Reporting of accessibility problems

Summary: How should the WAI accessibility guidelines be integrated into 
accessibility analysis and reporting tools like Bobby.

The WAI Accessibility guidelines, with Required and Recommended problem 
severity, should be the basis for a measure of URL accessibility. 
What measure should we recommend:

1. Should any failing of a Required guideline in the content of a URL 
   report that URL as an accessibility failure?

2. Should any failing of a Recommended guideline report a warning?

3. Should summary counts of total failings and warnings be in the report?

4. Should further breakdown of summary counts be by kind?

Current Bobby uses an opaque algorithm to rank a URL as 1 to 4 stars. 
Most results are either 1 star (bad) or 4 stars (assertedly accessible).
 
Sites that generate frame content (by script, applet, or activeX-control)
may have nothing in the body for Bobby to analyze, so may get 4 stars, 
even though they may be totally inaccessible.

5. Should the amount of accessible material be factored into the report?

6. Is there a tactful way to make such reports to sites other than your own?

Background:

My prior attempts at reporting Bobby results to different companies have 
been minimally effective, even though I provided summaries of the problems
encountered, encouraged Bobby use, and showed how to use it: 

*   Two passes three months apart over the home pages of about 50 SGML 
    Open member companies (showing most had problems, and a trend for 
    more getting worse than getting better, as their web pages evolved 
    into more layout-intensive graphics, frames, imagemaps, and layout 
    tables.)

*   One pass over the home pages of all the companies supporting the Microsoft
    Content Data Format. I sent this to key players at Microsoft for further
    distribution. No one at Microsoft acknowledged that combined report.

Josh Krieger says he will soon be revising Bobby analysis to recognize the 
problems identified in the accessibility guidelines.

The questions I pose above should help with that new Bobby analysis.

Regards/Harvey Bingham

Received on Monday, 9 March 1998 23:12:57 UTC