See also: IRC log
<alessio> hallo
<Steven> hi there
<markbirbeck> (guessing)
<Steven> zakim IP is Alessio
<alessio> thanks steven
<Steven> Scribe: Gregory
<Steven> ScribeNick: OedipusWrecked
SP: announcements: questionaire for coming
calls - which weeks people taking vacation;
... please fill out form if not already
... also questionaire about next Face2Face in Spain -- PLEASE fill in whether
coming or not
... possibility of remote attendance, but need a head-count A.S.A.P.
SP: have an action item to ask for a wiki; told that w3c is going to transfer / migrate to mediawiki, which is probably preferable; put action item on hold until migration complete
SP: XHTML1 Basic assertions need to be
explained -- why is one thing orange and the rest not?
... next step is CR -- assuming tomorrow's call goes well, will be in CR in a
few days; need to ping jan -- need to discuss pattern for testing
... had comment from W3C team that should have tests for input mode and
should have at least 2 implementations that ACTUALLY do something with it
... possible to ignore, but not possible for us to ignore;
... feature should be tested and excercised at least twice
... that is fine - need test to try out and implementations that actually do
what the feature specifies
Shane: putting stuff in mobile UAs that take a
while; working on implementations here - perhaps can get implementation on
our side
... here?
Mark: don't know
SP: put the keyboard in a state so that you get particular characters by default -- capitals, alphanumeric characters
<ShaneM> XHTML Basic editors draft has been updated and is at http://www.w3.org/MarkUp/2007/ED-xhtml-basic-20070711/
MB: would they be available by themselves -- can have braille; build for what's available and switch to that mode when necessary
SP: good point to bring up in XForms group; have mobile implementations there
<yamx> zakim+??P37 is yamx
SP: traditional call tomorrow at regular time -
link in the agenda, please pass an editorial eye over it; had a comment from
within w3c team to exit CR need tests that test optional features -- in
particular input mode -- 2 implementations that do input mode
... does that seem likely in short time frame?
Jan: i think so
SP: do you have a feeling for when would be likely to exit CR (2 implementations for each feature)
Alessio: 1.0 and 1.1 test space;
SP: by october?
Alessio: next february was target, but have 1 to 1.1 tests; new tests will use mobile profile 1.1 will already be published by then
SP: last test results you sent to list, most of tests not run; have to have more tests of the tests
Alessio: test for mobile 1.1; in isreal have test for input mode; rest will be derrived from ARIA profile
SP: need potential date for coordination call -- october or 3 months later
RESOLUTION: target to exit CR with completed tests: October 2007
scribe's correction - not a coordination call, but a joint call
<yamx> OMA has test fests in September (TestFest21) and November(TestFest22).
SP: replied to schema group - posted proposed reply; if anyone knows modularization PLEASE take a look at my comments and make sure that you agree -- looking for help, especially shane and mark and anyone else who knows schemas and modularization
SP: access module - shane must send proposal; for role, were going to go to last call but...
Shane: ready to go as far as i am concerned - put a draft up and see what happens
<Steven> http://www.w3.org/MarkUp/Drafts/#xhtml-role
SP: have public draft and new editor's draft from this month (july 2007); is everyone happy with us taking it to last call
Rich: yes
SP: would anyone object to us taking this to last call?
RESOLUTION: take role module to last call; SP will set in motion after negotiating space and time with Hypertext CG;
SP: we have a face2face where can deal with
last call issues at face2face; if we suggest a last call that ends on the
last day of august, gives SP time to chase down people and collect all
comments?
... six weeks reasonable?
yes
<ShaneM> Note that the spec does NOT have a schema implementation in it. My last call comment will be that it needs one, but that we need M12N 1.1 ready first.
<scribe> ACTION: SteveP - prepare last call draft for the role module [recorded in http://www.w3.org/2007/07/11-xhtml-minutes.html#action01]
SP: shane - interesting last call comment; could put on hold and give priority to getting modularization 1.1 out the door
Shane: no No NO!
SP: then have to meet last call for roles
Shane: if don't have schema implementation, will get formal objection
Rich: i concur
Shane: modularization 1.1 is ready; made changes we are going to implement
SP: add schema comments to database; then ask for transition call on modularization
Shane: going on holiday - can we wait until i get back?
SP: yes, but just to be clear, next step is to add comments to schema
Shane: yes
SP: are we happy with "next id in target area"? more sensible to say, yes, next one of those in order; take list in attribute order, of take list and find first one in document order; need to agree which is best
Shane: targetrole takes mulitple values; have them all do the same thing; need to resolve potential confusion
Rich: documentorder necessary
SP: either way is implementable; let's not make
decisions based on easiest for author, need it to be correct
... can't replicate in any other way, already there
... or using another feature, such as TABINDEX;
... should be sure feature doesn't collide with next focus - default action
key in a way
... if i have target = a and land on a which has next focus of b, follow
document order in that route?
??: document order not order of the list -- order that one navigates, should be plusNextFocus; other possiblity list of icons to be cycled through using a key definition
Rich: pass over normal focus?
SP: yes
??: id or role?
SP: didn't know if could have multiple ids; one roll gives many
??: understand your arguments, but if have doc with roles a b & c - TargetRole a b c - when activate access key, where does it go -- to all the As and then the Bs or movev onto the Bs
Rich: want multiple accesskeys for each role
... next in sequence is any one of the following
SP: rule for id
??: just wanted consistency -- people have discussed mapping in different fora - can always come back to it; like a flag that says these items qualify for focus, then provide option to go down list of As or tab to next set of Bs
SP: example i used to show NextFocus working: in a table want to navigate in different order than UA; specifying with NextFocus is something that overrides document order
MB: good example - if have table, can go to any row or column, set an accesskey for SeekTheCell -- knock off the head, just have the cell, should still do in row order or column order
SP: default document order; if redifine using
NextFocus to do column order navigation - using accesskeys should honor that
route through the table
... next focus redefines what document order is
... allows author to skip non-navigatable sections - but non-navigatable
section still available; tab from one table to next table without
interspersed text
Rich: when end of document wraps around standard?
SP: yes
MB: some browsers, go off the document into the chrome; actually a bug; we defined it in XHTML2
SP: should declare going to chrome in tabnavigation is invalid
Rich: incorrect behavior
GJR: +1
SP: so treat tab navigation to the chrome as abberant behavior?
MB: what next?
SP: where the definition should go - module is independent of NextFocus; could have lang with accesskey but not next focus; put in access module what happens when have NextFocus like mechanism, or put into NextFocus module, say this redefines concept of document order; should we use the term "navigation order"? -- perhaps that is the answer
Rich: should be "navigation order"; have a flowto mechanism in ARIA - if there is one id, go to next id, can skip the order -- this is Assisstive Tech, not UAs; if multiple IDs, user should be given a choice of which to navigate to
SP: only 1 ID per element
MB: do we not have a concept of navigation order?
SP: not as such - should introduce concept
MB: if conflict in navigation order - sequence
defined in explicit way; role and ids allow to stop if matches role in list,
if not hop over it
... could allow both attributes at the same time
<Tina> I'm having a REAL bad line.
SP: yes,
... id main and any element that have corresponding roles; couple of menues
with IDs some with roles
... sub-sets of navigation order
shane: just don't know what to do next
RESOLUTION: targetrole and targetid can take multiple values; accessvalue can have one or more; defines subset of navigation order, which is something we must define;
Rich: in HTML4 context, TABINDEX
SP: in XHTML2 will be tabindex plus document order
<markbirbeck> ...plus nextfocus.
SP: may have seen email to MB; question about
multiple ids on single element; XML Schema group has decided to allow
multiple IDs on an element; don't know when that behavior will be defined, so
can do what we need without breaking rules
... no reason for single ID restriction; nothing really breaks if something
has 2 names; sent comment to XML WG answered, yes, this is a bug, needs to be
fixed; resolved to change it, so heads up
Rich: like what done in XHTML2
SP: discussed endlessly and came up with solution; HTML5 case - allow image to have content - not to be empty element, but element with children; if source fails, image not available, image loading off, can use content of the image element
Rich: image shows can't navigate content, so where going to get description? ATs can get text for you; description relative to an item on same page -- describedby; objection to longdesc is someone has to create whole new web page and maintain it
MB: image as container fallback text
... text might say "you have image loading turned off, please turn on
SP: something that performs the image if not there -- that's fallback -- alternative content
MB: longdesc implies this is a description of this item; 2 exist at same time; not same thing if multiple fallbacks
SP: the whole point of fallback in XHTML2 is to say these 2 things are equivalent, end of story; move longdesc into document
Rich: like XHTML2 and describedby what need to push
GJR: will add links to minutes to wiki page on longdesc; and point to describedby URI test page
meeting adjourned
<Tina> Who is 'MB' again?
MB is Michael Birbeck (sp?)
<Steven> Mark Birbeck
<markbirbeck> Steven--if we were to give special status to text used in fallbacks then we could cover all use-cases. We could simply say in XHTML 2 that the UA should make fallback text 'available' regardless of which of the items is actually used. Then it really does become like @longdesc, and it also means that Rich doesn't need to have a separate area for his descriptive text.
URI for HTML Wiki discussion of longdesc: http://esw.w3.org/topic/HTML/LongdescRetention
<Steven> I agree Mark, good idea
This is scribe.perl Revision: 1.128 of Date: 2007/02/23 21:38:13 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/??/MB/ Succeeded: s/??:/MB:/ Found Scribe: Gregory Found ScribeNick: OedipusWrecked Default Present: Roland_, Steven, ShaneM, [IPcaller], +004670855aaaa, TIna, Gregory_Rosmiata, +0208761aabb, markbirbeck, Alessio, Rich Present: Roland_ Steven ShaneM [IPcaller] +004670855aaaa TIna Gregory_Rosmiata +0208761aabb markbirbeck Alessio Rich Agenda: http://lists.w3.org/Archives/Public/public-xhtml2/2007Jul/0010 Got date from IRC log name: 11 Jul 2007 Guessing minutes URL: http://www.w3.org/2007/07/11-xhtml-minutes.html People with action items: stevep WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option.[End of scribe.perl diagnostic output]