The Linux Foundation


From The Linux Foundation


Minutes of the Open A11y Expert Handlers SIG Call 2007/10/01


  • Neil Soiffer (NS/chair)
  • Vladamir Bulatov (VB)
  • Pete Brunet (PB)
  • Gregory J. Rosmaita (GJR/scribe)
  • Janina Sajka (JS)
  • Shawn Djernes (SD/IRC only)
    • regrets: none logged

Agenda Review

Approval of Minutes from 24 September 2007 Meeting

NS: any objections?

none articulated

NS: then, without objections, the minutes approved

RESOLVED: minutes from the 24 September 2007 Expert Handlers Conference Call approved.

Minutes of Expert Handlers Conference Call 2007/10/01

Topic 1: CSUN 2008 Panel Proposal Finalization

no comments on last draft at present -- action item JS/NS to draft revised draft continued;

  • deadline: new draft to be posted to the Handlers list by 7 October 2007

Topic 2: Testing Protocol for Expert Handlers

JS: testing -- not only publish specifications, but have testing for people to evaluate and -- perhaps -- for certification purposes; standard procedure coming out of LSB (Linux standard base -- need to conform); AT-SPI spec will have testing suite/protocol which will provide proof-of-concept; that's what we're working on in the main Open A11y WG; trying to obtain funding to generate tests

JS: if we succeed will be the first -- only web content (WCAG or 508) has conformance testing, and most of that is either private or proprietary; testing of ATK/AT-SPI and IAccessible2 will ensure that they are able to interact with, understand, and not interrupt specialized applications or specialized content parsers

NS: long way from there; let's turn to the question of whether we need a separate input handler?

Topic 3: Is there a need for an Input Expert Handler Use Case?

NS: discussion of use case for Expert Handler that deals with other forms of input; dubious about absolute need, but we should discuss

JS: case for the Use Case: politically, good to cover all different ways that present barriers to using and editing content accessibly; covered technology as to how you handled it, navigation and structured navigation -- on-screen keyboard, mouse morse input; smart content navigation -- describing problem is politically smart thing to do, and should trigger feedback from people for whom input is the issue

NS: input derived from navigability use cases?

JS: no real distinction; simply need paragraph or 2 about perspective of someone using alternate input;

NS: ok, so a navigability sub-set?

JS: perhaps a small section, stating: "for the purpose of someone who uses AT for input, we provide navigability to put point of regard as close to where focus needs to be, AT can pick up from there and establish focus for editing

NS: is there really something that specific about expert software that makes this a particular concern? trying to understand why not a general purpose issue out of Expert Handlers interface's scope?

JS: the need is for navigability of content and a standardized way to present that content and allow user interaction with the content, in particular in the editing context; need standard to perform specific actions so as to drive a specific AT through a generic handler

NS: software needs to think about non-keyboard methods of transversal

JS: yes, it is another group of users who need smart navigation; another set of users that need smart navigation for different reasons than non-visual users; in scenario, can perceive content, but can only interact with and edit it via an alternative input mechanism

NS: so, in music, a marked up score could have buttons to walk by note, bar, phrase, etc. -- using buttons or keyboard equivalents; what is making handler software special for non-keyboard users as opposed to any generic software component?

JS: dealt with in some sort of specialized way to AT -- needs to be exposed by IAccessible2 or ATK/AT-SPI so the 2 can interchange information and commands; GOKis all about getting to a particular place in an accurate expeditious manner; more than simply forward, back, etc. -- in music example, there is not much need for output handler using AT's output -- that isn't an issue if user can read, but cannot input; if user needs to track the score of a symphony where each staff is a the part for a different instrument, user may want flow-to through the french horn part as user moves through the document or makes changes to the score -- editing needs smooth traversal; blind user might want french horn part brailled or played against backgrounds of other notes to hear how part fits into overall context of the score better; if composing, need to get to right spots in each of the horizontal lines on which music is written -- staves a complex issue from editing perspective, but it is being done through kludges right now

NS: will you take an action item?

JS: just another pointer to importance of navigation for another set of users; providing levels of granularity -- is that sufficient? smart navigability and supporting AT through IAccessible2, LibATK, AT-SPI

ACTION JS: Propose text that stresses importance of navigation for discrete sets of users using smart navigability and supporting pertinent AT through IAccessible2, LibATK, AT-SPI

Topic 4: Web Site Update & Issues

NS: should there be links to internal reference page with point-of-regard to referenced markup pointer plus contextual information or just pointers to ML specs themselves (as is currently the case)

JS: similar to an explanatory footnote -- all the contextualizing doesn't belong in text; yet, explanation in context important, not specification details of cited example; keeps in context of Expert Handlers -- if user wants to jump into specification, can do, otherwise can stay here and peruse other resources, especially local ones

GJR: naming convention if need point to a specific ML, mathml-info all small cellml-info -- sample wikified link text follows:


RESOLVED: GJR will change links in mission statement to internal links

GJR: encourage members to use the "Discussion" pages to propose and/or answer questions about a document, its assertions, etc.; helps to have comments available in context

Topic 5: Identify Agenda Items for 2007/10/08 Conference Call

NS: concerned about rate of progress: we need to discuss deeper issues; any suggestions for next week?

JS: prioritize; Use Cases pretty much finished - just need some clean up and polish to take to Open A11y as a whole for feedback and approval

NS: would like a discussion as to where we see Expert Handlers fitting in relationship to AT and other standards? should AT use MSAA or IAccessible2 to find markup and then use the standards we define to pass pertinent info to an Expert Handler, which then returns something via the AT, or should an Expert Handler enable direct interaction between AT and specialized content? what and which are we proposing?

JS: should either go to main A11y working group as far as linux goes and to IAccessible2 SIG for Ia2 and MSAA questions

PB: can we get someone from the linux side on a Handlers call?

JS: person i want is Bill Haneman or Willie Walker on call; next week is Columbus Day, which clashes with GNOME a11y summit, so may not be feasible for next week

PB: we could at least give WillieW and/or BillH description of problem statement so can discuss at conference

JS: good idea; need to email WillW before Friday (mozilla a11y meeting)

PB: perhaps you and Will can talk at summits?

JS: or GeorgeK (gk4)

NS: do you just want to talk to them, or should i send an email request?

PB: anyone understand problem clearly enough to write it up?

JS: it would be helpful if NS can expand on that

PB: willie will have a lot of good input on that

ACTION NS: write something concerning where Expert Handlers fit in chain/flow of control

ACTION GJR: when NS pushes to list, push to wiki and send URI to list

NS: should we have an October 8 meeting? JS will you be present?

JS: hoping to get back between Saturday and Sunday

NS: anyone have problem with meeting on Columbus Day

none enunciated or recorded

NS: ok, we will meet on the eighth; i will be away on the fifteenth

JS: need to take care of CSUN proposal -- due October 15, 2007

NS: should be able to tackle that

NS: other issues?

none enunciated or recorded

Wrap Up

Summary of Action Items

  • ACTION JS: Propose text that stresses importance of navigation for discrete sets of users using smart navigability and supporting pertinent AT through IAccessible2, LibATK, AT-SPI
  • ACTION NS: write something concerning where Expert Handlers fit in chain/flow of control
  • ACTION GJR: when NS pushes to list, push to wiki and send URI to list

Summary of Resolutions

  • RESOLVED: GJR will change links in mission statement to internal links

[Article] [Discussion] [View source] [History]