Session Start: Mon Jun 26 23:01:26 2006 Session Ident: #twiki_edinburgh [23:01] * Now talking in #twiki_edinburgh [23:01] * Topic is 'http://twiki.org/cgi-bin/view/Codev/EdinburghReleaseMeeting2006x06x26' [23:01] * Set by SvenDowideit on Mon Jun 26 11:43:29 [23:01] Good evening/afternoon [23:01] hello Lavr [23:02] excellent preparation in release meeting topic, thanks a lot [23:02] hi kenneth [23:02] Indeed! [23:02] Thanks [23:03] here too [23:03] * CDot has joined #twiki_edinburgh [23:03] hi crawford [23:04] 'evening c [23:04] steffen, i have seen that you pretty much cleaned up the support web while i was gone [23:04] hi Peter [23:04] good holiday? [23:04] bravo, excellente [23:05] yeah, wanted to see if it was possible [23:05] yes, very nice holiday [23:05] I think I'm closing in on 10 open items or so, so it was :-) [23:05] the yellowstone park is a place i recommend anyone to visit at least once in a lifetime [23:06] Hi Stef [23:06] all the colors, shapes and smells [23:06] PeterThoeny: did you climb El Cap? [23:06] or just stick to Half Dome? [23:06] i am not a climber [23:06] oh, you are talking about yosemitee (sp?) [23:07] do you remember that canadian chap from WikiSym [23:07] very tall chap (I forget his name) [23:07] the former is in idaho, where all the geysirs are [23:07] the latter is in california [23:07] oh, ok. Both start with a "y" ;-) [23:08] i mixed them up too [23:08] *g* cleaning out the hash too early? [23:08] don't rember the canadian person [23:08] SteffenPoulsen: the hash? hash pipe, or hash browns? [23:08] * MartinCleaver has joined #twiki_edinburgh [23:08] hi martin [23:08] hi Peter [23:09] we have criticall mass to start [23:09] lynnwood is missing though [23:09] he is a good facilitator [23:09] is Kenneth here? [23:09] I was thinking only one entry left in "y"-hash-style .. not that the brownies doesn't sound interesting :-) [23:09] Yes [23:09] Preparing minutes. [23:10] yes, let's get started. Thanks, Lavr [23:10] Facilitator? [23:11] i can do it if you wish [23:11] that's settled then :-) [23:12] ok, kenneth for mm, me for talking [23:12] aka facilitating [23:12] Before we start - the agenda is a bit confusing. So you need to clarify Peter. The one liners do not match the sections further down [23:12] anything to add to the agenda? [23:12] http://twiki.org/cgi-bin/view/Codev/EdinburghReleaseMeeting2006x06x26 [23:13] * Revisit rule of changing API in non-backwards-compatible way - was agreed last meeting. Is this a carry over error? [23:14] I'd guess so. [23:14] oh, yes, that is a doc bug [23:14] just look at the toc [23:14] anything to add to the agenda? [23:15] nope, nothing for the agenda from me [23:15] ok, lets start [23:15] Who added Removal of !WebPreferences (an admin-oriented item) from normal users' left bars [23:15] That is a new one [23:15] martin i think [23:16] ---+ 1. Review Previous Action Items [23:16] TWiki Doc Users Focus group [23:16] It is number 5 then. I have saved the topic so it is updated [23:16] y, I did [23:16] * Soronthar has joined #twiki_edinburgh [23:17] * Twilight\ has joined #twiki_edinburgh [23:17] Welcome Twilight. That is a nick I do not know. A name for the minutes? [23:17] arthur, do you want to summarize our first meeting? [23:18] doc meeting [23:18] The doc group has met, and we have set a number of action items - [23:18] Twilight\ is n=ask@17.84-48-138.nextgentel.com * Anders Steinlein [23:18] Twilight\ on #twiki_edinburgh #twiki #tomcat #svn #django #postgresql [23:18] Twilight\ using irc.freenode.net http://freenode.net/ [23:18] Twilight\ is identified to services [23:18] Twilight\ End of /WHOIS list. - [23:18] We will create drafts of twiki.org home, Plugins home and Support home [23:19] we'll leave Codev as is for a while [23:19] Lavr: i'm just a curious twiki user hanging out in #twiki, coming here just to lurk ;) [23:19] Lavr: Anders is my name [23:19] You are very welcome [23:19] :) [23:19] Twilight\: You are very welcome, if you have a registered username at twiki.org, we will put a link to it in the minutes under "lurkers" :-) [23:20] I have a proposal for 4.1 that may/will impact the doc work in Plugins web [23:20] the main idea is to create 'easy' entrances for new visitors, evaluators and such [23:20] do you want to discuss that now, or later? [23:20] to to write new 'overview' topics [23:20] the doc team is probably meeting again this week, date not set yet [23:20] crawford, let's discuss that at the 4.1 focus [23:21] and to remove existing links that go directly into Codev and Plugins [23:21] so we've set our focus, but need to do our homework :-) [23:21] * ArthurClemens presses ourselves [23:22] (writing is hard, actually) [23:22] i think we had a good initial meeting [23:22] We have more time for doc work now that 4.0.3 is out of the door. [23:22] but a lot fo work ahead [23:22] "lot of" [23:23] If we follow my proposal timing wise later I plan to not test for a month and mainly focus on doc work [23:23] we've also suggested to use Google to index Support web [23:24] good idea [23:24] which is a new idea for TWiki [23:24] anything to add to the doc focus group part? [23:24] still we need better support for topic summaries [23:25] and topic titles [23:25] for your interest: http://developer.jot.com/WikiHome/2006-5-31-SeoPlugin [23:25] have you captured your goals somewhere? [23:25] (read another time) [23:25] Yes. I found that the Category searches in our distro docs return summaries that are best case useless and worst case confusing. So there is a low hanging fruit to persue there [23:25] yes [23:25] so we all know what you are working towards? [23:25] the doc group has a playground web [23:26] to write drafts [23:27] And avoid confusing users with unfinished proposals that we may throw away [23:27] should we make it public? [23:27] at some point, yes [23:28] we should consider it next meeting I guess, see where our homework has brought us [23:28] perhaps public but obfuscated: If you know where to look, you deserve to check what's in it [23:28] but i think it is better to let the doc team work out a 25% draft or so [23:28] I think having the doc team work in private is a recipe for disaster, FWIW [23:29] you will roll out something that everyone hates. [23:29] release arly, release often [23:29] early [23:29] so by all means have it read-only [23:29] but please share what you are doing. [23:29] yeah release early is the right thing, only history shows that this is a sensible topic so we need to reach a somewhat internal agreement first [23:29] thanks for the feedback, the doc team will discuss when to open it up [23:30] when we have highlevel goals / vision in place we will release I think [23:31] I don't have a problem in opening up now [23:31] If we have a meeting again end of this week we can open up right after. [23:31] that sounds good, let's put it on the agenda for next meeting [23:32] yes, the doc team will discuss this [23:32] next? [23:32] basically i am open for opening it up early if we get constructive feedback [23:33] ok, lets move on [23:33] ---++ TWiki 4.0.3 release [23:33] Sven did a great job on that [23:33] i have seen that all went well [23:33] with a follow up [23:33] yes, great job sven [23:34] For next release we need to automate making the -changes file. [23:34] And we need to make it as part of release candidates so we can test that also [23:34] even then we had some last minute changes that deserved more scrutiny [23:34] there are a few items left [23:34] as indicated in the Codev.ProductionReleaseChecklist [23:35] a.i. for peter: followup on 4.0.3 release [23:35] such as index.html update [23:36] anything to add to the 4.0.3? [23:36] add to the zip? [23:36] nope not from here - thanks to everyone involved in doing the 4.0.3 release also from me [23:36] no, the action item review [23:37] righty [23:37] yes, a big thank you to all involved :-) [23:37] it is nice to see that we have documented the process "good enough" [23:37] documentation of issues will be next (last?) in the meeting [23:38] yes, that is later, "visibility" [23:38] ---+ 2. Next Release 4.0.4 or 4.1 [23:38] at +38 min [23:38] depends what's in it [23:39] actually, we didn't document it well enough, there are still big holes [23:39] New features. Not just bugfixes! [23:39] I would not want to ask a less experienced person to build a release yet :-( [23:39] the build process? [23:39] y [23:39] The decision is simply to agree that next release is a minor where enhancements are allowed. [23:39] multitasking, CDot? [23:40] MartinCleaver: Kenneth did a proposal in the topic [23:40] vote: yes [23:40] * MartinCleaver checks again [23:40] I think we need to establish better what an "acceptable enhancement" is [23:41] otherwise some people might take it as carte blanche to include all sorts of changes [23:41] so we need a review process? [23:41] yes, that makes sense [23:41] vote: yes [23:41] i think we need a review process of new features [23:41] or at least some way of making changes highkly visible [23:41] create a Codev topic, explain the enhancement [23:42] let others give input [23:42] no, that isn't working [23:42] but who decides? [23:42] we don;t get input [23:42] perhaps express questionable features in some kind of "risk" evaluation? allow low / medium risk / low impact, deny high risk high impact [23:42] ArthurClements asked just the right question. [23:42] a report that shows all minor enh may be useful [23:42] I would rather have it driven by "customers" [23:42] A timeline when 4.1 would have to be finished would be great so that occasional programmers can see whether they've a chance to finish a contribution [23:43] CDot: continue [23:43] I would be quite happy, for example, to have to justify changes to Kenneth [23:43] or another similar user or set of users [23:43] CustomerTeam? [23:43] right [23:43] It is a balance. Someone need to want it. And someone needs to make it. [23:44] only, please don;t call them customers [23:44] we do have loyal users in Support [23:44] CriticalUserTeam [23:44] what's the operational difference betwen this non-named CustomerTeam and the "old" role of the CoreTeam? [23:44] with focus on kiss and unsability [23:44] unsanity, yes [23:44] the core team was a developer team [23:44] operational difference [23:45] "was"? [23:45] perhaps a new ReleaseManagerRole with a vote counting 4 or 5, other votes counting 1 .. for the cases where discussion ends up in non-agreement [23:45] (besides the fact that is the developer who will integrate the patch once approved) [23:45] and then changes the code :-) [23:46] well, strict rules don't apply [23:46] I think voting is not a good approach [23:46] I would be happy to submit to a board (whatever name it) [23:46] I would rather have a once-only review of ideas [23:46] me neither, we've seen voting periods extend over months [23:46] not a continuous process [23:46] as long as there is a mechanism to prevent the stagnation and frustration that happened pre-Dakar [23:46] We should start with the low hanging fruit. We should list the pending already coded - but not well tested - enhancements that are already in DEVELOP and agree which are merged in [23:46] right [23:46] i think that is a better approach, yes [23:46] we need a list of features [23:47] but the process stays the same [23:47] taht are in develop [23:47] fruit or not [23:47] most are no-brainers to put nto 4.1 [23:47] some need a discussion [23:47] there are very few features in DEVELOP [23:47] no-brainers make sense to add [23:47] so, the current ad-hoc process is "those in the meeting choose the features" [23:47] the question is how to identify the "some features" [23:47] that are not in 4.0.3 [23:48] Soronthar: yes [23:48] how about this: [23:48] Good. I can live with that [23:48] we create a report that shows the proposed features [23:48] too slow and ungainly [23:48] give 2 week time or so (until next meeting) for people to ask for a discussion of a feature(s) [23:49] we need to be able to *show* the features [23:49] so that people can try them [23:49] the features are in DEVELOP [23:49] and say "yes, i want that" [23:49] some of them, not all [23:49] the features are already in develop, so they are visible [23:49] and there is code in DEVELOP I do *not* want in a release [23:49] yes, that is why we need a vetting process [23:49] I would rather pull selected features into twiki4 [23:50] There are probably also some API stuff which is not "visible" on the user interface. And there is Merediths TAGS that we need to decide on. [23:50] separate the no-brainers from the ones that need a discussionm [23:50] I agree, we should avoid voting if possible .. but then we should be aware that we will de facto have a procedure that says "discuss / no commits until consensus" .. there's a difference between the daily added features / discussion and having an agreement on milestones set up front [23:50] hang on... this review process is actually rather simple [23:50] right, but new features can be in DEVELOP, then tested, before merged into TWiki 4.1 [23:50] it depends fundamentally on someone having a piece of work they would like to release [23:51] "the sponsor" [23:51] the review has to say "yes ok" or "no thats silly" [23:51] nothing more [23:51] there are also gray tones [23:51] "the release manager (group)" [23:51] it could be "yes, but we need to fix this and that" [23:51] fair point [23:52] * MartinCleaver has to leave in 10 mins [23:52] right now I am reluctant to use DEVELOP as an integration platform [23:52] it has had too many changes made in it without merging forward fixes from 4.0.3 [23:53] until someone puts the effort into mergeing those fixes, DEVELOP is essentially dangerous [23:53] agree [23:53] I would *not* base a reklease on it [23:53] (as I've experienced) [23:53] but features can be merged into TWiki 4.1 [23:53] any way to identify the 4.0.3 pieces that are not merged into develop? [23:54] for that reason i would rather wor forward on TWikiRelease04x00 [23:54] PeterThoeny: that is the work required to do a merge [23:54] I wroe a script to support that (to highlight the differences) [23:54] That is also my proposal. Current TWiki4 branch which is stable continues as the branch for 4.1 [23:54] but the main thing it shows is the difficulty of mreging [23:54] Lavr: good [23:55] so, basically, we proceed as follows: [23:55] someone has a proposal on Codev for an enhancement for 4.1 [23:55] they proceed to implement on TWikiReleas04x00 [23:55] knowing that they may be asked to revert it [23:56] The first thing in my proposal was that we all agree that we no longer need to reserve the TWIki4 branch as a bugs only branch because we do not plan a 4.0.4. And we plan a 4.1 in a not too distant future. [23:56] I support that. [23:56] casted mine [23:56] yes, kenneth's proposal is good [23:56] wait, can we interrupt and talk about martin's proposal now, since he has to leave [23:57] ok, so, we push features into 4.1, but carefully, knowing we may revert any of them [23:57] And the timing I propose - without any idea what is in 4.1 is to target August for beta and september for release. Just to give an idea of how much can go in. [23:57] ---+ 5. Removal of WebPreferences (an admin-oriented item) from normal users' left bars [23:57] this is a small item [23:57] martin, could you quickly elaborate? [23:57] (for "larger" features the sponsor should fork twiki4 branch and preferable use svk or similar to automagically merge twiki4 development to it often (daily)) [23:57] discuss in Codev -> raise Item -> do work [23:58] we do not need to discuss that in forum [23:58] hold on please [23:58] ok [23:58] that's right, next item is ok [23:58] lets give martin a chance [23:58] basically most people don't need to see the preferences of the web [23:58] so its info overload [23:58] see it if you are an admin [23:58] not otherwise [23:58] reduces clutter [23:59] oh, show it conditionally? [23:59] y [23:59] do we have IF member of AdminGroup? [23:59] I second Martin's suggestion - just remove it. [23:59] small change for kiss [23:59] shounds good [23:59] "sounds" [23:59] WebPreferences isn't changed, nor read, very often [23:59] any voices against? [23:59] ArthurClemens: I think I saw one in QuickMenuSkin [23:59] can consolidate on SitePreferences Session Time: Tue Jun 27 00:00:00 2006 [00:00] some people have vowed for graying out hidden items [00:00] If you look at the Preferences topic then there are some TWikiVariables in it. [00:00] Which are not documented elsewhere [00:00] but I would like to see it hidden [00:00] Lavr: The variables should get their VarFOO.txt topics [00:00] So that CDot's splitting of TWikiVariables makes more sense [00:00] i remember, colas does not like pages to look differently depending on users [00:00] SteffenPoulsen: good point on svk and personal branches [00:01] well children shouldn't see the same as adults [00:01] but this is a typical admin thingy [00:01] Neither do I actually. Exception is items visible only to adins [00:01] and marketing don't have the same interests as engineering [00:01] admins [00:01] not one user group vs another [00:01] +60 min [00:01] ok, I gotta run [00:01] byee [00:01] bye M [00:01] * MartinCleaver has left #twiki_edinburgh [00:01] ok, a.i. for martin: make the link conditional, and review other links for the same [00:02] But tailoring topics is so easy for anyone so I wonder why it is such an issue in the first place. [00:02] kenneth: good first time impression [00:02] we need to show the example sometimes [00:02] OK, FYI, the proposed enhancement I have for 4.1 is a restructured =configure= [00:02] i think it is good to show only relevant links [00:02] it is an interesting test-case [00:02] as it touches quite a lot of code [00:02] harald brings a good point to create VarFOO topics [00:03] for %BR%, %RED% etc [00:03] but as addition I would add a heading "Admin topics" [00:03] to make it clear to the admin as well [00:03] arthur: in every web? [00:03] to support colas [00:03] or just twiki web? [00:04] Actually it is TWikiPreferences that has the many variables. Not WebPreferences. [00:04] where admin topics are listed now [00:04] admins have the sitemap to access all webpreferences topics [00:04] it is about the left bar [00:04] i think it is ok to not show admin links in the sidebar [00:04] except for twiki web [00:04] but admins might want to add shortcuts [00:04] and bother noone [00:05] those admins who want shortcust can add them [00:05] the question is, do we need that by default? [00:06] our doc group will introduce a left bar with Newby topics, Skilled user topics, etc [00:06] :-) [00:06] yes, that is twiki.org specific [00:06] we can add 1 example for WebPreferences [00:06] A new TWiki ships with Main, TWiki and Sandbox. I have a hard time finding the issue of seeing a few links on the WebHome of those. New webs are normally created by WIPING the WebHome and start from scratch [00:07] yes, admin links on webhome is fine [00:07] admins can tailor it [00:07] so agreed to ermove admin links in sidebar? [00:07] "remove" [00:08] we can discuss this in detail later on [00:08] Absolutely [00:08] on Codev [00:08] On Side bar - No problem. I remove it on my TWikis also [00:08] do we need more discussion on this? [00:08] small change for kiss [00:09] if there are no objections i suggest to make it conditional [00:09] (I had missed the point that it was the default WebLeftBar we talked about. I thought about WebHome.) [00:09] _default, Main and Sandbox [00:09] not TWiki web [00:09] in distro [00:10] So minuted :-) [00:10] so, ok like this, or need more discussion? [00:10] objections? [00:10] no, unless more topics get involved [00:10] then we need more talk [00:10] ok, so lets do that [00:11] as for VarBR, VarRED etc [00:11] anyone takes that a.i.? [00:11] basically the ones in twikipreferences users might want to use [00:12] ...and related to that: Do we want to be VarXXX topics standalone? [00:12] can you explain? [00:13] Their "related" links in VarXXX topics don't work since they link to #OTHERVAR [00:13] instead of VarOTHERVAR or TWikiVariables#OTHERVAR [00:13] oic, good point [00:13] I was worried that you would hide the TWikiPreferences topic from normal view also in TWiki Web. If not then it is less an issue. [00:13] a simple solution is to only use TWikiVariables#OTHERVAR links [00:13] HaraldJoerg: add an Item; it's a 2 minute change [00:14] IMHO it ought to link to the varXXX topic [00:14] though..... what happens when the doc is extracted as a monolithic HTML? [00:14] do you still do that, Peter? [00:14] CDot: Exactly that's the problem! [00:15] If I use TWikiVariables, then linking to #othervar is just a browser issue [00:15] navigating in the big twikiavriables topic is much faster of links point to the same topic [00:15] there is another alternative [00:15] there is the big references manual that includes the important topics [00:15] which is to use one of our Publish technologies to generate the HTML [00:16] it includes only the twikivariables topic, not the individual VarXXX [00:16] so we are ok [00:16] no, we are not OK [00:16] can you elaborate? [00:16] because harald is proposing changing those links to link to the VarXXX topics [00:16] instead of anchor links [00:17] I'm not actually *proposing* that - I'd like to discuss what's better [00:17] since the VarXXX topics are not part of the monolithic HTML, the links will be broken [00:17] HaraldJoerg: hey, you stuck your head up.... ;-) [00:18] Yes, 'cause the current links are broken in VarXXX... [00:18] right [00:18] ok, that needs some more investigating [00:18] can we park this? [00:18] Yep. [00:18] +88 min [00:18] raise a topic in Codev [00:18] So does Harald take the action to investigate further= [00:18] ? [00:18] Yep [00:18] thanks [00:18] ok, lets go back to [00:19] ---+ TWikiVariables#OTHERVAR [00:19] sorry [00:19] ---+ 2. Next Release 4.0.4 or 4.1 [00:19] ok [00:19] the thing I was trying to highlight when I was intrrupted was the proposed changes to configure [00:20] which involves an "Extensions browser" [00:20] that allows you to browse and download extenisions from configure [00:20] it has two issues; first, it depends on a site being able to access t.o. [00:20] (which is a small issue) [00:21] the much bigger issue is that it depends on plugins being packaged using BuildCOntrib [00:21] Not when you have an authenticated firewall like we have. Then it becomes a bug issue [00:21] big issue [00:21] good point [00:21] so that it can auto-install them [00:21] at this meeting can we focus on the process? [00:21] you can still install "the old way" [00:21] and discuss individual features in the next meetings? [00:22] Peter, we just spent nearly 20 minutes discussing a minor change to the bloody menus! [00:22] ] [00:22] we can surely afford a small amount of time to discuss a potentially large impact issue like this! [00:22] crawford, if you feel like to discuss this feature in this meeting i am fine [00:23] ok, thank you. [00:23] yes,continue [00:23] but lets first agree on the process on taking featurs into 4.1 [00:23] OK, what I need to know is is it worth pursuing this [00:23] because the impact on the Extensions area is potentially huge [00:23] it would look very bad if 90% of the extensions couldn't be downloaded [00:23] because they were not "properly" packaged [00:24] any way to support the simple zips? [00:24] no [00:24] The configure script was a major improvement in TWiki4.0. If we can improve futher 1. First time installation and 2. Extension installation - then it will be a good thing. [00:24] What about Extension Configuration? [00:25] And it is the sort of thing that does not break topic compatibility at all. [00:25] also handled [00:25] CDot: Excellent! [00:25] what is the problem of built extensions? [00:26] crwaford, could you elaborate a bit more? [00:26] ArthurClemens: you can download and instal them, but it is high risk, for several reasons [00:26] I won;t go into it here [00:26] sorry, ok [00:26] the first is that the installer targets installations that may be configured over several dirs [00:26] Sounds like a candidate for a Codev feature topic - if not already there. [00:26] e.g. cg-bin instead of bin, libs in an odd place, data and pub miles away [00:27] zip defeats that [00:27] the second is permissions [00:27] the third is post-install. The zips do that step manually [00:28] ok, I have raised awareness; i will start a codev topic. [00:28] another one, updating existing topics, preserving history [00:28] excellent .. definetely an interesting feature [00:28] yes, good point [00:29] a.i. for crawford: raise topic in codev [00:29] back to process [00:29] back on the process for feature selection [00:29] 1. demonstrate feature in twiki4, being preopared to revert it if told no [00:29] Once you have a auto install tool for extensions you are very few steps away from auto installation of hotfixes. [00:29] 2. at this meeting, say yes or no to features [00:30] Lavr: yep [00:30] yes, we have identified two ways [00:30] proposal 1 is higher risk for sites that run from svn [00:30] I meant 1 and 2 as part of the same process [00:30] they shouldn't run svn from a develop version [00:31] propsal 2 is more time involved [00:31] not if the feature is already demoed [00:31] sounds ok to try this [00:31] CDot: Demoing hard work and then get it declined isn't very satisfying [00:31] that way the proposer gets feedback sooner rather than later [00:31] oic, you mean to merge into 4.1, then discuss [00:31] HaraldJoerg: I know; would you rather fight with stuff thrown over the wall? [00:31] ok, that is proposal 1 [00:32] PeterThoeny: zactly [00:32] I'd not give up the old-fashioned "Propose in Codev" approach before I start coding [00:32] nothing stopping that [00:32] I'd rather make a Codev topic *mandatory* for anything considered for 4.1 [00:32] Well. No matter what process. I hope people still will write feature topics on Codev and add agenda points on release meetings before starting a major development [00:32] I just feel a lot of proposals in Codev get lost [00:33] proposal 2: [00:33] - create report of features ready for 4.1 [00:33] - give time for people to "to be discussed" raise flag (two weeks?) [00:33] - discuss the raised ones only [00:33] - merge into 4.1 [00:33] evaluating / discussing already demonstrated features sounds like bliss in comparison to speculate on warm air [00:33] CDot: Many proposals in Codev get lost because they don't have a developer [00:33] true [00:34] exactly [00:34] If a developer writes a proposal in Codev and doesn't get objections he should go forward [00:34] that's the problem with TWiki (or OSS?) [00:34] good ideas, noone to create [00:35] So we can catch dependencies between proposals [00:35] HaraldJoerg: the only problem with that is that some recent proposals have been a bit.... unclear [00:35] which can make it difficult to comment [00:35] what is the feel on the process for merging features already in develop branch? [00:35] CDot: Absolutely! So it's up to us to ask for elaboration [00:35] PeterThoeny: I think the merge is a no-brainer, myself; it's just a question of "who" [00:35] how do we check quality of code? [00:36] no, not the merge process, the process to raise flags [00:36] ArthurClemens: that's a tough one [00:36] PeterThoeny: yes, I understand; i am saying, your process is fine, but "who"? [00:37] there si a lot of work in mergeing those features across [00:37] because the codebases have diverged (mostly due to the Tags) [00:37] isn't it simply the developer who implemented it in develop branch? [00:38] Would *you* like to try and persaude her? ;-) [00:39] "the sponsor" will need to do it, can be the original creator or anyone interested in seeing a feature in a release [00:39] yes, the tags one is something that needs to be discussed [00:39] The whole TAGS feature idea has never been validated from a performance point of view. [00:39] and from a usability point [00:39] usability? [00:40] i do not want to go into the tags detail at this meeting [00:40] when? [00:40] the usability of "TAGS", as opposed to Plugins, is exactly the same. [00:40] yes [00:40] each feature that needs to be dicussed can be raised in an upcoming meeting [00:40] let's elaborate on the process .. I think keeping it in two parts is fine 1st part: The feature is not done yet, 2nd part: the feature is demonstrated [00:41] whoops, forgot "in a codev topic" somewhere in that sentence [00:41] i am trying to figure out a process what works for all [00:41] Soronthar: would you like to sponsor tags? [00:41] make a case for it's inclusion in 4.1? [00:41] I am sympathetic, BTW, as I like the idea [00:42] btw, i have a hard stop at +120 min (in 20 min) [00:42] What do I need to to to sponsor tags? [00:42] merge to TWikiRelease04x00, and flag it for discussion [00:42] make a case for its inclusion [00:42] ok [00:42] hold on please, lets first decide on the process [00:42] and bet prepared to revert it if rejected [00:43] 1. merge then discuss, or [00:43] tags are "backported" from develop into Twiki04 via a plugin. [00:43] 2. give grace period, then werge if no objection [00:43] Merge first and then discuss???? Of a feature we can already see in DEVELOP?? [00:44] Lavr: the key point is *demo* first [00:44] i think proposal 1 is high risk [00:44] PeterThoeny: For new topics I think it depends on the code volume to be developed [00:44] A developer only has to try to waste hours of work ONCE to reject that idea in future. That won't work. [00:44] Lavr: the alternative is to accept *everything* if it is already coded [00:44] which is *not* IMHO an option [00:45] we *must* be prepared to reject code [00:45] crawford: the alternative is to merge after a grace period [00:45] CDot: Correct. Documentation must be a requirement before accepting a proposal [00:45] If it is code why merge? We can see it. If not coded - noone will code for a week and then have it rejected. Not without war and shouting. [00:45] how do you see that working? [00:46] Which is why people need to propose and document the idea FIRST. [00:46] crawford: the proposal i illustarted 15 min ago [00:46] Lavr: I think it is very hard to assess the value of an idea that way [00:46] * CDot looks back 15 mins [00:47] PeterThoeny: you mean proposal 2? [00:47] yes [00:47] I am worried about the rejection part. People need to know if the community buys in on an idea before you reject the whole thing. [00:47] But even if an idea is good the code can such and get rejected later naturally [00:48] PeterThoeny: sorry, i though you meant that for existing stuff in DEVELOp only [00:48] i think we need to distinguish between code alrady in develop and code to be developed [00:48] Lavr: if somone codes without discussing it first, they have to expect rejection [00:48] we are trying to discuss too many things at once, we need to split the process into manageable parts in a codev topic .. but a good idea would be to consider supporting both "doc first" and "code first"-style approaches, and perhaps only have a inclusion decision process for "doc done, code done" features [00:49] proposal 2 is for existing code in devlop [00:49] which is basically just Tags, i think [00:49] code to be developed should be discussed in codev and tracked in bugs web as usual [00:50] we still need the process for go/no go inclusion in 4.1 [00:50] I as a minute taker is not even sure what we discuss right now. We need - as always - a proper preparation on COdev to make a principle decision like this. And we do not sound too far apart anyway. It is practical things we discuss [00:50] Lavr: remember, rejection doesn't mean total rejection; it just means "not in this release" [00:50] it seems like we need some more time to think about the process? [00:50] the process is for deciding what gets into 4.1 [00:51] yes, sure, but let's not leave it too long [00:51] we are heading into holiday season, and it would be a shame to waste develioper efforts [00:52] If you like I will take an a.i. to draft a process [00:52] sounds like a deal [00:52] based on this discussion [00:52] Who takes the a.i. to start a Codev topic on this. .... thanks CDot. As I typed it you said it [00:52] done deal, appreciated [00:54] summary on "Next Release 4.0.4 or 4.1": [00:54] - agreed on kenneth proposal to wrok on 4.1 [00:54] crawford to prepare a proposal on process to accept code/doc [00:54] what about timeframe? [00:54] Yep. Already captured [00:54] first beta in aug as suggested by kenneth? [00:55] I proposed this for 3 reasons. [00:55] 1. We need to pass the summer holidays. [00:55] 2. It is close enough to catch up on bugs. [00:56] 3. it is late enough to get some actual work implemented and tested [00:56] i think this is sensible [00:57] any other input in timeframe? [00:57] nope, this sounds fine - this was ultimo August / primo September? [00:57] yes [00:57] yep [00:57] ok, we are almost at +120 min [00:58] Visibility of hotfixes for our releases is last item [00:58] shall we continue with remaining two items or stop here and call meeting next week? [00:58] two more items: [00:58] - 3. Next big release focus [00:58] - 3. Next big release focus [00:58] sorry, [00:58] - 4. Visibility of hotfixes for our releases [00:59] we can skip 3 for now [00:59] I thought the configure and feature process was 3 [00:59] That is where I captured the minutes for that discussion anyway. [00:59] release focus sounds a bit bigger to me [00:59] yes, but too much for now [00:59] i need to sign off in 1 min [01:00] if we continue, who takes the facilitator role? [01:00] Oh yes. It is big enough for 10 meetings. I propose at least touching the last item. We can then continue in the doc team [01:00] I say we defer remaining items [01:00] I'm tired, so would prefer to defer [01:00] if you feel like please continue with the visibility question [01:01] ok, then lest defer to next week [01:01] the last item will take a long time to discuss as well [01:01] we cannot decide this without your input here [01:01] OK. Then a 30 seconds decision. Are everybody OK to live with current TWiki403 release topic until we discuss further? [01:01] ok, meeting next week then [01:01] next meeting in one week, not two weeks? [01:01] any chance of a more sociable hour? [01:01] since it seems Australia can't make it [01:01] this time was for sven [01:01] ok, next: 2006-07-03 [01:01] zactly [01:01] too early [01:01] (for Sven that was) [01:02] can we go for 20:00 GMT? or even 19:00? [01:02] please discuss here and raise in codev for people to find a suitable time [01:02] i need to sign off, cheers [01:02] An hour earlier would for sure be better for me. And two even better [01:02] thanks for the good meeting [01:02] see you later Peter [01:02] bye Peter [01:02] See you Peter. [01:02] ciao [01:03] I'm in for both an hour and two hours earlier as well [01:03] but that will definitely remove Sven from the meeting [01:03] The only way to include Sven would be to have the meeting some hours *later* [01:03] (personally I would have it earlier as well) [01:04] He has already said he will not be here. We cannot change way the planet is made. [01:04] we could get up at 5 in the morning :-) [01:04] Early Morning Europe sucks in US but fits down under. [01:04] later will not work for me, but I don't mind Sven setting a time for every second meeting or so, see if it changes who is on board [01:05] http://twiki.org/cgi-bin/view/Codev/WhatIsIn04x01 [01:05] but if Sven already said he cannot join at this time, we should start earlier, definetely [01:05] Surely I can get up early. And I can usually also stay home until 9:45 AM. [01:06] ArthurClemens: that's fine with me, but 5 your time is 4 my time..... [01:06] you just stay awake [01:06] :-) [01:06] CDot: looks good [01:07] no, I need my beauty sleep. I *really* need my beauty sleep ;-) [01:07] Why Beauty sleep? It never helped. Give up [01:08] So what about two hours earlier next time - and let Sven suggest a time after that? [01:08] sure [01:08] +1 for Haralds suggestion [01:08] 19:00 GMT it is [01:09] OK. [01:09] fine [01:09] 21 CEST [01:10] 20 GMST or whatever that is called. [01:10] delta: -2, Sven for next delta [01:11] jolly good. Are we done? [01:11] I'm done, need some sleep now ... c you around [01:11] believe so. was quite fun today [01:12] Europeans are leaving for bed [01:12] ciao [01:12] Yes. It is bed time now. Thanks. Good night. [01:13] * ArthurClemens has left #twiki_edinburgh ("Leaving") [01:13] * CDot has left #twiki_edinburgh