21:13:45 #startmeeting 21:13:53 Meeting started Wed Apr 27 21:13:45 2011 UTC. The chair is lightguard_jp. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:13:53 Useful Commands: #action #agreed #help #info #idea #link #topic. 21:14:04 #chair mojavelinux 21:14:08 Current chairs: lightguard_jp mojavelinux 21:14:08 git [12core] push 10master7 fa5b2d1.. 6Lincoln Baxter, III Added 'wait' plugin for scripting convenience 21:14:12 git [12core] push 10master URL: http://github.com/forge/core/compare/75acf61...fa5b2d1 21:14:27 lincolnthree: were you waiting for the mtg to start before committing that? 21:14:32 meeting? 21:14:38 oh.. 21:14:42 right 21:15:24 of course he was, he needs to earn his vanity badge 21:15:39 #topic Report on action items from last meeting - Jason 21:16:23 Sorry 21:16:36 #info All action items from last week have been finished 21:16:47 The modules that requested JIRA integration have it 21:17:07 how's the testing of that going? working well? 21:17:12 Please feel free to try it out, you may need to use [ISSUE-###] format 21:17:18 I haven't done any yet. 21:17:40 You may also notice a new JIRA about formatting 21:17:48 We're trying to get all the formatting done by the 1st 21:18:01 Leads, please do your best to make that happen. 21:18:27 rank and file for jbw :) 21:18:31 I think I got all the modules that had formatting issues, but may have missed some, please double check your module if you don't have a JIRA for formatting 21:18:48 and forge should use the jboss community formatting :P 21:18:59 all aboard 21:19:04 yes sir 21:19:08 very good sir 21:19:20 right away sir 21:19:24 Also emails for pull requests should go out now for each module 21:19:29 that's the price of using danternet 21:19:34 But only to the one who authored the code :) 21:19:47 formatting only concern java files ? or are there rules for pom and other xml files too ? 21:19:52 I noticed the reference docs are still using the old fomratting in the example codes bits, we should make an effort to update those too 21:20:08 yes, good point 21:20:14 We don't have any rules listed, but I think it's three spaces? 21:20:19 lightguard_jp: does that mean the module lead won't get an e-mail notification for pull requests? 21:20:38 no, it's the thank you when the pull request is merged 21:20:49 or, in github terms, closed 21:20:53 It's a thank you email to the original author 21:20:57 ah, no problem 21:21:20 that's a nice touch 21:21:26 yes, this is a very important step to a) showing appreciation and b) keeping the author in the loop, hence keeping the conversation going 21:21:30 indeed 21:21:34 I'm sure we could jazz it up a bit more, but it works :) 21:21:39 #info source code formatting should apply to code in docs 21:21:45 First bit of python code for me, not too bad :) 21:21:52 #info XML files also get 4 spaces, no tabs 21:22:17 We need to document that 21:22:24 In build and on the site too 21:22:30 Any takers? 21:22:45 that is on the site 21:22:58 Okay, we need to update the info in build as well 21:23:03 the build module* 21:23:19 http://seamframework.org/Seam3/DevelopmentGuidelines 21:23:24 Not sure if that info is exported in eclipse and netbeans 21:23:45 I think it's default in netbeans (4 spaces in XML) 21:23:52 might not be, I've never understood why IDEs (particularly eclipse) only have formatting profiles for Java, seems shortsighted 21:23:58 yep, it's the default in netbeans 21:24:20 one of the main reasons we are choosing these settings is that they are close, if not the actual, default 21:24:24 IDEA has profiles for all sorts of langs 21:24:31 idea ftw! 21:24:35 :) 21:24:40 I think it's default in IDEA as well 21:24:44 Good thing Max isn't here ;) 21:24:48 I plan on giving IDEA a try real soon 21:24:53 hahaha 21:25:12 the guidelines could use note about jira reference in commit message 21:25:36 bleathem: Ask Dan for a licence for Seam work 21:26:46 bleathem: have you tried JBoss Tools? 21:26:55 should be 4 21:27:00 yes, it has only one problem 21:27:04 We need to at least not the action item, hopefully someone will volunteer for it :) 21:27:08 it's built on eclipse :P 21:27:13 oops, ignore that 21:27:23 max's team has drastically improved the quality of the eclipse tooling, no question about it 21:27:28 from the maven plugin to the cdi experience 21:27:50 not saying necessarily there isn't anything better, but I commend his team's work, they definitely have the right vision now 21:28:04 and that's a good start 21:28:31 but I'm also excited to have tooling that is agnostic of IDE (**forge**) 21:28:40 is that all from last week followup? 21:28:52 #action the information in the build module needs to reflect spacing for xml files 21:28:57 Yep, I'm good 21:29:02 #topic Seam Spotlight - Shane 21:29:22 oh, i'm up 21:29:43 so, we're looking for a few more volunteers to write spotlight articles 21:30:22 so far we've had jason, john 21:30:29 and brian 21:30:42 i think gunnar volunteered to do the next one 21:30:51 however we have no-one lined up after that 21:31:07 so, any takers? 21:31:14 Dan could write up about some of the cool events in the servlet module ;) 21:31:38 i could probably do one for remoting 21:31:50 Or security 21:31:56 right 21:32:00 +1 for security 21:32:07 security is probably used more / has more questions right now 21:32:12 also, their being on in.relation.to is a temporary hosting solution 21:32:18 ok, security it is then 21:32:28 +1 for security 21:32:33 +1 security 21:32:37 I'd like to read one on Seam Persistence 21:32:41 We should also reach out to the Errai team and see if they have someone who would like to write about CDI integration 21:32:46 would like to see a "get started" example 21:32:54 speaking of which, has anyone seen stuart online recently? 21:33:01 Nope 21:33:06 I can do one on persistence, though not until a week after jbw (I need to find my sanity again) 21:33:10 holidays I think 21:33:14 ah 21:33:21 well cbrock could write one for errai/cdi 21:33:33 that would be good, also christian could do one 21:33:37 he did this 21:33:42 stuart was last seen on IRC almost two weeks ago 21:33:46 i'm sure a lot of people would find it interesting 21:33:51 http://errai-blog.blogspot.com/2011/04/latest-enhancements-to-errai-cdi.html 21:34:00 ah 21:34:04 but I don't think people are finding those entries, we could always syndicate them both places 21:34:08 this is exactly why we need the seam university site 21:34:14 that entry needs a bit more meat 21:34:19 to collate all these blogs that are spread out everywhere 21:34:52 #action sbryzak will author the seam spotlight on Seam Security 21:34:56 ok, so any takers for writing the next article after security? 21:35:19 persistence would be good, of mojavelinux or stuartdouglas want to write one 21:35:26 That's two weeks, right? 21:35:31 I'll do one for persistence after that (or stuart if I can manage to delegate it to him) 21:35:46 yep, we have validation this week, then security 21:35:54 so i'll lock in persistence after that 21:35:59 plus, I've recently done some work with persistence in my cdi tutorial, so i can just port that likely 21:36:04 and either dan or stuart can write it 21:36:09 #action mojavelinux or stuartdouglas will write about Seam Persistence next 21:36:34 We should see if we can get Jozef to done one on Rest 21:37:32 perhaps we should have either a wiki page or google calendar to keep track of this pipeline 21:37:45 sorry 21:37:49 git [12core] push 10master7 a998d40.. 6Lincoln Baxter, III Updated wait plugin output format 21:37:54 git [12core] push 10master URL: http://github.com/forge/core/compare/fa5b2d1...a998d40 21:37:58 it's tracked in the minutes 21:38:02 we can keep it semi-private if you don't want to build up expectations unnecessarily 21:38:06 right, I just mean a longer pipeline 21:38:55 i'll follow up on that 21:39:05 to me this sounds like it kind of falls into dev methodologies in the sense that we need an area to "plan" that is either private or semi so 21:39:36 nice segway into the next topic kenfinnigan 21:39:41 :) 21:39:46 hahah 21:39:50 #topic Briefly discuss development methodologies (please come prepared to discuss methodologies that you have used which worked) 21:40:10 + 1 for waterfall. We need design specs people! 21:40:20 This will be further talked about next week at JUDCon for those attending 21:40:36 so it's become fairly obvious in recent discussions reflecting on the state of seam that we have pretty much been shooting from the hip for about...forever 21:40:52 very true 21:41:14 It was fine when it was just a slow moving JBoss only project 21:41:21 as long as any methodology we choose doesn't have much overhead 21:41:28 remember, we have limited resources 21:41:35 We now have over 20 (over 30?) modules and it's simply not working well anymore :) 21:41:55 now, it's very important to keep the perspective that this is a community project and by no means do we want to get all bossy on people 21:42:01 Right, we need something easy to use / follow with minimal touch points 21:42:06 so agile is the word 21:42:13 And it works with a distributed team 21:42:18 +1 agile 21:42:28 +1 agile, or variation there of# 21:42:43 agile is a buzzword though. Which form of agile? 21:42:54 agile has a whole heap of principles 21:42:58 but I think where we are breaking down is that it's actually quite inconvenient for part-time volunteers that we don't have any real target points, then all of a sudden we have a super urgent one 21:43:05 we need to be specific about which ones we want to adopt 21:43:09 Does anyone have suggestions that have worked well for them in the past? 21:43:26 I actually watched a presentation recently that made a ton of sense, and really spawned this idea 21:43:31 RUP! (only kidding) 21:43:36 and in reflection, actually solved our problem directly 21:44:10 in the past, we have made promises of target dates, then had no real plan to actually get there 21:44:15 in this talk 21:44:19 is the presentation on the web that you can pass around? 21:44:58 of course we had a plan... work our asses off to meet our deadlines 21:45:06 it worked ;) 21:45:14 I like the idea of fixed timebox releases 21:45:19 Not very sustainable though :) 21:45:23 the speaker was saying that the idea is that set some goals, iterate (for a period of time) then revisit those goals, but two things are important 21:45:28 people like regular release schedules 21:46:09 1. don't assume that over time you will get better at estimating, because the fact is, the creative process just isn't that predictable 21:46:27 2. stay consistent with your iterations, but not necessarily forcing every feature you hoped in 21:46:43 can relate to 2 21:47:18 so the timebox release does seem to be the recommended way 21:47:22 I'd say we're better off having a larger number of smaller releases, than waiting for big catch-all releases 21:47:27 a la Jenkins 21:47:34 but maybe not to that extreme :P 21:47:45 then, also, people can plan their schedule for these releases 21:47:49 that's essentially the plan anyway 21:47:55 our bundled releases are due every 6 months 21:48:00 +1 to smaller but regular releases 21:48:11 they include whatever latest versions of modules are available at the time 21:48:16 just as an example, I think Ubuntu is popular largely because every 6 months, like clockwork, you know you are going to get something new 21:48:29 sbryzak: First time I've heard that 6 mo release schedule 21:48:43 so it's up to each module lead to get a release out before the bundled release date, if they have new features or bug fixes 21:48:48 right, but I think our modules should release at least 2 - 4 times in that period 21:48:52 maybe 2 -3, not sure the exact number 21:48:58 i think that's highly dependent on the module 21:49:06 a lot depends on the volatility of the module though 21:49:10 if a module is mature, it may have no releases 21:49:14 e.g. seam-servlet 21:49:23 some modules will become feature complete at some point, no? 21:49:33 well, naturally, yes 21:49:37 bleathem: exactly my point 21:49:41 is there ever such a thing as feature complete? 21:49:49 so perhaps it's more that we should round robin module releases every 2 weeks or so 21:49:54 there's always something new to add on 21:50:03 i think in some cases, yes a module can be feature complete 21:50:07 kind of like with the spotlight, we always have activity 21:50:15 for example, a module that integrates spring with cdi 21:50:24 and for those interested in multiple modules, it gives you a chance to change up your focus 21:50:28 once it works, there's no requirement for any additional features 21:50:49 mojavelinux: more inter-module development would be good. They are turning into silos 21:50:54 i think that point could be argued, but let's just assume it will or it won't 21:51:03 yeah, so perhaps we can start by saying 21:51:07 with a move to short releases cycles, I would propose we also need to re-evaluate the release process with a view to streamlining it 21:51:19 let's have a module release every 2 weeks on avg (to start, we can see if this is too long or short) 21:51:31 sorry, at least one 21:51:51 some modules have only one person committing changes. 2 weeks might be a little steep in those cases 21:52:06 let me clarify furthur 21:52:18 i think that saying we'll have a module release every 2 weeks is putting an artificial constraint on our modules 21:52:27 maybe we can recommend a range, and the individual modules can make their own committment to a regular release schedule inside that range 21:52:38 of all the seam modules, we should have at least one module release happen every two weeks 21:52:43 that's 1 goal 21:52:51 In two weeks we have a servlet release, then in two more we have a catch release, then in another two have a faces release, that's six weeks for faces to get things ready, probably plenty of tim 21:52:55 time* 21:53:05 yes, exactly 21:53:11 now, that's a goal, not a restriction 21:53:23 it's a goal for the sake of appearances though 21:53:28 that keeps things happening, and something that jason, shane and I can keep an eye on 21:53:44 i think we need to evaluate this on a per module basis 21:54:28 Is this enough to strike up a larger conversation next week? 21:54:34 I don't think it's arbitrary, because it keeps people interested in seam and keeps encouraging people to participate 21:54:40 Sounds like we have an idea where we want to go 21:54:45 by all means, this is just the start of the conversation on this 21:54:50 didn't have to have a full solution this meeting 21:54:58 sure, let's continue it over the next week or so 21:55:04 nope, so that should get you thinking 21:55:08 yes, plenty of food for thought 21:55:12 a document would be good 21:55:17 brainstorming 21:55:25 We didn't have any infos 21:55:29 lightguard_jp: do we want to record any of these ideas in the minutes? 21:55:35 Should probably put something in the notes so we don't lose it 21:55:40 :) 21:55:45 working on it... 21:55:50 Okay 21:56:13 btw, jozef has (possibly) volunteered to lead the compatibility module 21:56:17 #info we plan to release the seam stack every 6 months, but we need to est the first target date 21:56:24 however i don't think we're going to be able to implement the way we originally planned 21:56:40 because of the class conflicts issue 21:56:54 Ah, for Seam2 compat 21:56:59 #info we are considering setting a goal that we have at least one module release happen every two weeks to keep seam active and to give community new bits 21:57:03 yes 21:57:07 fantastic! 21:57:20 the original strategy was to just deploy the seam 2 app inside of a cdi container 21:57:28 and allow bi-directional DI between Seam 2 and CDI 21:57:49 that would allow the developer to gradually convert his seam 2 components to CDI beans 21:57:53 however 21:58:11 what we didn't consider was that there would be a conflict in classes 21:58:19 many of the seam 2 classes exist in seam 3 also 21:58:30 e.g. org.jboss.seam.security.Identity 21:58:35 there's many others 21:58:39 bummer 21:58:43 I guess a package rename for Seam2 would be out of the question? 21:58:48 #info we are looking for other ideas that would make the seam development process more tangible, but which would respect the distributed nature of the team and the volunteer-based membership 21:58:57 yes, a package rename won't work 21:59:04 so we're back to plan A 21:59:14 that would mean the app would have to be re-written that uses Seam 2 21:59:25 which is for the compatibility module to process all the seam 2 components and shoe-horn them into CDI 21:59:46 lots of questions last night at VanJUG re: migrating from Seam 2 21:59:50 i don't see any major problems doing it this way, it's just more work 21:59:55 jboss modules perhaps? 22:00:09 My job just got harder 22:00:13 mojavelinux: good idea to explore at JUDCon 22:00:20 mojavelinux: JBoss modules that is 22:00:24 indeed 22:00:46 two more topics, one quick one, speaking of judcon 22:00:56 anyway, i'll write up an e-mail for seam-dev and we can discuss it further there 22:01:01 #info next week's meeting 22:01:05 oops 22:01:09 #topic next week's meeting 22:01:49 #info no meeting next week on IRC, but for those at JUDCon, we will have an informal gathering/brainstorm/meeting after the hackfest monday ~21:30ish 22:02:22 we might even start during the hackfest...I'm looking at places where we can get a section of the bar/restaurant for eats, drinks and good times 22:02:30 #topic git workflow 22:02:38 Briefly discuss git workflows (possibly looking at implementing http://nvie.com/posts/a-successful-git-branching-model/) 22:02:57 now that we aren't in complete alpha/crisis mode, we can slow down and take branches more seriously in git 22:03:12 the URL proposes a really clean way to manage the development branch for modules 22:03:28 agreed, official feature braches would be useful 22:03:36 lightguard_jp: is going to pilot it in catch, feel free to give it a go for your module 22:03:43 it's something that totally sucked in SVN 22:03:48 and now that we have git, it's finally possible 22:03:58 can you add that link to the minutes? 22:04:03 that would then allow collaboration on a feature in a branch correct? 22:04:15 #link git workflow proposal http://nvie.com/posts/a-successful-git-branching-model/ 22:04:19 absolutely 22:04:25 that is cool 22:04:29 #link http://nvie.com/posts/a-successful-git-branching-model/ 22:04:34 #info it would allow collaboration on a branch, and it will keep master stable 22:04:41 I like it 22:04:54 mojavelinux: uri first, then desc 22:05:02 It's backwards 22:05:06 so that master branch only ever has commits for the release? am i reading that right? 22:05:14 Yep 22:05:23 like that idea 22:05:27 Mostly they'll be merge commits and maven spam 22:05:48 yep, and ci should likely build from develop 22:05:57 because master will be a tag, and that's already tested 22:06:06 Ah, good point 22:06:12 I need to change that in catch 22:06:16 #info ci should build from develop in this proposed git workflow 22:06:30 okay, we are going long, so we'll wrap it up 22:06:35 one last topic/recommentation 22:06:39 #topic documentation 22:06:49 if you haven't watched it yet, see http://blip.tv/file/4881071 22:06:56 #link http://blip.tv/file/4881071 writing good documentation 22:07:11 it's not the same old song and dance 22:07:15 Max posted that on twitter a couple of weeks ago, very great presentation about documentation 22:07:19 hello ! 22:07:25 gastaldi: Hey, welcome 22:07:38 it basically gives you an appreciation for the fact that if you don't document it, you might as well not wasted your time writing it 22:07:43 yikes, getting homework from the Seam community meetings! 22:07:47 because it's just the same as if it doesn't exist 22:07:51 lightguard_jp: Thanks ! How ´re doing ? 22:07:56 :) 22:08:00 I'll mention one point that he doesn't 22:08:05 mojavelinux: I heard that your presentation was great 22:08:09 #info writing documentation/blogs makes you famous and get jobs :) 22:08:13 true, no point having a feature if no one knows its there 22:08:18 hehehe 22:08:22 gastaldi: still in the community meeting, one sec 22:08:26 nop 22:08:52 #idea We should probably schedule a documentation review soon 22:09:05 sbryzak: Did we ever hear anything from the doc team at Red Hat? 22:09:14 indeed, that will be something jason you can I can start with, then when organized, take it to the team 22:09:29 okay, gotta roll, there are drinks beaconing :) 22:09:39 lucky you! 22:09:43 plus, gotta go talk up all the great work you guys have done 22:10:00 lightguard_jp: they didn't get around to reviewing our docs 22:10:08 doh 22:10:23 about writing doc. Do you guys have resources on docbook 22:10:29 arquillian talk went great at ete, tomorrow lincoln and I showcase java ee 6, cdi and forge! 22:10:38 tools, helpers, etc.. 22:10:51 * lightguard_jp stepping away for a sec 22:10:57 that's a good question antoine_sd and something that would be good for the handbook 22:11:06 just talked to ken rimple and he praised oxygen 22:11:11 not free though 22:11:16 i've used serna and it's very good as well 22:11:30 I saw topics on the ml long ago on the subject 22:12:27 mojavelinux: is meeting closed? 22:12:34 #endmeeting