15:07:46 <infapi00> #startmeeting
15:07:46 <Services> Meeting started Thu Feb 12 15:07:46 2015 UTC.  The chair is infapi00. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:07:46 <Services> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:07:53 <infapi00> #topic Progress towards 3.16
15:08:31 <infapi00> #info infapi00 still has in his TODO list the GDM bug and the event reordering bugs for gnome-shell
15:08:59 <infapi00> #info infapi00 added a comment ping on bug 729603, without no response (that bug has a patch)
15:09:10 <Services> 04Bug http://bugzilla.gnome.org/show_bug.cgi?id=729603 normal, Normal, ---, gnome-shell-maint, UNCONFIRMED , Regression: GDM user list buttons no longer have accessible labels
15:09:16 <infapi00> #info during this week I will try to ping them on IRC
15:09:17 * clown looks
15:10:01 <infapi00> #info there is also a discussion about bugzilla products, but that is not exactly "progress towards 3.16". I will raise that point on misc time
15:10:02 <infapi00> done
15:10:40 <joanie> #info Joanie is still working on the generic-izing of Orca's support for Web content, fixing many bugs in the process and improving performance.
15:11:03 <joanie> #info She is not sure if that will be done in time to have an Orca-controlled caret for WebKitGtk by 3.16, but she'll try.
15:11:24 <joanie> #info Failing that, we can still "advertise" the improvements for Firefox support.
15:12:13 <joanie> #info Mike has a new patch for Evolution (which she will let him discuss). But at some point we should probably decide what the meaning of AtkValue is, and what sorts of things it does and does not apply to (which is also related to some W3C stuff).
15:12:18 <joanie> (done)
15:12:25 <clown> can you explain "generic-izing"?  In 25 words or less?
15:12:29 <joanie> sure
15:12:57 <joanie> Right now, Orca has the means of controlling the caret (including piecing together lines, dealing with embedded objects, providing structural navigation, etc., etc.) for Gecko content.
15:13:04 <joanie> oops, close to 25
15:13:06 <joanie> but almost done
15:13:23 <joanie> For WebKitGtk, however, Orca has been relying upon WebKitGtk native caret nav, etc.
15:13:30 <joanie> That is problematic for a couple of reasons:
15:13:45 <joanie> 1. Doesn't work like Windows (or even Orca in Firefox)
15:14:05 <joanie> 2. Hidden/offscreen stuff are never found (and sadly everyone seems to expect it to be found)
15:14:32 <joanie> 3. (Ok, more than a couple) Things like MathML in WebKit is done via flexbox which totally breaks native caret nav.
15:14:48 * clown makes note to read up on flexbox.
15:14:51 <joanie> Thus what was just a sad hack (thousands of lines worth of code) for Gecko
15:15:03 <joanie> is going to be a non-hack "feature" of Orca with Web content
15:15:14 <joanie> I can also use this to work around evince caret-nav bugs
15:15:17 <joanie> etc., etc., etc.
15:15:28 <joanie> so genericizing really is generalizing into a new module
15:15:31 <joanie> make sense?
15:15:37 <clown> yes.
15:15:48 <joanie> sorry for the > 25 words
15:16:02 <clown> does it mean that Orca will not rely on WebKitGtk native caret support going forward?
15:16:14 <joanie> It will always have the option to
15:16:19 <joanie> just like it does with Gecko
15:16:21 <clown> yes, makes sense.
15:16:29 <joanie> but it will also have its own
15:16:37 <joanie> which by default, once it's working, will be enabled
15:16:41 <joanie> so as not to confuse users
15:16:53 <clown> okay, I think I got it.  Thanks!
15:17:00 <joanie> thank you for your interest :)
15:17:04 <infapi00> ok, joanie, thanks for the explanation
15:17:14 <infapi00> so more questions, comments, stuff on this topic?
15:17:35 * joanie looks to mgorse and says, "value interface for status?"
15:18:15 * joanie has mixed feelings about this
15:18:26 <joanie> JOANIE_STATE_INTEDERMINATE
15:18:35 <clown> lol
15:18:41 <mgorse> #info mgorse proposed a revised patch for the status column in evolution, that exposes the value of the underlying enum via AtkValue, although it is a bit of a hack and might not be the best way to do it. An alternate hack, that Joanie proposed, would be to have the "read" icon have no name, so that orca doesn't read anything
15:19:07 <mgorse> or maybe I proposed it originally
15:19:08 <joanie> (for the record, I realizing my idea is also a hack)
15:19:24 <joanie> s/realizing/realize/
15:19:38 * joanie needs spellcheck in irssi
15:19:56 <infapi00> so, who will decide which hack to use?
15:20:00 <infapi00> evolution reviewers?
15:20:04 <infapi00> mgorse vs joanie?
15:20:14 <joanie> infapi00?
15:20:17 <joanie> all of us?
15:20:35 <mgorse> So far, the evolution maintainers have been silent, although that could be because they're figuring that we're still trying to decide what to do
15:20:39 <infapi00> hmm, well, I lack most of the context, and probably here is not the best place
15:20:41 <infapi00> so
15:20:58 <joanie> mgorse: I think they reviewed your patch recently
15:20:59 <infapi00> #action infapi00 will take a look to the evolution bug, see both solutions, and at least, give an opinion
15:21:10 <infapi00> joanie is that good enough?
15:21:19 <mgorse> bug 743404
15:21:30 <Services> 04Bug http://bugzilla.gnome.org/show_bug.cgi?id=743404 normal, Normal, ---, evolution-shell-maintainers, UNCONFIRMED , Status column of message list exposed poorly via atk
15:21:44 <joanie> infapi00: sure(ish). The bigger convo is what all does AtkValue apply to? Currently it's been for widgets with a range-y nature
15:22:03 <joanie> this bug, and some W3C stuff, raises the question of whether or not that's what we want/need
15:22:15 <joanie> if the answer is, yes, broaden it. Ok. I'll adjust.
15:22:16 <clown> perhaps a dumb question, but what happens with Thunderbird?  Does it have a status column?
15:22:21 <mgorse> joanie: oh, yeah, there's a comment from Milan, as of today
15:22:38 <joanie> if the answer is no, don't broaden it, the problem is we have no good API for it and should create some.
15:23:05 <joanie> clown: Yeah. And there are bugs I filed related to it
15:23:09 <joanie> I don't recall what they did
15:23:26 <clown> ok, thanks.
15:23:32 <joanie> i think perhaps it's to not show an icon for read
15:23:46 <joanie> anyhoo....
15:23:58 <infapi00> joanie, so there is a possibility that we need to review AtkValue API again?
15:24:14 <joanie> infapi00: It's not the AtkValue API
15:24:22 <joanie> it's a question of what should implement that API or not
15:24:29 <infapi00> ah ok
15:24:31 <infapi00> well, yes
15:24:32 <joanie> Orca is expecting it's on ranges
15:24:39 <joanie> are docs suggest it's on ranges
15:24:40 <infapi00> that is one of the problems of ATK
15:24:42 <joanie> ugh
15:24:44 <joanie> s/are/our/
15:24:46 <infapi00> there are several interfaces
15:24:47 <joanie> migraine....
15:25:09 <infapi00> and each developer would need to use his common sense to know when they apply
15:25:19 <joanie> anyhoooooo.... We are seeing use cases (and developers) saying, "Well, we should use the Value interface for <insert non-range widget here>"
15:25:27 <joanie> I'm saying we need to decide which it is
15:25:33 <joanie> and then act accordingly
15:25:40 <infapi00> and again, our team unicorn: it would be good to have a developer guide to answer that kind of stuff
15:25:48 <infapi00> but ok
15:25:53 <joanie> we could add it to our docs
15:26:04 <infapi00> as mentioned, I will take a look to that bug
15:26:04 <joanie> but we need to decide first
15:26:04 <infapi00> and at least, give an opinion
15:26:05 <joanie> as it's bigger than just Evo.
15:26:09 <infapi00> ok
15:26:16 * joanie coughs "W3C" again
15:26:19 <mgorse> Ideally, we might have a "give me a textual description of what is displayed by this widget, in a way that isn't translated and can thus be keyed by an AT" function, and I don't think we really do
15:26:28 <joanie> clown: this is... Was it Jon's idea?
15:26:30 * clown gives joanie a cough drop
15:27:18 <clown> waa what Jon's idea?
15:27:20 <joanie> mgorse: what you describe sounds like a state
15:27:25 <clown> and, Jon Gunderson?
15:27:36 <joanie> clown: I don't recall the exact discussion. But it was some meeting (ARIA I think).
15:27:54 <joanie> And the consensus was he (if it was Jon, then yes, Gunderson) would write a proposal
15:28:13 <joanie> but it was about using valuetext for something similar to what mgorse is describing
15:28:20 <joanie> and it was not a range
15:28:21 <clown> ah.  I could check the mailing lists to see if he has.  Or the tracker to see what his actions are.
15:28:22 <joanie> ohhhhh
15:28:27 <joanie> switch
15:28:31 <joanie> it was switch
15:28:35 <joanie> not just on/off
15:28:41 <joanie> male/female
15:28:43 <joanie> etc.
15:28:48 <clown> right.  There was a suggestions that role="switch" has a value (on or off)
15:28:58 <joanie> but a switch ain't a range
15:29:04 <joanie> any more than email status is a range
15:29:06 <clown> and a worry as to where to put that value in the AAPI.
15:29:23 <joanie> it's a different flavor of the evo issue
15:29:28 <mgorse> Yeah, it sounds similar to these "toggles" that evolution uses for mail status
15:29:31 <joanie> what does the Value interface apply to?
15:29:45 <joanie> mgorse: it's the same freaking problem
15:29:59 <joanie> so we must decide if AtkValue applies to non-ranges or not
15:30:07 <joanie> if so: fine, we adjust (orca, documentation, etc.)
15:30:15 <mgorse> As far as evo goes, I'm wondering if it makes sense to just not have a name for "read" for now, and revisit it later possibly, since I'm not sure we'll end up adding new API for 3.16
15:30:16 <joanie> if not: fine, we create new API
15:30:22 <clown> aria-valuetext isn't that bad an idea, but I don't recall what it maps to in AAPIs.
15:30:45 <joanie> in ATK/AT-SPI2 right now, it's an object attribute
15:30:57 <joanie> ATK has new API for it
15:31:01 <joanie> AT-SPI2 doesn't yet
15:31:08 <joanie> and I took an action to create it
15:31:16 <clown> and, ATK's new API puts it, where?
15:31:18 <joanie> but still, it assumes a range widget
15:31:22 * joanie pulls up new API
15:31:23 <clown> "it" = aria-valuetext.
15:31:34 <joanie> though infapi00 is going to hit us for the deep dive soon ;)
15:31:44 <clown> yes, we should discuss elsewhere....
15:31:46 <infapi00> yeah, I was thinking on that
15:31:58 <joanie> please don't hit me, fearless leader!
15:32:00 <joanie> :P
15:32:02 <infapi00> in fact, due that I will change the agenda on the flight just in case
15:32:08 <infapi00> s/flight/fly
15:32:16 <joanie> clown: https://developer.gnome.org/atk/unstable/AtkValue.html#atk-value-get-value-and-text
15:32:23 <infapi00> so, as this topic diverted a little to w3c discussion
15:32:28 <infapi00> lets close it with
15:32:28 <joanie> :)
15:32:52 <infapi00> #info Evolution bug raised some concerns about AtkValue and when/how apply it, that will be discussed on the following days
15:33:03 <infapi00> so, ok, to move to next topic?
15:33:07 <clown> sure
15:33:15 <infapi00> #topic Bugzilla products
15:33:23 <infapi00> so that is the modified agenda
15:33:56 <infapi00> #info Recently bugzilla got updated, and somehow it led to bugzilla maintainers to realize an (somewhat) old bug about how a11y components are distributed, products-sake
15:33:57 * clown isn't sure what the topic means.
15:34:08 <infapi00> #info bug 740075
15:34:19 <Services> 04Bug http://bugzilla.gnome.org/show_bug.cgi?id=740075 normal, Normal, ---, bugzilla-maint, UNCONFIRMED , java-atk-wrapper should be filed under atk
15:34:19 * clown waits for explanation
15:34:30 <infapi00> #info In short, the problem is that right now on bugzilla there is a product "at-spi"
15:34:59 <infapi00> #info and bugs for pyatspi2, at-spi2-core, at-spi2-atk and java-atk-wrapper are included there, making everything somewhat messy
15:35:11 <infapi00> #info it was asked to split on products
15:35:36 <infapi00> #info recently andre kappler mentioned that a good general rule is that  git repository == bugzilla product
15:36:03 <infapi00> #info just after magpie and infapi00 agreed that it was a good approach (and that in fact, it was exactly what asked)
15:36:13 * clown wonders if bugzilla does "sub-topics".
15:36:40 <infapi00> #info akappler mentioned that there is a product on freedesktop bugzilla for at-spi2-core/at-spi2-atk, and suggested that it would be a good idea to use just one
15:36:56 <infapi00> #info infapi00 and joanie mentioned that it would be better to use gnome one (as is the one active) and close the other
15:37:03 <infapi00> #info magpie seems to suggest the same, but not sure
15:37:15 <infapi00> so, as we are here, and in order to have the opinion of the maintainer
15:37:20 <infapi00> mgorse, what it is your opinion?
15:37:41 <mgorse> I agree about closing the fdo bug. Is there a thread somewhere where I should be commenting?
15:38:24 <mgorse> As far as splitting the gnome bugzilla, that would seem to make sense, although I'm worried about bugs being filed for "at-spi" and being sent to the unmaintained CORBA AT-SPI
15:39:09 <clown> is CORBA AT-SPI still supported?
15:39:10 <infapi00> well, but I guess that that
15:39:19 <infapi00> is a problem inherent to new versions
15:39:26 <infapi00> I mean, I really think that a lot of gtk3 bugs
15:39:36 <mgorse> clown: no, it went away with GNOME 3
15:39:43 <infapi00> are wrongly ticketed as gtk2 bugs, and viceversa
15:39:55 <mgorse> infapi00: makes sense
15:39:55 <joanie> mgorse: in the GNOME bug (740075) there is discussion
15:40:03 <mgorse> joanie: ok, thanks
15:40:06 <joanie> including a link to a new bug I filed at fdo
15:40:10 <infapi00> so if someone opens a bug for at-spi, and it is in fact an at-spi2-core bug, is just about move the bug
15:40:16 <joanie> requesting they close that at-spi2 product
15:40:25 <joanie> https://bugs.freedesktop.org/show_bug.cgi?id=89106
15:40:31 <Services> 04Bug 89106: normal, medium, ---, sitewranglers, NEW , Remove at-spi2
15:40:35 <infapi00> mgorse, so if you don't mind, it would be good if you add a comment ont hose bugs with your maintainer hat on
15:40:45 <mgorse> ok
15:41:19 <magpie> it's not clear what modules andre is talking about
15:41:34 <infapi00> magpie, as far as I understand
15:41:46 <infapi00> he just want to avoid having two bugzillas
15:41:49 <infapi00> tracking the same product
15:42:01 <infapi00> so I think that as far as we get the one not used removed
15:42:06 <infapi00> he will not put any problem
15:43:02 <magpie> ah right. That seems to make sense. Except that the wrapper is also tracked on launchpad
15:43:07 <magpie> where does it end?
15:43:25 <infapi00> hmm, well, but launchpad is for ubuntu canonical
15:43:32 <infapi00> I mean that it is their downstream bugzilla
15:43:56 <infapi00> fedora also has a downstream bugzila
15:44:05 <magpie> like what does this problem have to do with the bug which is to make these modules products instead of components?
15:44:11 <infapi00> I were pinged some times on those bugzillas
15:44:27 <infapi00> the issue here is that both gnome and freedesktop are upstream bugzillas
15:44:45 <infapi00> magpie, no, I don't think that andre is talking about that
15:45:04 <joanie> magpie: see my comment on the bug :)
15:45:08 <infapi00> again, his concern is about having a product on fdo bugzilla, and the same product on gnome   one, that I agree that can be confusing
15:45:10 <joanie> it shouldn't have ANYTHING to do with that
15:45:16 <infapi00> or read what joanie already said
15:45:19 <magpie> he just prefers to move all the accessibility module bug tracking to free desktop?
15:45:27 <joanie> but because we're talking about it, now it magically becomes related
15:45:41 <joanie> magpie: I don't *think* he wants that
15:46:00 <infapi00> magpie, he didn't suggested anything at all
15:46:08 <infapi00> he just pointed that there is another bugzilla
15:46:09 <joanie> but he's been reminded of this other tracker because of our request locally
15:46:31 <joanie> and I think that we *may* be blocked on our request, until we resolve the fdo tracker issue
15:46:42 <joanie> hence I filed a bug to remove the at-spi2 product from fdo
15:46:49 <magpie> ah ok
15:46:57 <joanie> if they do that, Andre shall have nothing to point out or complain about ;)
15:47:25 <magpie> so once this is removed from free desktop we can get back onto the bug itself?
15:47:37 <joanie> magpie: that is my belief and hope :)
15:47:58 <joanie> we can say "see, look here, problem solved!" "now can we have our request fulfilled?"
15:48:01 <joanie> :)
15:48:04 <mgorse> I hope someone from freedesktop responds soon. That was part of the reason that the git repo was moved to gnome in the first place
15:48:20 <magpie> ok makes sense, thanks for filing the bug.
15:48:41 <infapi00> ok, so as everybody here agrees
15:48:44 <infapi00> ok to move?
15:48:49 * joanie nods
15:48:53 <magpie> sure
15:48:55 <infapi00> #topic W�C updates
15:48:58 <infapi00> #topic W3C updates
15:49:04 <infapi00> clown, ?
15:49:13 <clown> I prepared my info's in advance.
15:49:23 <clown> I will copy and paste…
15:49:32 <clown> #info Last week's call dealt with @aria-describedat
15:49:32 <clown> #info Representatives from the e-publishing vendors were present.
15:49:32 <clown> #info They want aria-describedat as a mechanism for having additional information about a subject, but not having that information interfere with the current page layout.
15:49:33 <clown> #info aria-describedat is an URL to reference the external information.
15:49:33 <clown> #info Main objection from browser vendors is that aria-described should not alter main user interface.
15:49:34 <clown> #info James Craig commented that this appears to be "a solution in search of a problem", and pointed to other ways to accomplish the same thing:  http://cookiecrook.com/longdesc/
15:49:37 <clown> #info Although not resolved, it was generally felt the attribute would be useful, but it shouldn't affect primary UI.
15:49:40 <clown> #info The discussion continues; there is still an issue about "discoverability".
15:49:42 <clown> #info Minutes of the meeting: http://www.w3.org/2015/02/05-aria-minutes.html#item01
15:49:44 <clown> #info aria-describedat in the spec:  http://rawgit.com/w3c/aria/master/aria/aria.html#aria-describedat
15:49:48 <clown> done
15:49:50 * clown waits while that is all read.
15:51:32 <infapi00> as James Craig mentioned, Im having a kind of longdesc deja-vu
15:51:40 <joanie> we all are
15:51:42 <joanie> ;)
15:52:21 <clown> yes, but, (1) aria-describedat is only for AAPIs, and (2) it applies to more that images (longdesc is for images only).
15:52:25 <clown> But, point taken.
15:52:46 <infapi00> so if that is included too
15:52:57 <infapi00> I foresee an "aria-describe-everything-at"
15:53:01 <infapi00> that deprecate both
15:53:15 <infapi00> to avoid having two with the same meaning
15:53:19 <infapi00> or too similar meaning
15:53:24 <infapi00> but probably Im just being mean
15:53:39 <clown> there are some who suggest there be a general "describedat" attribute (non aria) for use by everyone, not just AT users.
15:53:54 <joanie> clown: BUT, the discussion from the last meeting including references from the publishers about it not being just for accessibility
15:54:01 <clown> so, not completely mean, infapi00
15:54:26 <joanie> so I myself question if it belongs in ARIA
15:54:26 <infapi00> so the council defining aria suggest to include that as a non-aria attribute?
15:54:38 <infapi00> is just me, or sound like hiding the dust under the carpet?
15:54:40 <clown> yes, joanie, that is the question.
15:54:56 <clown> or a question.
15:54:56 * joanie nods
15:55:02 <clown> In fact Alex, brought it up in a related way.
15:55:07 * clown looks for quote.
15:55:08 <joanie> infapi00: we haven't concluded anything yet
15:55:33 <infapi00> ok
15:55:45 <clown> "It should not be aria-describedat if it has UI; should just be describedat."
15:56:02 <clown> That quote was from alex
15:56:07 <clown> from the meeting.
15:56:46 <clown> acttually, the discussion led to a "check" of the aria documents regarding advice about UI.
15:57:11 <joanie> My question is: Are we considering aria-describedat because (assuming we get consensus) we can shove it into ARIA more easily than we could, say, HTML or some other spec?
15:57:28 <clown> #info As a result of the discussion around aria-describedat, the ARIA documents are being searched to remove any suggetions regarding ARIA affecting main stream UI.
15:57:45 <clown> joanie, yes, that is a factor.
15:58:05 <joanie> In which case, it's the spec equivalent of a hack (imho)
15:58:07 <clown> also ARIA is language neutral.
15:58:14 <joanie> true
15:58:25 <clown> what language does describedat belong to?
15:58:34 <joanie> all languages!
15:58:34 <clown> HTML?  SVG?  DPBB?
15:58:41 <clown> *DPUB
15:58:51 <joanie> I know. It's a valid issue, but that issue doesn't make it ARIA
15:58:59 <clown> how do you do that in a non-xml world, joanie?
15:59:07 <joanie> do what?
15:59:11 * infapi00 notes that the sands of time are running low
15:59:30 <joanie> and I have another meeting to run to
15:59:32 * clown shuts up.
15:59:43 <joanie> but all valid issues
15:59:44 <clown> move along, if you want.  I'm done
15:59:58 <joanie> it wasn't a shut up thang
16:00:05 <joanie> I hope I didn't sound like that
16:00:08 <infapi00> well, do you want to summarize this with an info
16:00:16 <infapi00> or it is implicit to your previous infoes?
16:00:22 <infapi00> clown, ^
16:00:47 <clown> I'll try to info a  couple of things.
16:01:11 <clown> #info part of the attraction is that aria-describedat is language neutral and can be used in any language.
16:01:40 <clown> #info wheras, describedat would have to be introduced to each language (html, svg, etc) separately.
16:02:44 <clown> #info which was a problem that XML solved — mix and match language features — but it has fallen out of favour.
16:02:45 <clown> done.
16:03:24 <infapi00> ok, so, I think that unless jjmarin wants to say something about marketing
16:03:28 <infapi00> I will close the meeting
16:03:30 <infapi00> jjmarin, ?
16:03:38 <jjmarin> hi
16:03:53 <jjmarin> #info There is a section in the wiki for adding info for the release notes https://wiki.gnome.org/ThreePointFifteen/ReleaseNote
16:04:02 * jjmarin wonders if we can add anything of the work done/in progress in this page
16:04:13 <jjmarin> For example, mclassen has said in DevConf.cz that an High Constrast theme will be included in GTK 3.16 https://www.youtube.com/watch?v=Rdbw8NAnPFI (min. 7:08), and I think this can go in the release notes (I only know about this from this talk)
16:04:24 <jjmarin> If we have several lines, I think we can get together in a a11y section
16:05:20 <jjmarin> And about the annual report, because I need to figure out some info
16:05:29 <jjmarin> #action Juanjo will add some notes with items he thinks he can mention in the article for the annual report https://wiki.gnome.org/Engagement/AnnualReport/2014 and he will ask for opinions, corrections, love, rage and/or addtions
16:05:38 <jjmarin> done !
16:05:42 <infapi00> jjmarin, ok thanks
16:06:01 <infapi00> and sorry but the lack of feedback (at least lack of my feedback) on last year report
16:06:19 <jjmarin> np
16:06:52 <infapi00> so, having said so
16:06:56 <infapi00> as it is already +6
16:07:01 <infapi00> Im going to close the meeting
16:07:03 <infapi00> #endmeeting