Accessibility Minutes 2009 04 08

From MemberWiki

Jump to: navigation, search



  • Ann Abbott (IBM)
  • Jon Gunderson (University of Illinois)
  • Nathan Jakubiak (ParaSoft)
  • Vio Onut (IBM)
  • Tony Salcedo (ParaSoft)
  • Rich Schwerdtfeger (IBM - chair)
  • David Todd (IBM)



Topic: David's WCAG2 action item to investigate table captions

Jon: expressed concern regarding screen readers voicing table captions as part of gridcell.

David: re-read technique regarding Captions

David: if not providing Captions, Caption element is not required - is that correct?

David: wondering if anyone else has another interpretation?

JonGunderson: orientation is required?

Rich: this is a technique - is it limited to data tables?

JonGund: repair oriented technique? Doesn't seem to require Caption? Ambiguous.

David: if using Caption, use caption element.

JonGund: caption is indisguishable from first data cell.

JonGund: problematic for screen reader users

Rich: sounds like if providing Caption, use caption element

JonGund: JAWS table navigation - does it read?

JonGund: requirement or nice to do?

David: if using caption on table, use caption element

Rich: test case?

Rich: caption may not make it into HTML 5

Rich: is test case a warning?

JonGund: not having a caption would be a warning

JonGund: summary should be required and we should develop rules.

JonGund: will this help group we're trying to help - PwD?

David: table caption not used a great deal - messes up layout

JonGund: can be hidden

Rich: detect data table, fire warning that a caption may be required?

Rich: if caption is provided, recommend using caption element

JonGund: is there another element? Labelby?

Rich: don't need special elements

JonGund: problem with caption needs to be communicated to AT vendors and needs to be fixed.

JonGund: caption is used so infrequently, ATs may not choose to fix

Rich: leave it off?

JonGund: can you bring this up with Freedom? Leave it in for now. Will create test pg and can provide to vendors

Rich: We have violation, potential violation and potential recommendation - what is the correct

JonGund: potential recommendation - if using caption, should use caption element

JonGund: how to test to know if using Caption?

Rich: how about aria-labeledby?

JonGund: what will AT vendors do with aria-labeledby

Rich: willing to get in front of ATs

JonGund: prefers labeledby for consistency

JonGund: labeledby has broader usage

David: table gets focus, summary gets read

Rich: in case of screen reader, doesn't automatically read description.

Rich: data table using captions (WCAG2 now) or data table using aria-labeledby?

Nathan: leaning toward aria-labeledby

Rich: let's stick with aria-labeledby due to problems with AT handling of captions

David: need to be backward compatible?

Rich: JAWS - not a problem. Cant speak for other AT vendors

Rich: Summary - must be associated with table. Generate potential violation.

David: violation instead of potential violation?

JonGund: Data Table Rules

Rich: if it is determined to be a data table, generate a violation when a table summary is not provided or aria-describedby

Topic Inner HTML/ Document.write

David Todd's post

David: WCAG doesn't like it because of mime type

Rich: is this a violation for accessibility?

Rich: is this problem with XHTML only?

Rich: document.write can't be validated/parsed

Rich: unless mediat type is XHTML, leave it alone...

Nathan: flag as advisory but not sure it makes page inaccessible

David: is under Information Relationships/Technique A/programatically determinable


Rich: This is one technique for meeting WCAG 2 but doesn't personally doesn't support - no basis on Accessibility - checkponts prior require semantics - take to WCAG 2 working group as it doesnt ensure semantic markup

Nathan: we have warning but has limited use

Rich: would have to mark live regions on this page.



annabbott: Rich: this would be very noisey, you're right...

annabbott: Rich: if tied into browser and have updates on page - if area is in area in focus or captured by live region, we could flag warning which would be very verbose

JonGund: can't determine live region if not marked up unless something in a browser looking for content updates.

JonGund: does it generate dom mutation?

Rich: since test strategy, working with dynamic page, think aria should apply here

Rich: Will take up innerHTML and document.write() with the WCAG working group.

Personal tools