/irc-logs / mozilla / #accessibility / 2013-04-11 / end
Options:
- # Session Start: Thu Apr 11 00:00:01 2013
- # Session Ident: #accessibility
- # [00:11] * Quits: nhirata (anonymous@moz-BE33DA21.fw1.sfo1.mozilla.net) (Quit: nhirata)
- # [00:15] * Joins: nhirata (anonymous@moz-BE33DA21.fw1.sfo1.mozilla.net)
- # [00:26] * Joins: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP)
- # [00:27] * Quits: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP) (Ping timeout)
- # [00:56] * Quits: Gijs (gijs@2557E599.66715431.D25A875A.IP) (Ping timeout)
- # [01:01] * Joins: Gijs (gijs@2557E599.66715431.D25A875A.IP)
- # [01:01] * Joins: Jamie (Instantbir@moz-CA26021.jantrid.net)
- # [01:26] * Joins: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP)
- # [01:28] * Quits: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP) (Ping timeout)
- # [02:27] * Joins: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP)
- # [02:28] * Quits: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP) (Ping timeout)
- # [02:36] * Quits: logbot (logbot@moz-58CB32ED.glob.com.au) (Ping timeout)
- # [02:37] * Joins: logbot (logbot@moz-58CB32ED.glob.com.au)
- # [02:42] * Joins: scott_gonzalez (scott_gonz@moz-91C81A39.hrbgpa.fios.verizon.net)
- # [03:00] * Joins: surkov (surkov@B3C081AC.7AE20955.EBE09E3C.IP)
- # [03:00] * ChanServ sets mode: +o surkov
- # [03:22] * Quits: @surkov (surkov@B3C081AC.7AE20955.EBE09E3C.IP) (Quit: surkov)
- # [03:24] * Quits: nhirata (anonymous@moz-BE33DA21.fw1.sfo1.mozilla.net) (Quit: nhirata)
- # [03:27] * Joins: surkov (surkov@B3C081AC.7AE20955.EBE09E3C.IP)
- # [03:27] * ChanServ sets mode: +o surkov
- # [03:27] * Joins: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP)
- # [03:27] * Quits: @surkov (surkov@B3C081AC.7AE20955.EBE09E3C.IP) (Quit: surkov)
- # [03:28] <@firebot> akeybl@mozilla.com granted approval-mozilla-aurora for attachment 733213 on bug 857936.
- # [03:28] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=857936 nor, --, mozilla23, surkov.alexander, RESO FIXED, ARIA columnheader/rowheader should inherit editable state
- # [03:29] * Joins: surkov (surkov@B3C081AC.7AE20955.EBE09E3C.IP)
- # [03:29] * ChanServ sets mode: +o surkov
- # [03:30] * Quits: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP) (Ping timeout)
- # [03:32] * Quits: logbot (logbot@moz-58CB32ED.glob.com.au) (Ping timeout)
- # [03:33] * Joins: logbot (logbot@moz-58CB32ED.glob.com.au)
- # [03:36] * Quits: Gijs (gijs@2557E599.66715431.D25A875A.IP) (Ping timeout)
- # [03:41] * Joins: yzen (Adium@AA2C085E.10C8AEF8.9A8C35B4.IP)
- # [03:51] * Quits: scott_gonzalez (scott_gonz@moz-91C81A39.hrbgpa.fios.verizon.net) (Quit: scott_gonzalez)
- # [03:52] * Joins: davidb (davidb@moz-22F8547E.dsl.bell.ca)
- # [03:52] * ChanServ sets mode: +qo davidb davidb
- # [04:07] * Quits: yzen (Adium@AA2C085E.10C8AEF8.9A8C35B4.IP) (Quit: Leaving.)
- # [04:13] * Joins: yzen (Adium@AA2C085E.10C8AEF8.9A8C35B4.IP)
- # [04:28] * Joins: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP)
- # [04:30] * Quits: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP) (Ping timeout)
- # [04:39] * Quits: @davidb (davidb@moz-22F8547E.dsl.bell.ca) (Quit: davidb)
- # [04:58] * Joins: lizzard (ehenry@moz-5A4BE8E8.dsl.static.sonic.net)
- # [05:09] * khuey|away is now known as khuey
- # [05:13] * Quits: fxa90id (fxa90id@moz-DA50A290.dsl.dynamic.t-mobile.pl) (Ping timeout)
- # [05:14] * Quits: lizzard (ehenry@moz-5A4BE8E8.dsl.static.sonic.net) (Quit: lizzard)
- # [05:28] * surkov sets mode: +o tbsaunde
- # [05:29] * Joins: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP)
- # [05:31] * Quits: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP) (Ping timeout)
- # [05:49] * Joins: Gijs (gijs@2557E599.66715431.D25A875A.IP)
- # [06:19] * Joins: lizzard (ehenry@moz-5A4BE8E8.dsl.static.sonic.net)
- # [06:28] <@surkov> tbsaunde: ping me when you're around
- # [06:28] <@tbsaunde> surkov: hi
- # [06:28] <@surkov> oh!
- # [06:29] <@surkov> tbsaunde: I wanted to ask you to compare our text and gedit text on your example
- # [06:29] <@surkov> tbsaunde: I think whether I should move forward or I made mistake somewhere when I was reading the atk spec
- # [06:30] * Joins: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP)
- # [06:31] * Quits: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP) (Ping timeout)
- # [06:33] <@tbsaunde> surkov: I asked hub to look into it since I'm kind of busy
- # [06:33] <@tbsaunde> hub: how did that go?
- # [06:35] <@surkov> hub sleeps I guess
- # [06:36] <@tbsaunde> surkov: apparently, sorry
- # [06:38] * Joins: nhirata (anonymous@moz-1A50F7F8.hsd1.ca.comcast.net)
- # [06:45] * Quits: brankc (brankc@moz-29FFE1B5.me) (Ping timeout)
- # [06:45] * Joins: brankc (brankc@moz-29FFE1B5.me)
- # [06:46] * Quits: Gijs (gijs@2557E599.66715431.D25A875A.IP) (Ping timeout)
- # [06:49] * Quits: nhirata (anonymous@moz-1A50F7F8.hsd1.ca.comcast.net) (Quit: nhirata)
- # [06:51] <@firebot> yura.zenevich@gmail.com requested feedback from eitan@monotonous. org for attachment 736122 on bug 811307.
- # [06:51] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=811307 nor, --, ---, yura.zenevich, NEW, [AccessFu] Add mochitest for enabling
- # [06:53] <@firebot> yura.zenevich@gmail.com requested feedback from eitan@monotonous. org for attachment 736123 on bug 811307.
- # [06:58] * khuey is now known as khuey|away
- # [07:07] * Joins: icaaq (Adium@moz-D8A76420.cust.bredbandsbolaget.se)
- # [07:18] <icaaq> lizzard: regarding marcoz geekmeet, we will stream the event here http://geekmeet.creuna.se/
- # [07:24] <lizzard> icaaq: awesome, thanks
- # [07:25] <icaaq> np :)
- # [07:30] * Quits: lizzard (ehenry@moz-5A4BE8E8.dsl.static.sonic.net) (Quit: lizzard)
- # [07:30] * Joins: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP)
- # [07:32] * Quits: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP) (Ping timeout)
- # [07:35] * Quits: @hub (hub@E639FEDB.F04051C4.DC1243F8.IP) (Ping timeout)
- # [07:39] * Joins: hub (hub@E639FEDB.F04051C4.DC1243F8.IP)
- # [07:39] * ChanServ sets mode: +o hub
- # [07:42] * Quits: peteb-away (ptbrunet@moz-B51E1692.austin.res.rr.com) (Ping timeout)
- # [08:11] <@firebot> trev.saunders@gmail.com requested review from surkov.alexander@gm ail.com for attachment 736146 on bug 809338.
- # [08:11] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=809338 nor, --, ---, trev.saunders, NEW, make HTML select hierarchical
- # [08:12] * Quits: yzen (Adium@AA2C085E.10C8AEF8.9A8C35B4.IP) (Ping timeout)
- # [08:12] * Joins: yzen (Adium@AA2C085E.10C8AEF8.9A8C35B4.IP)
- # [08:14] <@firebot> trev.saunders@gmail.com cancelled review?(surkov.alexander@gm ail.com) for attachment 723778 on bug 809338.
- # [08:16] * Quits: yzen (Adium@AA2C085E.10C8AEF8.9A8C35B4.IP) (Quit: Leaving.)
- # [08:20] <Jamie> surkov: hi. was just reading your blog post re accessible text
- # [08:20] <Jamie> surkov: this line concerns me: "Do not use embedded characters for inline elements."
- # [08:21] <Jamie> surkov: If i understand that correctly, it will *definitely* breka NVDA
- # [08:21] <@surkov> Jamie: oh
- # [08:21] <@surkov> Jamie: regardless of that do you think it's a good idea?
- # [08:21] <@surkov> to not use embedded chars for inlines
- # [08:22] <Jamie> surkov: that debate has been going on for years :) My belief has always been that if it's going to be done, we should go all the way and make the entire document one big flat text string
- # [08:22] <Jamie> I don't see the point in going half way
- # [08:22] <Jamie> I think I've ranteda bout this on a bug somewhere
- # [08:23] <@surkov> Jamie: inlines are different, for example, <p>hel<a>l</a>o</p> is one word
- # [08:23] <@surkov> but you can't detect it if inlines are embedded chars
- # [08:23] <Jamie> That depends. *Why* are the different?
- # [08:23] <Jamie> It might be one word, but <p>foo</p><p>bar</p> is one document
- # [08:23] <@surkov> Jamie: like example above
- # [08:23] <Jamie> so why is word so special?
- # [08:23] <Jamie> why line? why paragraph?
- # [08:23] <@surkov> it's two words in case of paragraphs
- # [08:24] <Jamie> right, but my point is that you're saying word is special somehow
- # [08:24] <@surkov> having a one text document might be a big perf issue
- # [08:24] <@surkov> inlines are different from blocks
- # [08:24] <Jamie> if you select two paragraphs, you have the same problem you always have with words
- # [08:24] <@surkov> since sibling inlines can create one word
- # [08:24] <Jamie> yes, but in terms of text, they can all be selected continuously
- # [08:25] <Jamie> So again, why is "word" special?
- # [08:25] <Jamie> I'm partly playing devil's advocate here, but that needs ot be answered before there's any real benefit
- # [08:25] <@surkov> in case of sibling paragraphs I guess you can ask them separately for selection
- # [08:25] <@surkov> in case of sibling inlines they can be a single word and there's no way to get it
- # [08:26] <Jamie> you could, but essentially, you have ot do the same fiddly evil stuff you've always had to do; it's just different units
- # [08:26] <Jamie> sure there is. it's just very difficult
- # [08:26] <Jamie> I'm not arguing that walking between objects is painful
- # [08:26] <Jamie> isn't
- # [08:26] <Jamie> it really is very painful
- # [08:27] <@surkov> Jamie: it means screen reader should decide where the word start and where it ends
- # [08:27] <Jamie> my point is that if you don't go the whole way, at some point, you have no choice but to do it
- # [08:27] <Jamie> that's not true
- # [08:27] <@surkov> why?
- # [08:27] <@surkov> how you can traverse by words in case of <p>hel<a>l</a>o</p>?
- # [08:27] <Jamie> the algorithm is hugely complicated, but basically, you ask the bottom object for its word offsets, go up, ask the embedding object for its word offsets, and walk around until you find the start and end
- # [08:28] <Jamie> it is utterly hideous
- # [08:28] <@tbsaunde> Jamie: you need to do it sometimes, but not always, for example if we say word is always within one accessible no embeded objects, then logic to read words can ignore embeded objects
- # [08:28] <@surkov> not sure I get this alg
- # [08:28] <Jamie> tbsaunde: that suggests that words are special :)
- # [08:29] <@surkov> oh, difference in offsets
- # [08:29] <Jamie> tbsaunde: in abstract code, words are the same as any other unit
- # [08:29] <Jamie> surkov: so in <p>a<a>b</a>c</p>
- # [08:29] <Jamie> and you're on b
- # [08:29] <Jamie> you ask the link for its word offsets for b. then you go up and ask the paragraph for its word offsets at that point
- # [08:30] <Jamie> so there's a lot of hopping in and out of objects
- # [08:30] <@tbsaunde> Jamie: word offsets? what method do you mean?
- # [08:30] <@surkov> getTextAtOffset i think
- # [08:31] <@tbsaunde> Jamie: sure, but practically I suspect words are a common case so making them fast is probably good
- # [08:31] * Joins: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP)
- # [08:31] <Jamie> IAccessible::textAtOffset
- # [08:32] <Jamie> tbsaunde: if it's entirely performance, thats probably a valid argument. but if you think it's going to make life simpler for AT, it isn't, because they still have to handle the complex object hopping for anything other than words
- # [08:32] <Jamie> I'll grant performance is important
- # [08:32] <Jamie> I just want ot make sure the reasons are clear
- # [08:32] * Quits: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP) (Ping timeout)
- # [08:32] <Jamie> This is *not* simpler or tidier for all cases, only for words, and therefore, it is purely performance
- # [08:33] <Jamie> I also suspect we'll have ot rewrite a large chunk of NVDA's vbuf backend, not sure about other ATs
- # [08:34] <Jamie> Right now, we walk into an embedded object and we can recurse to render it
- # [08:34] <Jamie> with this, we can't; the text comes from the parent node, but we need the object to get semantic info; e.g. perhaps aria-label overrides the text
- # [08:34] <@surkov> Jamie: so the problem is you don't support embedded objects having normal text (rather than embedded char)?
- # [08:35] <Jamie> surkov: not when the text is on the parent
- # [08:35] <Jamie> btw, what will IAccessibleText on the inline object return?
- # [08:35] <@surkov> I think the text
- # [08:35] <@surkov> of inline object
- # [08:36] <Jamie> in that case, I guess we can just ignore the text on the parent
- # [08:36] <@surkov> if it makes sense to implement IAccessibleText on it at all
- # [08:36] <Jamie> that would certainly be an easy change
- # [08:36] <@tbsaunde> Jamie: I was expecting we'd only do it for things where we don't need a child accessible, but I'm not the one writing it ;)
- # [08:37] <Jamie> btw, <p>Hel<a aria-label="wow">l</a>o</p> should be fun for ATs to handle
- # [08:37] <Jamie> tbsaunde: define "thigns where we don't need a child accessible"
- # [08:37] <Jamie> tbsaunde: you need a child accessible for everything if you don't want to lose data
- # [08:37] <Jamie> tbsaunde: if you don't, firefox already strips it as is
- # [08:37] <Jamie> <p>blah<span>blah</span></p>
- # [08:37] <Jamie> the span gets stripped
- # [08:38] <Jamie> surkov: certainly, if it's just performance, I do agree it'll be faster; it'll make our rich text code much faster once it's done
- # [08:38] <@tbsaunde> Jamie: <p>a<i>b</i></p> we can expose text attributes for only part of the text, and afaik one accessible should be fine
- # [08:39] <Jamie> tbsaunde: Firefox already does that
- # [08:40] <Jamie> surkov: just making sure you're aware that it isn't a "this makes life 500 times easier for ATs" kind of change :)
- # [08:40] <@tbsaunde> Jamie: so, its not terribly clear how the aria-label thing should work
- # [08:40] <Jamie> tbsaunde: yeah, tell me about it
- # [08:40] <@surkov> Jamie: yeah for old like you probably not :)
- # [08:41] <Jamie> surkov: mm, even me aside, it won't; as I said, you still have to do that evil stuff where multiple paragraphs are selected. it's just going higher
- # [08:41] <@surkov> anyway the logic should be simpler in case of words
- # [08:41] <@tbsaunde> Jamie: like ok, aria-label effects name, but if it effects text vallue as well then how do you get text attributes for it, and if not then why is it fun for ATs?
- # [08:41] <Jamie> surkov: I guess i sort of follow your argument re inline being different; paragraph offsets are always confined to one block element, line offsets must always be within one block
- # [08:42] <Jamie> surkov: oh, except inline-block. that gets *really* fun
- # [08:42] <@surkov> inline-block? I don't recall how these look like
- # [08:43] <Jamie> tbsaunde: because people have this belief that aria-label should always override content
- # [08:43] <Jamie> tbsaunde: so in that case, we see a word, then we descend and find there's an aria-label
- # [08:44] <Jamie> tbsaunde: in other words, you always have ot descend anyway to check whether something might have overridden that area of text
- # [08:44] <@surkov> that'd be crazy
- # [08:45] <@surkov> I agree you should announce that aria-label on tabbing
- # [08:45] <@surkov> but using it instead text seems crazy
- # [08:45] <Jamie> surkov: I used ot agree with you, but it is expected now
- # [08:46] <@tbsaunde> Jamie: so, happens if I do <p>text <b>some <span aria-label="foo">more</span> text</b></p> and I ask for text attributes of the stuff in the span?
- # [08:46] <@surkov> well, I nothing understand in screen readers :)
- # [08:46] <@surkov> sometimes
- # [08:46] <Jamie> surkov: it's not us, trust me
- # [08:46] <@surkov> I believe
- # [08:46] <@tbsaunde> Jamie: I doubt orca does that, but would need to check fwiw
- # [08:46] <Jamie> surkov: I'd argue it is incorrect use of aria-label
- # [08:46] <Jamie> but apparently, the spec disagrees
- # [08:46] <@surkov> which spec?
- # [08:46] <Jamie> tbsaunde: we provide no attributes; we can't
- # [08:47] <@surkov> UAIG?
- # [08:47] <@tbsaunde> Jamie: specs usually have bugs :)
- # [08:47] <Jamie> well, ARIA suggests aria-label should be reported by ATs
- # [08:47] <Jamie> consider something like a div with a background image
- # [08:47] <Jamie> authors use aria-label to give it a label
- # [08:48] <Jamie> so we get cool stuff like <a aria-label="Widgets"><br></a>
- # [08:48] <@surkov> I see
- # [08:48] <Jamie> but it's a web document, so reading it on tab is not enough; you also need to read it while doing caret navigation.
- # [08:49] <@surkov> I'm curious how AT can maintain caret navigation inside aria-label
- # [08:49] <Jamie> they can't. that's why windows ATs use virtual buffers. despite all the flack we get for it, it provides a superior caret navigation experience
- # [08:50] <Jamie> well, that's one of the reasons :)
- # [08:50] <@surkov> it's something new for me I think
- # [08:51] <Jamie> surkov: but yeah, on the earlier topic, I guess doing this does mean we cna guarantee that a unit won't ever cross two objects
- # [08:51] * Quits: logbot (logbot@moz-58CB32ED.glob.com.au) (Ping timeout)
- # [08:51] <@tbsaunde> Jamie: I might be ok supporting authors providing text somehow, but not with aria-label like that
- # [08:51] <@surkov> Jamie: what does it mean sorry?
- # [08:51] <Jamie> surkov: paragraphs are always one block, lines are always inside one object, words are always inside one object
- # [08:52] <Jamie> I guess that does simplify the code a bit; it's one guarantee we never had
- # [08:52] <Jamie> it'd certainly cause me less headaches in the rich text editing stuff
- # [08:52] <@surkov> Jamie: if you mean accessible object then no, but if you mean accessible text then yes
- # [08:52] * Joins: logbot (logbot@moz-58CB32ED.glob.com.au)
- # [08:52] <Jamie> I mean text, yeah
- # [08:52] <@surkov> we keep the same hierarchy
- # [08:52] <@surkov> ok
- # [08:52] <@tbsaunde> Jamie: so, what happens if I do aria-label="foo bar\nbaz" ?
- # [08:52] <Jamie> err, is that even valid html?
- # [08:53] <Jamie> tbsaunde: the core problem si that authors and spec writers seme ot have this impression that screen reader users just tab around everywhere
- # [08:53] <@tbsaunde> Jamie: its a string why not?
- # [08:54] <Jamie> tbsaunde: afaik HTML strings don't permit backslash escapes like that
- # [08:54] <Jamie> they're just literal
- # [08:55] <@tbsaunde> Jamie: sure, so you just read \n, but what about spaces how do you handle navigating by words there?
- # [08:55] <Jamie> tbsaunde: Windows ATs don't use firefox caret navigation
- # [08:55] <Jamie> tbsaunde: we render the entire document into a buffer and handle the lot ourselves
- # [08:55] <@tbsaunde> Jamie: you still need to know where the words stop and end
- # [08:56] <Jamie> so we use an algorithm :)
- # [08:56] <@tbsaunde> Jamie: its crazy we and you both have an algorithm for that
- # [08:56] <Jamie> so in browse mode, lines and words don't agree with firefox's own idea of lines and words
- # [08:57] <Jamie> tbsaunde: sure, but what choice did we have? firefox's caret nav is still dodgy at best
- # [08:57] <Jamie> and that applies ot every other browser
- # [08:57] <@tbsaunde> Jamie: how is caret nav involved with finding word boundaries?
- # [08:58] <Jamie> tbsaunde: if we have ot render the whole document ourselves, it's easier to just handle everything ourselves
- # [08:58] * Quits: Jesse (jruderman@2557E599.66715431.D25A875A.IP) (Quit: Jesse)
- # [08:58] <Jamie> tbsaunde: so we suck up the document and then do our own navigation
- # [08:58] <Jamie> tbsaunde: checking for word offsets for every word would be a perf hit
- # [08:59] <Jamie> btw, we now use Windows Uniscribe to do the word finding logic, so at least we're using a common lib :)
- # [08:59] <@tbsaunde> Jamie: then it seems like the whole thing is a shit show and we just shouldn't bother?
- # [08:59] <Jamie> tbsaunde: shouldn't bother with what?
- # [09:00] <@tbsaunde> Jamie: more or less any text api
- # [09:00] <Jamie> tbsaunde: not quite
- # [09:00] <Jamie> tbsaunde: first, we need text attributes for browse mode
- # [09:00] <Jamie> tbsaunde: second, we don't use browse mode for *editable* text
- # [09:01] <Jamie> tbsaunde: third, orca needs this for both browsing and editing
- # [09:01] <@tbsaunde> Jamie: so, then what happens if I have <div contenteditable="true"> <a aria-label="foo">bar</a> ... </div> ?
- # [09:01] <Jamie> tbsaunde: ah. I have no idea.
- # [09:01] <Jamie> tbsaunde: NVDA doesn't support contentEditable very well. I gave up on that months ago due to bugs in Firefox's text implementation
- # [09:02] <Jamie> so in thunderbird, I can't write HTMl email
- # [09:02] <Jamie> tbsaunde: my comment re aria-label was more related to browsing. as far as editing goes, I think we'd ignore the aria-label; that's my feeling at this stage
- # [09:02] <Jamie> unless perhaps it was just a one character object
- # [09:03] <@tbsaunde> Jamie: I guess I'm miss understanding you, but my impression was you couldn't use our text support even if it didn't have bugs because of things it doesn't do like aria-label
- # [09:03] <Jamie> tbsaunde: someone in the ARIA spec community will shoot me for saying that, in which case they can provide a patch which fixes firefox, NVDA and the spec
- # [09:04] <Jamie> tbsaunde: errg. no, sorry. That was more of a side observation.
- # [09:04] <@tbsaunde> Jamie: ignoring it in one case butt not the other sounds crazy
- # [09:04] <Jamie> tbsaunde: In a land where only theory applies, I totally agree
- # [09:04] <Jamie> tbsaunde: however, the reality is that we generally edit things that are far less complex than we browse
- # [09:05] <Jamie> tbsaunde: I doubt you'd write a fully fledged web app using contentEditable alone
- # [09:05] <@tbsaunde> Jamie: except google docs / ether pad / the msft thing (office 360?)
- # [09:06] <@tbsaunde> Jamie: no, but you might write a word processor type thing
- # [09:06] <Jamie> even there, you don't tend ot do things like create a button with a background image and then replace it with an accessible label
- # [09:06] <Jamie> you might create4 a graphic, but that's okay, since a graphic is only one character
- # [09:06] <Jamie> another example is MathJax. they're trying to make math accessible by using aria-label
- # [09:07] <Jamie> it sucks, but since we don't have any way to do maths via a11y APIs, it's better than nothing
- # [09:07] <Jamie> so, <span aria-label="x^2">blah blah blah mathml here</span>
- # [09:07] * Jamie would love to know what Firefox caret nav does with MathML
- # [09:08] <@tbsaunde> Jamie: we should just come up with a reasonable api
- # [09:08] <Jamie> tbsaunde: we've been saying that for about 6 years now
- # [09:08] <Jamie> tbsaunde: no one has the funding, the resources or the people from different walks of life willing ot sit down and agree on anything
- # [09:08] * Joins: Jesse (jruderman@moz-537BCF9.hsd1.ca.comcast.net)
- # [09:09] <Jamie> tbsaunde: believe me, I'm not being painful. I agree with you. but it just hasn't happened and I don't see it happening any time soon
- # [09:09] <@tbsaunde> Jamie: that may be, but aria-label= isn't really a solution, though it may be a somewhat working hack for simple cases
- # [09:09] <Jamie> tbsaunde: we find it hard enough to find funding to keep NVDA alive let alone do stuff like this
- # [09:09] <@tbsaunde> Jamie: sure
- # [09:11] <@tbsaunde> Jamie: my point is consider what to do with \sum_{i=0}^{k} \frac{i}{(k-1)!}
- # [09:11] <Jamie> tbsaunde: if you care, http://community.nvda-project.org/ticket/1354 http://community.nvda-project.org/ticket/1362 http://community.nvda-project.org/ticket/2963
- # [09:12] <Jamie> those are all cases of people wanting aria-label or similar to be rendered instead of the real content
- # [09:12] <Jamie> I fought them for months
- # [09:12] <@tbsaunde> Jamie: maybe someday :) (proving your point sort of I know, but on other hand, maybe somebody with the need will decide to do it if they have no hacky alternative)
- # [09:13] <Jamie> Puristic and practical is a really hard balance
- # [09:13] <Jamie> Just recently, a site just had <a></a> with a background image and onClick logic
- # [09:13] <@tbsaunde> Jamie: so nnow there thing works in nvda, and maybe some comercial screen reader probably not orca and who knows about safari / mobile stuff
- # [09:13] <Jamie> I couldn't really use it, so I emailed and asked for aria-label
- # [09:14] <Jamie> without that option, it would ahve been "fix your link to have real content!" "but we don't want the text showing up on screen!" "so use off screen... oh wait, that's a terrible hack... um... use... something..."
- # [09:15] <Jamie> safari always takes aria-label
- # [09:15] <@tbsaunde> Jamie: true
- # [09:15] <Jamie> btw, what I haven't worked out is what the hell to do if someone does <div aria-label="blah">the entire body of the document here</div>
- # [09:16] <Jamie> I need to throw that at some supposed aria experts
- # [09:16] * Joins: marcoz (marco.zehe@moz-9D111104.dip.t-dialin.net)
- # [09:16] <Jamie> their response will doubtless be "don't override the content with aria-label". "oh, but you said aria-label should override the link name in all cases"
- # [09:16] * ChanServ sets mode: +o marcoz
- # [09:16] <@marcoz> Morning all!
- # [09:17] <Jamie> The ARIA community irritates me a great deal because of stuff like this
- # [09:17] <Jamie> marcoz: morning!
- # [09:20] <@tbsaunde> Jamie: == to the extent I have experience with them, but tbh I'm not sure what a reasonable API would be
- # [09:20] <Jamie> tbsaunde: oh, I'm not always sure either. it's just that they don't seem to understand the subtle complexities and then they get all "you should support the standard properly" when we don't do something right
- # [09:21] <Jamie> John Foliot has actually been known to suggest that no one should support NV Access at all because we didn't want to implement longdesc initially
- # [09:22] <@tbsaunde> Jamie: oic O.O
- # [09:26] <@marcoz> Jamie: WHAT? He actually said that? Unbelievable….
- # [09:29] <Jamie> marcoz: not quite those words. basically, he sent an email to the person managing html5accessibility.com saying that NVDA shouldn't be mentioned on the site because we don't support standards
- # [09:29] <Jamie> marcoz: think it was Steve Faulkner. I believe the response was fairly bluntly "no"
- # [09:31] * Joins: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP)
- # [09:32] <@marcoz> Thankfully so! Everybody is making such a big fuzz about longdesc. I'm quite indifferent myself, but it is a topic having people up in arms about it.
- # [09:33] <@tbsaunde> marcoz: guess we should service now you some more popcorn if your going to pay attention
- # [09:33] * Quits: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP) (Ping timeout)
- # [09:33] * @tbsaunde thread deletes bug mail when he sees bug numbers for longdesc
- # [09:37] <@marcoz> tbsaunde: I actually am not watching it too closely, but thanks for the offer. ;)
- # [10:06] * Joins: victorporof (victorporo@466CECCA.162B26E2.79933D60.IP)
- # [10:18] * Quits: Jamie (Instantbir@moz-CA26021.jantrid.net) (Quit: Instantbird 1.4a1pre -- http://www.instantbird.com)
- # [10:19] * Joins: rednaks (rednaks@24067163.74C6FD40.360EF119.IP)
- # [10:52] <@firebot> marco.zehe@googlemail.com set status-firefox22 to fixed on bug 857936.
- # [10:53] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=857936 nor, --, mozilla23, surkov.alexander, RESO FIXED, ARIA columnheader/rowheader should inherit editable state
- # [11:06] * Joins: habber (habber@514D186.F3BBB17D.144F44FA.IP)
- # [11:40] * Quits: rednaks (rednaks@24067163.74C6FD40.360EF119.IP) (Ping timeout)
- # [12:01] * Quits: @surkov (surkov@B3C081AC.7AE20955.EBE09E3C.IP) (Quit: surkov)
- # [12:04] * Quits: @marcoz (marco.zehe@moz-9D111104.dip.t-dialin.net) (Quit: Leaving.)
- # [12:18] * Joins: rednaks (rednaks@24067163.74C6FD40.360EF119.IP)
- # [12:43] * Joins: surkov (surkov@B3C081AC.7AE20955.EBE09E3C.IP)
- # [12:43] * ChanServ sets mode: +o surkov
- # [13:19] * Quits: rednaks (rednaks@24067163.74C6FD40.360EF119.IP) (Ping timeout)
- # [13:34] * Joins: scott_gonzalez (scott_gonz@moz-91C81A39.hrbgpa.fios.verizon.net)
- # [14:04] * Quits: @surkov (surkov@B3C081AC.7AE20955.EBE09E3C.IP) (Quit: surkov)
- # [14:15] * Joins: surkov (surkov@B3C081AC.7AE20955.EBE09E3C.IP)
- # [14:15] * ChanServ sets mode: +o surkov
- # [14:18] * Joins: yzen (Adium@AA2C085E.10C8AEF8.9A8C35B4.IP)
- # [14:18] * Joins: Justin_o (Justin_o@67828CC7.C1A51174.9D42CF23.IP)
- # [14:27] <@firebot> surkov.alexander@gmail.com requested needinfo from marco.zehe@google mail.com on bug 809338.
- # [14:27] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=809338 nor, --, ---, trev.saunders, NEW, make HTML select hierarchical
- # [14:43] * Joins: rednaks (rednaks@24067163.74C6FD40.360EF119.IP)
- # [14:55] * Quits: rednaks (rednaks@24067163.74C6FD40.360EF119.IP) (Ping timeout)
- # [14:55] * Joins: davidb (davidb@F2D29657.F60B0462.67AC9B1.IP)
- # [14:55] * ChanServ sets mode: +qo davidb davidb
- # [14:56] * Joins: rednaks (rednaks@24067163.74C6FD40.360EF119.IP)
- # [14:56] <@davidb> heyo!
- # [14:56] * @davidb ->coffee
- # [15:04] * Joins: marcoz (marco.zehe@moz-9D111104.dip.t-dialin.net)
- # [15:05] * ChanServ sets mode: +o marcoz
- # [15:05] <@marcoz> Good day all!
- # [15:06] * Quits: rednaks (rednaks@24067163.74C6FD40.360EF119.IP) (Client exited)
- # [15:06] * Quits: @surkov (surkov@B3C081AC.7AE20955.EBE09E3C.IP) (Quit: surkov)
- # [15:07] * Joins: clown (clown@67828CC7.C1A51174.9D42CF23.IP)
- # [15:07] <@firebot> yura.zenevich@gmail.com cancelled feedback?(eitan@monotonous. org) for attachment 736122 on bug 811307.
- # [15:07] <@firebot> yura.zenevich@gmail.com requested feedback from eitan@monotonous. org for attachment 736260 on bug 811307.
- # [15:07] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=811307 nor, --, ---, yura.zenevich, NEW, [AccessFu] Add mochitest for enabling
- # [15:08] <@firebot> yura.zenevich@gmail.com cancelled feedback?(eitan@monotonous. org) for attachment 736123 on bug 811307.
- # [15:08] <@firebot> yura.zenevich@gmail.com requested feedback from eitan@monotonous. org for attachment 736262 on bug 811307.
- # [15:10] <yzen> marcoz: thank for spotting those issues, i updated the patches :)
- # [15:11] <yzen> marcoz: by the way, I also tested it on the device and it seems to work as before
- # [15:15] <@marcoz> yzen: Thanks! I will do a test run of my own later today, too. :)
- # [15:17] * @davidb gulps coffee
- # [15:19] * Joins: peteb-away (ptbrunet@moz-B51E1692.austin.res.rr.com)
- # [15:22] * Joins: rednaks (rednaks@24067163.74C6FD40.360EF119.IP)
- # [16:06] * clown is now known as clown_mtg
- # [16:24] * Joins: lizzard (ehenry@moz-5A4BE8E8.dsl.static.sonic.net)
- # [16:28] * Quits: rednaks (rednaks@24067163.74C6FD40.360EF119.IP) (Quit: Téléportation !)
- # [16:36] * Quits: icaaq (Adium@moz-D8A76420.cust.bredbandsbolaget.se) (Ping timeout)
- # [16:38] * Joins: icaaq (Adium@moz-D8A76420.cust.bredbandsbolaget.se)
- # [16:43] * Joins: nhirata (anonymous@moz-1A50F7F8.hsd1.ca.comcast.net)
- # [17:10] * Quits: a-865 (fmcz@moz-8F21088B.cable.mindspring.com) (Ping timeout)
- # [17:15] <@firebot> surkov.alexander@gmail.com changed the Assignee on bug 860013 from nobody@mozilla.org to surkov.alexander@gmail.com.
- # [17:15] <@firebot> surkov.alexander@gmail.com requested review from trev.saunders@gmail .com for attachment 736300 on bug 860013.
- # [17:15] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=860013 nor, --, ---, surkov.alexander, NEW, navigation by words broken with orca
- # [17:17] * Quits: Justin_o (Justin_o@67828CC7.C1A51174.9D42CF23.IP) (Quit: Justin_o)
- # [17:18] * Quits: @marcoz (marco.zehe@moz-9D111104.dip.t-dialin.net) (Quit: Leaving.)
- # [17:22] * Quits: lizzard (ehenry@moz-5A4BE8E8.dsl.static.sonic.net) (Quit: lizzard)
- # [17:26] * Joins: Justin_o (Justin_o@67828CC7.C1A51174.9D42CF23.IP)
- # [17:27] * Joins: lizzard (ehenry@moz-5A4BE8E8.dsl.static.sonic.net)
- # [17:29] * Joins: marcoz (marco.zehe@moz-9D111104.dip.t-dialin.net)
- # [17:33] * Quits: habber (habber@514D186.F3BBB17D.144F44FA.IP) (Quit: habber)
- # [17:36] * Quits: nhirata (anonymous@moz-1A50F7F8.hsd1.ca.comcast.net) (Ping timeout)
- # [17:39] * Joins: nhirata (anonymous@moz-1A50F7F8.hsd1.ca.comcast.net)
- # [17:41] <@firebot> marco.zehe@googlemail.com cancelled needinfo?(marco.zehe@google mail.com) on bug 809338.
- # [17:41] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=809338 nor, --, ---, trev.saunders, NEW, make HTML select hierarchical
- # [17:44] * Quits: Justin_o (Justin_o@67828CC7.C1A51174.9D42CF23.IP) (Quit: Justin_o)
- # [17:45] * Joins: Justin_o (Justin_o@67828CC7.C1A51174.9D42CF23.IP)
- # [18:03] * Quits: nhirata (anonymous@moz-1A50F7F8.hsd1.ca.comcast.net) (Ping timeout)
- # [18:12] * Quits: marcoz (marco.zehe@moz-9D111104.dip.t-dialin.net) (Ping timeout)
- # [18:23] * khuey|away is now known as khuey
- # [18:29] * Quits: yzen (Adium@AA2C085E.10C8AEF8.9A8C35B4.IP) (Quit: Leaving.)
- # [18:45] * Quits: clown_mtg (clown@67828CC7.C1A51174.9D42CF23.IP) (Quit: Leaving.)
- # [18:57] * Joins: clown (clown@67828CC7.C1A51174.9D42CF23.IP)
- # [18:57] * clown is now known as clown_mtg
- # [19:04] * Joins: Gijs (gijs@2557E599.66715431.D25A875A.IP)
- # [19:05] * Quits: @davidb (davidb@F2D29657.F60B0462.67AC9B1.IP) (Ping timeout)
- # [19:05] * Joins: davidb (davidb@F2D29657.F60B0462.67AC9B1.IP)
- # [19:05] * ChanServ sets mode: +qo davidb davidb
- # [19:12] * Joins: yzen (Adium@AA2C085E.10C8AEF8.9A8C35B4.IP)
- # [19:14] * Joins: rednaks (rednaks@77F0CB85.5C1AD644.55FFA9B4.IP)
- # [19:28] * Joins: a-865 (fmcz@moz-8F21088B.cable.mindspring.com)
- # [19:44] * Quits: icaaq (Adium@moz-D8A76420.cust.bredbandsbolaget.se) (Ping timeout)
- # [19:44] * Joins: icaaq (Adium@moz-D8A76420.cust.bredbandsbolaget.se)
- # [20:25] * clown_mtg is now known as clwon
- # [20:25] * clwon is now known as clown
- # [20:29] * Quits: Justin_o (Justin_o@67828CC7.C1A51174.9D42CF23.IP) (Quit: Justin_o)
- # [20:31] * Joins: nhirata (anonymous@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net)
- # [20:31] * Quits: icaaq (Adium@moz-D8A76420.cust.bredbandsbolaget.se) (Ping timeout)
- # [20:32] * Joins: icaaq (Adium@moz-D8A76420.cust.bredbandsbolaget.se)
- # [20:34] * Quits: clown (clown@67828CC7.C1A51174.9D42CF23.IP) (Quit: Leaving.)
- # [20:40] * Joins: clown (clown@67828CC7.C1A51174.9D42CF23.IP)
- # [20:46] * khuey is now known as khuey|away
- # [20:57] * Quits: @davidb (davidb@F2D29657.F60B0462.67AC9B1.IP) (Ping timeout)
- # [20:57] * Joins: davidb (davidb@F2D29657.F60B0462.67AC9B1.IP)
- # [20:57] * ChanServ sets mode: +qo davidb davidb
- # [21:42] * khuey|away is now known as khuey
- # [21:55] * Quits: brambles (xymox@moz-969AAE9B.barwen.ch) (Ping timeout)
- # [21:57] * Joins: brambles (xymox@moz-969AAE9B.barwen.ch)
- # [22:19] * Joins: fxa90id (fxa90id@moz-D8C47538.dsl.dynamic.t-mobile.pl)
- # [22:29] <@firebot> gavin.sharp@gmail.com requested review from jaws@mozilla.com for attachment 735620 on bug 448945.
- # [22:29] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=448945 enh, P4, Future, nobody, NEW, minimum font size should offer 32
- # [22:41] * Quits: @davidb (davidb@F2D29657.F60B0462.67AC9B1.IP) (Quit: davidb)
- # [22:56] * Quits: icaaq (Adium@moz-D8A76420.cust.bredbandsbolaget.se) (Quit: Leaving.)
- # [22:57] * Joins: icaaq (Adium@moz-D8A76420.cust.bredbandsbolaget.se)
- # [23:02] * Parts: clown (clown@67828CC7.C1A51174.9D42CF23.IP)
- # [23:06] * Quits: yzen (Adium@AA2C085E.10C8AEF8.9A8C35B4.IP) (Quit: Leaving.)
- # [23:15] <@firebot> trev.saunders@gmail.com changed the Component on bug 860913 from Disability Access APIs to DOM.
- # [23:15] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=860913 nor, --, ---, nobody, NEW, Fix MSVC build warnings in content/html/content/src/ and mark it as FAIL_ON_WARNINGS-for-all-compile
- # [23:15] <@firebot> jaws@mozilla.com granted review for attachment 735620 on bug 448945.
- # [23:15] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=448945 enh, P4, Future, nobody, NEW, minimum font size should offer 32
- # [23:16] <@firebot> jaws@mozilla.com changed the Assignee on bug 448945 from nobody@mozilla.org to cykesiopka@hotmail.com.
- # [23:16] <@firebot> jaws@mozilla.com changed the Status on bug 448945 from NEW to ASSIGNED.
- # [23:17] <@firebot> jaws@mozilla.com changed the Target Milestone on bug 448945 from Future to ---.
- # [23:46] * Quits: Gijs (gijs@2557E599.66715431.D25A875A.IP) (Quit: brb)
- # [23:48] * Joins: yzen (Adium@AA2C085E.10C8AEF8.9A8C35B4.IP)
- # [23:48] * Quits: scott_gonzalez (scott_gonz@moz-91C81A39.hrbgpa.fios.verizon.net) (Quit: scott_gonzalez)
- # Session Close: Fri Apr 12 00:00:01 2013
The end :)