Accessibility Minutes 2012 03 26
From MemberWiki
Present
- Jon Gunderson (University of Illinois - Co-Chair)
- Nicholas Hoyt (University of Illinois)
- Rich Schwerdtfeger (IBM - Co-Chair)
- Prasanna Bale (University of Illinois)
- Marc Johlic (IBM)
- Ann Abbott (IBM)
- Philip Ackermann (Fraunhofer Institute for Applied Information Technology) - Scribe
Minutes
Jon: Any more feedback to the Accessibility Inspector?
Jon: Send any comments to me or to the list. Including the URLs of tested webpages in case of problems.
Ann: "Script having a problem / script hangs" kind of error
Jon: Currently re-evaluates the webpage every time it reloads
Ann: Try to reproduce the error and let you know.
Jon: Next agenda item: relationship between rules, rulesets and results
Jon: four different types of results: pass / fail / manual check / not applicable
Nick: the conceptual model that is necessary for modeling rules and rulesets is going to be complex if you want cover everything
Nick: send email to the list
Nick: two types of rules (required / recommended) --> different types of results
Nick:I would say that there are five types of results
Nick: one top level object is "rule-result"
Nick: results are tied to the different types of rules
Ann: what if one technique of a success criteria fails?
Ann: multiple sufficient techniques --> only one of them passes --> passing that rule for WCAG
Nick: conceptual model for web accessibility evaluation
Nick: see email for detailed description
Nick: understanding evaluation results in different context, one of the for example WCAG
Nick: list of rules --> mapping rules to WCAG into guidelines, criteria, techniques
Nick: what is a "rule"?
Nick: I would say an "interpretation of technique to meet a success criteria"
Ann: http://www.w3.org/WAI/WCAG20/quickref/#qr-navigation-mechanisms-mult-loc
Nick: how many of the success criteria have this case, where one technique is sufficient?
Nick: Do we have a rule that cover 2.4.5?
Nick: Landmark rule
Jon: manual checks, for example alternative text for images
Jon: maybe over a hundred warnings
Nick: what Mark was saying: recommended violation / recommended manual check
Nick: if the results are broken down, recommended separated from required ones
Nick: users can prioritize
Jon: do we need to have some kind of logic for success criteria, e.g. "two techniques required"
Nick: how do we map those?
Nick: need to think about this
Jon: at the moment rules can be associated with techniques, but no way to describe the logic mentioned above
Philip: similar ontology (in RDF) started at Fraunhofer, but it is not finished yet
Jon: will discuss the models with PB and NH and prepare a proposal for next week
