 2011 UTC. The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:35:34 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:35:34 #meetingname fesco 17:35:34 #chair mclasen notting nirik SMParrish kylem ajax cwickert mjg59 mmaslano 17:35:34 #topic init process 17:35:34 The meeting name has been set to 'fesco' 17:35:34 Current chairs: SMParrish ajax cwickert kylem mclasen mjg59 mmaslano nirik notting 17:35:41 yo. 17:35:47 * mclasen here 17:35:57 * SMParrish here 17:36:00 made it this week, sorry about my absence last week. 17:36:01 * notting is here 17:36:04 Hi 17:36:05 #info cwickert unable to attend today, left feedback in tickets. 17:36:06 * mmaslano here 17:36:15 ajax is many timezones away 17:36:48 ok. 17:37:05 I guess lets go ahead and dive in... 17:37:08 #topic #516 Updates policy adjustments/changes 17:37:08 .fesco 516 17:37:09 nirik: #516 (Updates policy adjustments/changes) - FESCo - Trac - https://fedorahosted.org/fesco/ticket/516 17:37:18 I had two ideas from the ideas container this week: 17:37:37 1) have being sponsored into packager add you to proventester as well. 17:38:06 we would want to have some means of notifiying people/gettting them to read the proventester docs... 17:38:39 http://fedoraproject.org/wiki/Proven_tester 17:38:52 i can see that being good for increasing the pool. but they are disparate skillsets 17:38:58 true. 17:39:17 maybe if maintainers asked for proventesters group? 17:39:48 it's pretty easy to join, but currently it requires seeking it... 17:40:14 I'd really hope that any packager be capable of performing the proventester responsibilities 17:40:16 so, many maintainers may be unaware it exists off hand. 17:40:38 I guess I'm +1 to this, if we can advertise it appropriately 17:40:53 many packagers don't even test 17:41:12 Do they test their own stuff, at least? 17:41:15 so I don't think they'll be good proventesters 17:41:16 doubtful 17:41:28 possibly i am excessively pessimistic. :) 17:41:48 as far as implementing, we should be able to have fas autoadd them to proventesters and have that send out a 'hey, read the proventesters info' email... 17:42:27 if we can do that (add some information), then I'm for it 17:42:36 * SMParrish agrees 17:42:48 nirik, heh, that's not a bad idea... could make it automatic but have the link at the bottom of the info email. :) 17:43:00 * nirik looks to make sure thats possible. 17:44:00 i would assume that in the general case people will be responsible, and we can, uh, educate, repeat offenders. 17:44:10 I think it is, but it would send the message to anyone who was added, so we would have to make sure qa doesn't mind that. 17:44:36 jlaska / adamw: happen to be around? Any objection to adding a join message to proventesters? 17:45:04 * mclasen thinks a join message is a good idea anyway 17:45:15 mclasen, indeed 17:45:19 so then do we mass add all current packagers? 17:46:35 1117 packagers. 62 current proventesters. 17:47:20 what about sending out invites to the packagers so they can sign-up? Something feels weird with mass adding all packages with no idea whether they all want it 17:47:26 the trick is, of course, not increasing the pool of proventesters, but increasing the *active* pool 17:47:27 does the qa team think that is a good idea ? 17:47:30 I think we should plausibly bring this up on devel-list first 17:47:32 notting: yeah 17:47:46 But I think jlaska's idea is good 17:48:08 yes, only the interested will sign up 17:48:13 is there enough infrastructure in the process currently to add interested folks? 17:48:41 (the process currently has a trac ticket filed and someone asking the person that they read the guidelines, then adding them) 17:49:06 nirik: yeah, that's potentially time consuming if we get flooded with hundreds of requests 17:49:38 yeah, although we could ask people to say that they already read it in their ticket? 17:49:45 I suspect we can adjust that process based on demand ... it was more just to avoid silent group joiners and to encourage folks to say hello on the list 17:49:52 nirik: definitely 17:49:55 (if we are sending a solicitation to maintainers) 17:50:25 so, with this proposal, we would not auto add, but instead try and get interested maintainers joining? 17:50:39 Ok 17:51:04 how about new packagers? add something to the join docs? 17:52:03 so, proposal: Send email to packagers noting proventesters group and how to join and ask interested folks to join. Additionally, add note to maintainers join doc about it. 17:52:15 +1 17:52:16 * notting is +1 to that 17:52:20 +1 17:52:22 +1. 17:52:24 +1 17:52:38 +1 17:52:43 yeah, seems ok to me too... +1 17:52:55 would anyone like to step up to send the email? to amend the doc? ;) 17:53:52 * nirik listens to crickets. 17:54:29 i can do it. 17:54:38 Hurray. ;) 17:54:42 Both items? 17:54:47 sure. 17:54:54 #agreed Send email to packagers noting proventesters group and how to join and ask interested folks to join. Additionally, add note to maintainers join doc about it. 17:55:05 #action kylem to send email and amend wiki docs. 17:55:16 ok, second item. A fun one: 17:55:25 2. allow maintainer's to +1 their own updates 17:55:36 This was discussed in the past and rejected. 17:56:00 There are some folks who think it's a good idea however, so I thought it would be good to revisit. 17:56:34 heh 17:56:43 Why would a maintainer not +1 his own update. So if we assume every maintainer +1's then really we are saying karma is +2 17:56:48 if the packager is also a proventester that could be funny. 17:57:16 the +1 should really represent 'I have tested this', not just 'It built ok' 17:57:26 my thought in the past was that we assume the maintainer has tested their update before submitting it. 17:57:40 Yes 17:57:43 but yeah, perhaps they just built it and expect others to test, then go back and test it themselves. 17:57:50 There shouldn't be a "I really have tested this" option 17:58:03 Because they should already have tested it 17:58:09 IMO if we allow this karma should be set at +4 so we still get 3 more sets of eyes on it 17:58:09 or perhaps they took upstreams word for a bug fix, then went back and confirmed it was fixed. 17:58:36 SMParrish: i'd rather just keep the restriction there and not worry about changing the threshold 17:58:49 notting: I agree 17:59:12 heh. 17:59:30 * nirik isn't sure what the state of bodhi is with this. I know we requested it deny people from adding +1 to their own updates. 17:59:36 i have no objection to not counting maintainer +1 towards karma... 17:59:51 nirik: Last I heard that was in vcs, but not released. 17:59:52 nirik: it doesn't enforce that at the moment 18:00:02 ok 18:00:12 I dont see any advantage to letting the maintainer +1 their own updates 18:00:25 * mclasen agrees with SMParrish 18:00:31 ok, so shall we vote then? 18:01:34 -1 18:01:45 -1 18:01:53 what's the specific proposal we are voting on? 18:02:10 allow maintainer's to +1 their own updates 18:02:17 ok. -1. 18:02:18 :) 18:02:19 so -1 :) 18:03:14 #agreed the proposal to allow maintainer's to +1 their own updates is rejected. 18:03:20 sorry, I was -1 as well. 18:03:22 * gholms counts −4 18:03:24 Oh 18:03:55 -1 18:04:04 ok. I don't think we have anything on the next two items, but we can quickly touch on them... 18:04:11 #topic #515 Investigate a "features" repo for stable releases 18:04:11 .fesco 515 18:04:12 nirik: #515 (Investigate a "features" repo for stable releases) - FESCo - Trac - https://fedorahosted.org/fesco/ticket/515 18:04:27 I think cwickert was going to try and work on this. as well as some folks from the devel list... 18:04:31 but no news yet. 18:05:06 so, will move on unless anyone has thoughts on this... 18:05:25 #topic #517 Updates Metrics 18:05:25 .fesco 517 18:05:27 nirik: #517 (Updates Metrics) - FESCo - Trac - https://fedorahosted.org/fesco/ticket/517 18:05:33 also, not sure where we are here... 18:05:35 ah, crud, i was supposed to talk to lmacken about that. 18:05:39 but it totally slippedm y mind. 18:05:46 no worries. 18:05:53 i'll get on that for next week, promise. 18:06:16 ok, sounds good. 18:06:23 #topic #539 Meeting with the Board regarding strategic goals 18:06:23 .fesco 539 18:06:26 nirik: #539 (Meeting with the Board regarding strategic goals) - FESCo - Trac - https://fedorahosted.org/fesco/ticket/539 18:06:35 Do fesco meetings not use #action items? 18:06:38 so, on this one... perhaps everyone could list their 5 fav items? 18:06:53 * mmaslano 1,4,13,14 :) 18:06:53 gholms, we do 18:07:10 could all post their goals? 18:07:11 1 3 4 12 15 18:07:37 gholms: we do, I forgot. I blame the cold meds. ;) 18:07:44 ;) 18:08:05 1 3 5 12 15 18:08:08 *1*, 3, 7, 9, 15 18:08:19 cwickerts: 1, 3 18:08:24 1 3 12 14 15 18:09:14 1, 3, 4, 5, 14 18:09:40 * mclasen_ has to jump to a different meeting, will answer in the ticket 18:10:27 so, 1 - 7 votes, 3 - 7 votes, 4 - 3 votes, 5 - 1 vote, 7 - 1 vote, 9 - 1 vote, 12 - 3 votes, 13 - 1 vote, 14 - 3 votes, 15 - 4 votes. 18:10:34 (unless I miscounted) 18:10:42 Crap, you beat me to it. 18:10:53 5 has 2 votes. 18:11:21 1, 3, 15, 12 14 4, 5, 7 9 18:11:58 so, sounds like 1, 3, 15 are highly thought of, but 12 14 4 are all tied in the next rank. 18:12:41 so, should we just go with talking to the board about 1, 3, 5... or do we want to do one or all of 12, 14, 4? 18:13:09 I guess concentrate on the ones we agree on most strongly, and then see what kind of resources that takes 18:13:31 ok... 18:13:54 #info item voting: 1 - 7 votes, 3 - 7 votes, 4 - 3 votes, 5 - 2 votes, 7 - 1 vote, 9 - 1 vote, 12 - 3 votes, 13 - 1 vote, 14 - 3 votes, 15 - 4 votes. 18:14:06 ok, lets start with 1. 18:14:17 #info GOAL #1: Improve and simplify collaboration in the Fedora Community 18:14:25 how can we attain or help attain this? 18:14:27 git clone git://git.sugarlabs.org/dextrose3/dextrose3.git dextrose3 18:14:31 miss 18:15:20 encouraging more group maintenance? 18:15:27 pretty nebulous 'goal' 18:15:36 we all agree with it, but what is 'it' :) 18:16:07 yeah. 18:16:54 We have a pretty full array of mailing lists, irc channels and such... perhaps we could look at a way to reduce them some to simplify the amount of things people need to join/follow? 18:17:21 or have a better way to direct people to the communications channels they really are looking for. 18:17:51 maybe improvement in our infastructure 18:17:59 i agree with the second thing there. 18:18:12 perhaps 'tags' for mailing lists, right now they seem to be pretty much 'foldered' 18:18:16 encouraging co-maintainers is good too, but not sure that fits into this goal. 18:18:59 topics? or a different interface? 18:19:10 not that i've looked at our mailman interface in the last fewy ears 18:19:12 mo had some ideas for alternate mailing list interfaces 18:19:42 yeah, mailman has 'topics'. You can define a list, and then people can use those to filter emails... 18:20:32 well, the board wanted us to meet with them about the most important goals, right? (not have plans in hand) 18:20:46 yeah. 18:21:04 so, I guess we go on? if anyone can think of additional items for this, add to ticket? 18:21:08 there are mentioned spins in goal 1 18:21:24 nirik: we can certainly discuss it - do we have any other agenda items? 18:21:33 notting: nope. This is it. ;) 18:21:34 we should have opinion on spins 18:21:40 oh, right: see https://fedoraproject.org/wiki/Meeting:Board_meeting_2010-12-13 18:22:01 ugh, spins 18:22:11 +1 to deleting all spins and forcing everyone to use xmonad. ;-) 18:22:20 * nirik chuckles. 18:22:31 kylem: interesting proposal ;-) 18:22:50 hehe. 18:22:51 On the one hand, I really like spins, because I want to have a way to point people to Xfce and have them try it before they install, or any desktop/fedora for that matter. 18:23:03 I dislike that they are yet another install method thats not as flexable. 18:23:03 In terms of shipping code, spins are clearly useful 18:23:21 and that the people doing them aren't very involved, so I worry that they will be busted. ;) 18:23:21 In terms of introducing further QA for the Fedora brand, I'm not sold 18:24:19 It's okay to have examples of how you'd help implement this -- the important thing is that you're giving the board a prioritised list of goals that fesco can help with in some way. 18:24:34 creating easy process to add your own spin into Fedora would be perfect for goal 1 ;-) 18:25:08 mmaslano: Is making it easier to add different versions of Fedora an improvement in collaboration? 18:25:12 Would that be likely to fragment the community? 18:25:17 the problem is that if it's too easy, we will ship things that are not well done, which reflects poorly on us. ;( 18:25:50 10 different versions of Fedora could get confusing really quickly. 18:26:10 * nirik just had an idea, but not sure how feasable it would be... 18:27:21 have desktop, kde, lxde, xfce spins. Then, make some method of having an 'add on' iso/image/whatever for the things that are not desktops. Then you could for example download desktop + electronics-lab. 18:27:39 and electronics-lab is not a full spin, but just a package collection and some setup info. 18:27:41 nirik: split media installs are the devil. 18:27:43 hmm. 18:27:56 pjones: indeed. 18:28:04 just a thought. 18:28:23 nirik: say, a package group? 18:28:39 i'm not sure that having fedora be more and more different things is the answer 18:28:44 anyhow, I don't know that we will solve this here. But we could tell the board that we would work on it if people want... 18:28:54 Anaconda supports package groups; why not just let people push checkboxes like they already can? 18:28:55 notting: well, yes, but then you have to install first. 18:29:00 Oh 18:29:21 because you either need 64GB of ram, lots of persistent storage, or to have installed first 18:29:29 gholms: but you have to download some iso first 18:29:38 mmaslano: Like the install DVD? 18:29:44 also: how is having more separate fedora things increasing collaboration? 18:30:15 gholms: for example downloading dvd with gnome doesn't help me much ;-) 18:30:16 notting: it's not simplifying any for sure. ;) 18:30:32 lol. 18:30:38 sigh, next goal? 18:31:04 ok... 18:31:31 #info GOAL #3: Improve and encourage high-quality communication in the fedora community. 18:32:06 I think the items listed at the board meeting on that are good... 18:32:25 we could/should employ summarizers more. 18:32:42 and for meetings, docs would be good... 18:32:49 yeah 18:33:03 took me a bit to figure out best practices when hsoting the fesco townhall with meetbot 18:33:06 tribal knowledge and all that 18:33:37 what else can we do to help communication? 18:33:52 there is mention of fedora talk as well... but thats not practical for some folks. 18:34:03 and more than a few people on a call gets confusing fast. 18:34:09 we might clean up wiki 18:34:20 does ensuring we have a positive environment fit into this? 18:34:32 It may... 18:34:52 The CWG is working on a code of conduct and some general enforcement guidelines... 18:34:58 i've also heard rumblings that ftalk doesn't work as well as we'd like 18:34:59 cool 18:35:07 CWG? 18:35:18 Community Working Group 18:35:23 appointed by the board. 18:35:26 Ah 18:35:55 * drago01 hopes the "enforcement guidelines" don't end up being "hall monitors 2.0" 18:36:34 any other ideas for improving communication? 18:36:42 not here. 18:37:21 #info GOAL #15: Improve developer experience 18:37:28 This is one that should be in our domain... ;) 18:37:48 haha. 18:37:53 making rawhide more stable would be nice, but not sure I like too many of the ideas for that. 18:37:58 also nebulous though 18:38:20 #15 and #14 are kind of in conflict if that's the sense you take it at :) 18:38:34 I'd like to see us collect more feedback from maintainers... what's causing them issues, whats unclear, what part of their workflow is slow/anoying. 18:38:41 yeah. 18:39:10 drago01: perhaps Avatars of Excellence? 18:39:23 a straw poll of developers to see who's actually /using/ rawhide (as opposed to just keeping a vm around) would be interesting. 18:39:44 getting more people in some areas I think might help out... more rel-eng folks doing buildroot overrides, more infrastructure people doing hosted requests, etc 18:40:09 it's not only about rawhide, we could improve our tools like - fedpkg, pkgdb, ... 18:40:21 yeah, I think having a stupid poll of the week/day would add to fun/inclusion of community, etc. 18:41:23 hmm 18:41:25 i like that idea 18:41:31 mmaslano: i agree, and wonder how we, as fesco, can encourage/motivate/bribe people to work on our tools 18:41:48 mmaslano, any specifics on whats lacking? 18:41:49 notting: I ment we could ask maintainers what they are missing in tools 18:41:54 ah. 18:42:12 something like 'release notes' for rawhide could be interesting. 18:42:29 kylem: you mean a 'whats broken today' ? 18:42:46 a bit? 18:43:01 mmaslano: hm, so more frequent surveys? certainly doable. did we ever deploy limesurvey? 18:43:06 there was a rawhidewatch blog thing that warren did a while back. 18:43:06 something like this though: http://openbsd.org/plus.html 18:43:16 notting: it's stuck in review I think. ;( 18:43:19 but collaborative instead 18:44:06 kylem: you could collate the rawhide reports and rpm changelogs. but omg too much data 18:44:22 right. 18:44:32 and people don't generally do a good job with changelogs. 18:44:57 yeah, although improving that would be nice. 18:45:51 oh... a few more: 18:46:05 Try and clean out our needsponsor queue. 18:46:28 It would be very to get people contributing before they get bored and wander off. 18:46:32 very nice 18:46:42 corollary - fix group signup so you dont have so many drivebys 18:47:12 and another big one: revamp our non responsive maintainer process. 18:47:18 Is there anything that can be done to increase sponsor activity? 18:47:29 * notting is a crap sponsor. -ENOTIME 18:47:55 Something tells me a great deal of sponsors aren't very active in that respect for one reason or another. Would it help to find out why? 18:48:08 gholms: yeah, I think a lot of them are just busy... 18:48:25 I did mail sponsors a while back asking for them to try and sponsor someone, but not much came of it. 18:49:00 we have currently 79 people waiting to be sponsored (approx) 18:49:31 so, working on improving that would be good, IMHO. 18:49:35 some of that may also go back to target audience/motivation - if a particular person doesn't see 'more packages' as a end goal in and of itself, they're unlikely to be sponsoring random people unless they're working specifically in the area they're interested in 18:49:50 maybe we could get sponsors who have free time to post somewhere and then people can approach them directly 18:50:12 notting: yeah, some SIGs I think have sponsors that are active, but others don't... 18:50:22 nirik: what's the search method for finding packages that have packagers that need sponsors? 18:50:39 http://fedoraproject.org/PackageReviewStatus/NEEDSPONSOR.html 18:50:57 grouped by name and reviews. 18:51:56 any other thoughts? 18:52:03 who all can make the Board meeting next monday? 18:52:27 * notting can 18:52:33 me. 18:52:34 * nirik should be able to. 18:52:36 i think. 18:52:43 I can 18:53:10 * mmaslano probably can 18:54:07 Afraid not 18:55:27 ok, so anything further to add? shall we take those 3 to the board? or do people want to discuss some of the other ones? 18:57:25 * kylem checks again 18:57:32 abadger1999: does the board think 3 is a good number? 18:57:48 notting: Yes definitely 18:58:18 [Insert Holy Hand Grenade counting joke here] 18:58:21 The board will be narrowing the goals to focus on to 3 if possible before announcing/drumming up people to work on them. 18:58:32 *down to 18:58:32 I was wondering in fact if one wouldn't be a good number. ;) Ie, have each release a "overriding goal" that everyone looks at and tries to help with that release... 18:59:22 I think the Board was attempting a 2 release cycle but the same concept of "a goal to get people to look at improving in this timeframe" 19:00:42 yeah. 19:01:45 heh 19:01:47 fair. 19:03:24 ok, anything else, or shall we stick a fork in this meeting? 19:04:04 Open floor? 19:04:20 Not that I have anything for it, but... 19:04:55 nirik: fpc were asking about something? 19:05:05 #topic Open Floor 19:05:12 * nirik needs less or more cold meds. ;) 19:05:32 or a beer 19:05:38 yeah. 19:06:08 so, FPC was wanting fesco to list/approve exceptions to a 'shouldn't start by default' query on services/units 19:07:48 they don't want to do it? 19:08:00 seems not. 19:08:04 abrt was a sticking point... 19:08:24 I thought we pushed it off to them, but if they don't want it, I think we should go ahead and decide. 19:09:37 well if they won't or don't wan't to we can do it, but they have to live with our choices :) 19:10:13 they want decision about abrt or all of them? 19:10:21 * mmaslano thought what was on, will be on 19:10:45 I thought it was all of them, but perhaps we should wait until we get their request? 19:10:54 spot / abadger1999: what is the FPC requesting of us? ;) 19:11:04 opinion was divided about whose responsibility it was but the end vote was for fesco to make the list of exceptions. 19:14:07 So what fpc voted for was[09:34:24] #action No, but FESCo provides exceptions (and has the option of tasking FPC to determine exceptions). (+1:5, 0:0, -1:1) [09:32:15] abadger1999: We can certainly ask them if they want us to do it (like the bundling exceptions), when they review the list 19:14:07 so, exceptions to what? "Off by default" ? 19:14:25 Sorry for the bad formatting there. 19:14:25 The guideline voted was No [to any service being on], but FESCo provides [a list of] exceptions (and has the option of tasking FPC to determine exceptions). 19:14:25 So I guess there is wording there that fpc would make the list but some people thought that it was not in fpc's charter while other people did. 19:14:25 spot: Could you talk about that part since you're on the "It's not in fpc's charter" camp? 19:14:25 nirik: Correct. 19:14:41 nirik: If it is to be on, then it needs to be in an exception list. 19:15:10 nirik: We had the start of a list here: https://fedoraproject.org/wiki/User:Spot/DefaultServices 19:15:48 nirik: But fpc had heavy disagreement with 1) the things not in the explicit list of exceptions (blanket exception for local services, for instance). 19:15:59 nirik: 2) abrtd 19:16:14 The other things on the explicit list are a good starting point. 19:16:30 And the blanket exceptions are a good starting point for thinking about further exceptions. 19:16:42 pesky freenode 19:17:43 ok. I fear we will need to look at all that and discuss it next week... 19:18:03 Things like: if mysqld is configured to only bind to localhost, it would be acceptable under the blanket local exception; do we want to worry about local-user exploits? were things that stopped us from using the blanket local services exception. 19:18:05 19:18:56 So I guess there's two questions: 1) FPC or fesco makes the list (b/c it could be seen as an expansion of fpc's charter) and 2) if fesco makes the list, then what's the list? 19:19:10 And yeah, big topic, next week is fine. 19:20:42 abadger1999: ok, could you create ticket for us? 19:21:02 will do 19:21:18 thanks 19:21:31 ok. 19:21:31 everyone back now? 19:21:38 Looks like 19:22:14 cool. Anything else for open floor? 19:23:16 * nirik will close out in a minute if nothing else. 19:24:43 thanks for coming everyone! 19:24:47 #endmeeting