20:25:27 <joshmoore> #startmeeting First governance meeting
20:25:27 <PyTablesBot`> Meeting started Mon Aug 29 20:25:27 2011 UTC.  The chair is joshmoore. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:25:27 <PyTablesBot`> Useful Commands: #action #agreed #help #info #idea #link #topic.
20:25:46 <joshmoore> #link https://github.com/PyTables/PyTables/issues/6
20:26:14 <joshmoore> So I was thinking first we should brain storm about things we want to cover (but then again, I was thinking there would be more people)
20:26:21 <joshmoore> avalentino: anything particular on your list?
20:26:34 <joshmoore> scopatz had wanted to discuss the release schedule
20:26:49 <avalentino> yes, release scheule
20:27:13 <avalentino> and rodmap for release > 2.3
20:27:17 <joshmoore> agreed.
20:27:32 <joshmoore> There was also mailing lists.
20:27:41 <joshmoore> And then the question of how often we want to try to have these meetings.
20:27:50 <joshmoore> Sound like enough (or too much) to tackle?
20:28:16 <avalentino> yes
20:28:30 <joshmoore> you get the first pick :)
20:30:44 <avalentino> regarding release schedule for 2.3
20:30:54 <joshmoore> #topic 2.3 Release Schedule
20:31:17 <avalentino> IMHO we should release the firs beta as soon as possible
20:31:32 <joshmoore> right, at this point, you're basically just fixing bugs, right?
20:31:37 <joshmoore> (and warnings, etc)
20:32:24 <avalentino> we only miss release doc update
20:32:37 <avalentino> the rest is almost done
20:32:38 <joshmoore> I saw that and wanted to ask: what all's missing?
20:33:02 <joshmoore> The integration of pro is ok, no?
20:33:08 <joshmoore> (in the docs, I mean)
20:33:16 <avalentino> yes
20:33:45 <avalentino> the UG needs some lifting (at reStruvturedText level)
20:33:52 <avalentino> but is is OK
20:34:02 <joshmoore> #link https://github.com/PyTables/PyTables/issues/80
20:34:27 <joshmoore> How much time do you think? Can you tell me where it's off and I can get my hands dirty? Or is it something you've almost got done?
20:34:50 <avalentino> no 0%
20:35:20 <avalentino> we should add all new features fron the pro version
20:35:28 <avalentino> ... in the release doc
20:35:58 <joshmoore> ok, so we do that before the 1st beta?
20:36:03 <joshmoore> (so people can find the features)
20:37:04 <avalentino> yes IMHO
20:37:25 <faltet> IMO, the new features are basically 2: the new OPSI engine (indexing) and enhanced caching
20:38:21 <avalentino> make sense just point users to PyTables Pro page?
20:38:25 <faltet> enhanced caching is internal stuff, so no need to document a lot (perhaps some new params, tahts all)
20:38:35 <joshmoore> Just generating the latest docs. .... where were you thinking of adding the text, avalentino? There's certainly some from faltet already, no?
20:39:35 <avalentino> ?? RELEASE_NOTES.txt ?
20:39:50 <faltet> linking to the Pro page could be enough initially, I think
20:39:50 * joshmoore smacks head.
20:40:07 <joshmoore> Understood. I was thinking something was missing from the documentation for release.
20:40:32 <joshmoore> I can take a swing at adding a summary then if that's the goal.
20:40:47 <avalentino> perfect
20:41:00 <faltet> and all the indexing stuff is fully documented in the UG, so you can put some links to there too.
20:41:09 <joshmoore> #action joshmoore takes gh-80 and updates the RELEASE_NOTES.txt
20:41:20 <joshmoore> faltet: will do.
20:42:24 <joshmoore> ok. I'll push that tomorrow or Wednesday for review. Anything else before we push 2.3b1?
20:42:26 <faltet> one important thing: where are you going to hung the new UG?
20:43:19 <joshmoore> I was just looking into pages.github.com
20:43:24 <joshmoore> #link https://github.com/blog/272-github-pages
20:43:44 <joshmoore> So we could host one copy there. I don't know how easily you can work them into pytables.org, faltet.
20:44:07 <avalentino> also pypi can host docs
20:44:21 <joshmoore> then there's http://readthedocs.org
20:45:03 <avalentino> don't know much about it
20:45:11 <joshmoore> http://read-the-docs.readthedocs.org/en/latest/getting_started.html#import-your-docs
20:45:26 <joshmoore> several github/sphink projects that I've looked into recently use it.
20:45:37 <joshmoore> virtualenv, pip
20:45:54 <avalentino> I like it
20:45:59 <faltet> pytables.org is a virtual server owned by me, but I don't think is a good idea to continue using it
20:46:13 <joshmoore> readthedocs does look quite nice.
20:46:17 <faltet> hosting the manual in github-pages looks good
20:46:21 <joshmoore> faltet: understood.
20:46:47 <joshmoore> the github-pages would also be where I put the logs from MeetBot (if you're ok with that)
20:47:31 <avalentino> ok
20:47:32 <joshmoore> (here comes a little bit of MeetBot spam)
20:47:46 <joshmoore> #item 2.3b1 as soon as possible
20:47:55 <joshmoore> #item location for doc pages
20:48:52 <joshmoore> ok, I just added a new issue for that last one: https://github.com/PyTables/PyTables/issues/95
20:49:11 <joshmoore> What else are we going to need to do to get 2.3b1 out the door?
20:49:17 <faltet> the http://read-the-docs.readthedocs.org seems nice too
20:50:19 <faltet> I saw some tickets from antonio about Python 3 compatibility
20:50:21 <joshmoore> I'll have to look into it some more; I think you can do both.
20:50:43 <joshmoore> #item python3 compatibility
20:50:50 <faltet> do you plan to have Python 3 support for 2.3?
20:51:35 <avalentino> IMHO to much work for 2.3
20:51:39 <joshmoore> Personally I wouldn't think so.
20:52:10 <faltet> ah ok. just curious
20:52:24 <avalentino> there is some preliminar step we could address in 2.3
20:52:35 <joshmoore> I was actually going to ask about that under "What comes after 2.3?". Do we go to PyTables 3.0 for Python 3 support?
20:53:19 <avalentino> IMHO 3 (we also have a lot of deprecated features now)
20:53:44 <avalentino> Numarray, numeric, HDF5 1.6 (??), netcdf ..
20:54:22 <faltet> well, PyTables 3 could be a nice match for Python 3 support ;)
20:54:32 <joshmoore> Agreed. :)
20:54:42 <joshmoore> And while we're doing all that, it might be necessary/beneficial to have a 2.4. We'll just have to wait and see.
20:55:02 <avalentino> agreed
20:55:36 <joshmoore> Ok, so assuming the notes and docs were ready by next week, would we want to do 2.3b1 then?
20:55:45 <faltet> I'd also deprecate HDF5 1.6 for PyTables 3
20:56:00 <avalentino> +1
20:56:17 <avalentino> and what about python 2.4  and 2.5
20:56:17 <faltet> but would continue supporting it (HDF5 1.6) for 2.x series
20:56:23 <joshmoore> I'm adding a 3.0 milestone and an issue for that, faltet.
20:58:26 <joshmoore> we (openmicroscopy) still have 2.4 support planned, but at some point it will end.
20:58:30 <joshmoore> hey scopatz!
20:58:32 <scopatz> Hey Everyone
20:58:35 <faltet> don't know about python 2.4 and 2.5. my suggestion would be to keep compatibility with them in 2.x series, and restrict for 2.6 (perhaps 2.7) for PyTables 3
20:58:35 <scopatz> Sorry I am late
20:59:01 <avalentino> hi anthony
20:59:07 <faltet> welcome Anthony!
20:59:15 <scopatz> Hello all!
20:59:28 <scopatz> SciPy still supports 2.4
20:59:34 <scopatz> which I think is a mistake.
20:59:48 <faltet> :)
21:00:04 <avalentino> faltet +1 for keeping support in 2.x series
21:00:16 <avalentino> so probably we also need a 2.4
21:00:17 <faltet> probably a lot of scientists would think like this
21:00:33 <faltet> *would not*
21:01:30 <faltet> IMO, keeping Python 2.4 compatibility in PyTables 2.x series would be nice for users
21:02:05 <scopatz> Hmm the advantages to moving to 2.6-2.7 are really high, though.
21:02:12 <joshmoore> scopatz: I just sent you the IRC log.
21:02:20 <scopatz> (thanks)
21:02:31 <joshmoore> (np)
21:02:36 <faltet> more advanced development can take place in PyTables 3.x branches
21:02:40 <joshmoore> scopatz: can you outline the advantages?
21:02:49 <joshmoore> #item py24 / py25 support
21:03:29 <scopatz> with statement, new style formatting, set and dictionary comprehensions, argparse
21:03:37 <scopatz> New AST module
21:04:05 <scopatz> These are probably the main ones that I use daily.
21:04:18 <joshmoore> argparse we could ship (openmicroscopy does)
21:04:27 <faltet> and PyTables 2.x series would remain mainly for maintenance (perhaps receiving some improvements from 3.x)
21:05:17 <joshmoore> in a way, this isn't much of a release other than being Pro. (barring the documentation beautification which is great!)
21:05:19 <faltet> but manitaining both 2.x and 3.x branches is much more work, of course...
21:05:21 <scopatz> faltet, If that is the case, then yeah I see no problem with supporting old version
21:05:32 <joshmoore> So, it would certainly seem a bit unfair to pull the python24/25 rug out from people.
21:06:34 <scopatz> Yeah...
21:06:42 <faltet> it basically depends on you, the new maintainers
21:06:42 <scopatz> I think you have me convinced.
21:07:32 <joshmoore> so jumping back a question or two: when do we want to shoot for 2.3b1?
21:07:52 <joshmoore> (scopatz: in case you haven't finished skimming the log, that essentially means getting release notes ready and the documentation up somewhere)
21:08:41 <scopatz> (thanks)
21:09:25 <scopatz> Would sometime in the next week or two be reasonable?
21:09:58 <avalentino> that was my idea
21:10:12 <joshmoore> I think so. I'll do the doc dance before travelling this weekend. If you guys can handle any of the other loose ends, then we should be good.
21:10:13 <scopatz> faltet, how much longer will pytables.org be available?
21:10:46 <faltet> 8 months aprox
21:11:18 <scopatz> OK, it would be nice to migrate some of the misc docs on the website to the sphinx docs too (FAQ, etc.)
21:11:25 <scopatz> Should be pretty easy.
21:11:35 <scopatz> (faltet: Ok thanks)
21:11:37 <joshmoore> for 2.3?
21:12:00 <scopatz> Maybe, it could wait too.  it isn't 2.3 critical
21:12:26 <avalentino> agreed
21:12:39 <scopatz> That is basically if we want to drive our entire webpresense from the sphinx docs.
21:13:10 <joshmoore> scopatz: which could make sense? I've noted it down for a later meeting.
21:13:24 <scopatz> So maybe we should shoot for a 2.3 release Thursday of next week?
21:13:24 <joshmoore> then that mostly leaves testing for 2.3, eh? didn't anyone get to look at the CI builds?
21:13:34 <joshmoore> #link https://github.com/PyTables/PyTables/issues/7
21:13:45 <faltet> I don't know for 2.3, but for future release, I'd merge some of the blurb in pytables.org/moin/PyTablesPro in the Introduction chapter of UG. that should be easy
21:14:06 <scopatz> faltet: definitely.
21:14:28 <joshmoore> ah man! That's what I was going to take for the release notes. ;)
21:14:46 <scopatz> haha you still can ;)
21:15:48 <scopatz> I am still getting used to jenkins.  it seems like it is building OK though.
21:16:20 <joshmoore> I still need to turn on heavy
21:16:22 <joshmoore> #link https://gist.github.com/1174568
21:16:30 <joshmoore> And nosetests completely went sproing on Windows.
21:16:49 <faltet> https://python.g-node.org/wiki/schedule
21:16:58 <faltet> oops
21:17:28 <joshmoore> we'll strike that from the record ;)
21:17:50 <faltet> I meant that having a new release of PyTables soon would be nice for our next Python school in the middle of September
21:18:08 <joshmoore> Then let's do it!
21:18:12 <joshmoore> so is there anything else in the way of testing we need/want?
21:18:32 <joshmoore> scopatz: do you want notifications from the jenkins builds?
21:18:40 <joshmoore> #item Testing / Jenkins builds
21:18:51 <scopatz> Yeah, that's be great
21:18:59 <scopatz> *that'd
21:20:10 <joshmoore> I just kicked off a heavy build. We'll see how that looks in the morning.
21:20:33 <faltet> josh, do you think that the testing platform could cope with heavy tests?
21:20:56 <faltet> I'd relax that and just run the light suite
21:20:57 <joshmoore> what do *you* think?
21:21:13 <joshmoore> Ok.....*grabs a bucket of water*
21:21:31 <faltet> in my experience, the heavy suite should be run once or twice before a release, not more
21:21:45 <scopatz> +1
21:21:58 <joshmoore> Ok. So we do that next week.
21:22:06 <faltet> 3 hours running tests is quite a lot. 5 minutes is handier, IMO
21:22:07 <joshmoore> #action run heavy tests before 2.3b1
21:22:56 <faltet> another thing: you insist with 2.3b1. I suggest 2.3rc1
21:23:08 <joshmoore> heh. We have on the order of dozen machines for testing, so it's certainly manageable once a day.
21:23:14 <joshmoore> Ah, better.
21:23:17 <joshmoore> 2.3b1 is habit.
21:23:25 <joshmoore> I'd vote for that.
21:23:40 <joshmoore> all?
21:23:47 <faltet> rational: 2.3 is basically 2.2.1pro, which was completely stable
21:23:58 <faltet> but there is not new functionality to test
21:24:06 <scopatz> yes, I agree.
21:24:12 <joshmoore> just documentation changes to proof read. :)
21:24:28 <scopatz> and a couple of bugfixes
21:24:31 <faltet> but it is your first release, so I understand thta you want to be on the safe side
21:24:46 <joshmoore> no, I'm always like this. *lol*
21:25:02 <avalentino> who long is the rc period?
21:25:13 <joshmoore> #item 2.3rc1
21:25:14 <avalentino> ... usually
21:25:41 <faltet> for me, rc1 would take a week, and a rc2 another week. final would come a week before...
21:25:51 <scopatz> It varies widely.  I would be fine with a week
21:26:35 <avalentino> rc code have to be considered completely frozen?
21:26:54 <faltet> three weeks for final in total after rc1 release
21:27:43 <scopatz> Antonio, no the point of the rc is for people to hammer on it and find any critical bugs that need to be changed before the final
21:27:49 <faltet> yes, completely fron, except for fixing bugs, of course :)
21:28:00 <faltet> *frozen*
21:28:20 <avalentino> OK
21:28:30 <scopatz> freature freeze, not code freeze ;)
21:28:47 <faltet> exactly :)
21:29:09 <scopatz> *feature
21:29:42 <avalentino> well I was thinking to improve py3k compatibility a little more for 2.3
21:30:00 <avalentino> bot maybe it is not the a goo idea at this point
21:30:06 <avalentino> *but*
21:30:37 <faltet> well, improving Py3 this is a feature, clearly...
21:31:08 <joshmoore> I'd concur. More code changes means I'll get nervous again. And none of us wants that, do we?
21:31:37 <avalentino> OK :)
21:31:47 <scopatz> Yeah it could probably wait until after the release.  Of course, you could always work on it in your own fork or branch and just wait to merge it back into master.
21:32:23 <faltet> IMO, I'd create a new branch for improving py3k compatibility, that will eventaully become PyTables 3.x...
21:33:32 <joshmoore> avalentino: there's a py3k milestone, so let's add a few branches and run with it.
21:33:54 <joshmoore> Ok. 2.3rc1 to go out next week which is very cool, py3k in a branch which is also cool. We're at a bit over an hour. Can I suggest we handle some of the less exciting things?
21:34:41 <avalentino> ok
21:34:45 <scopatz> Go for it.
21:34:53 <joshmoore> #topic mailing lists
21:34:54 <faltet> sure
21:35:14 <joshmoore> Assuming, we don't have these meetings too often, we're going to probably have some channel of communication.
21:35:17 <joshmoore> (for governance)
21:35:27 <joshmoore> There was also the suggestion of moving pytables-users to google groups.
21:35:56 <joshmoore> We don't have to deal with that right now, but if we're adding a new list, might make sense to pick where we want to eventually end up.
21:36:33 <joshmoore> Alternatively, we can discuss things vit github and periodically do IRC. Just setting up this meeting was a bit of a rollercoaster. If we could make it easier (and transparent to the community) that'd be really great.
21:36:58 <scopatz> We could have a users list as well as a dev list
21:37:07 <avalentino> +1
21:38:23 <faltet> before there was not much point in having a dev list, but with three developers now, it makes a lot of sense
21:38:45 <scopatz> the dev list should be public though as well.
21:38:56 <scopatz> I think we want as many devs as possible ;)
21:38:57 <joshmoore> scopatz: agree completely. the more people the better.
21:39:05 <avalentino> of curse
21:39:18 <joshmoore> more true than you intended, avalentino!
21:39:29 <faltet> sure
21:39:34 <avalentino> :)
21:39:45 <joshmoore> ok. are we going for google-groups then?
21:39:50 <joshmoore> are there any other outstanding choices?
21:40:10 <scopatz> I am fine with googlegroups, if only because it is easy and not sourceforge
21:40:54 <faltet> and has google visibility, which is *important*
21:40:58 <avalentino> googlegroups is OK
21:41:13 <joshmoore> ok, that sounds like a decision. Creating pytables-dev now....
21:41:24 <joshmoore> #action create pytables-dev
21:41:32 <joshmoore> Simiarly, when do we want the next meeting? how often?
21:41:36 <joshmoore> Only when needed?
21:41:56 <scopatz> well probably whenever we want to do releases.
21:42:08 <scopatz> I am in favor of a regular release schedule
21:42:16 <scopatz> once every 3-6 months.
21:42:35 <scopatz> 3, 4, or 6 are really the only ones that make sense
21:43:46 <faltet> why regular releases?
21:44:28 <scopatz> It motivates developers to fix bugs, update docs, etc if they know that there is a release coming
21:45:11 <scopatz> Maybe that is not needed as much with mature packages like PyTables.
21:46:28 <avalentino> maybe we could set a target release date for PyTables 2.3 + 1
21:46:33 <scopatz> I would be willing to play this by ear, but we should probably meet at least once per release.
21:46:41 <avalentino> and then fix the release period
21:47:08 <scopatz> That works for me.  I don't think this has to be decided now.
21:47:55 <joshmoore> Like I said earlier, I'm in a chat room all day with my dev team and we have 3 conference calls per week, so I'm really fine with almost any frequency of meeting that you guys are comfortable with.
21:48:00 <joshmoore> :)
21:48:54 <faltet> ok then, so it seems you are in agreement :)
21:48:59 <scopatz> Haha I am not up for that much! ;)
21:49:12 <joshmoore> :)
21:49:13 <scopatz> faltet: Yeah, I think so
21:49:34 <faltet> nice to see that we will have a 2.3 release anytime soon
21:49:58 <joshmoore> Hmmmm....ok. Then I'll take it upon me to bother everyone at some point for a next meeting (it'll be easier with the dev list).
21:50:07 <faltet> that will make me (and my pupils!) more confident for my next course
21:50:20 <joshmoore> here, here.
21:50:51 <faltet> ok guys, time for some sleep
21:51:21 <faltet> good luck with the release, and don't be afraid to ask if you need some help from me!
21:51:26 <joshmoore> agreed.
21:51:29 <joshmoore> faltet: will do.
21:51:52 <faltet> bye
21:51:54 <avalentino> thank you Francesc
21:52:11 <scopatz> Thans faltet!
21:52:24 <joshmoore> scopatz / avalentino: http://pytables.github.com/ is up. readthedocs is building. and googlegroup invites have been sent.
21:52:54 <joshmoore> if there's anything you don't have access to that you think you should, speak up.
21:52:57 <scopatz> Fantastic!  Thanks.
21:53:14 <joshmoore> i'll post the notes here shortly.
21:53:18 <joshmoore> #endmeeting