- From: Hanno Lans <Hanno.Lans@Kinggemeenten.nl>
- Date: Tue, 10 Jun 2014 11:32:45 +0000
- To: "public-auto-wcag@w3.org" <public-auto-wcag@w3.org>
Automated WCAG Monitoring Community Group Teleconference 2014-06-06, 15:00 - 16:00 h (CET) Attendees: Annika Nietzio, Kamyar Rasta , Wilco Fiers, Hanno Lans Annika: propose to write down action points also in the wiki, with permanent links - (https://www.w3.org/community/auto-wcag/wiki/Action_items EARL Annika: investigated the mapping of the template model to EARL. You can see the results here: https://www.w3.org/community/auto-wcag/wiki/Result_template It is mapping of basic single tests and it maps well. both for succes criterion, or lower level: test case in EARL terminology. You can have one single EARL outcome Wilco: what is the difference between the info part and a description part? Annika: description is more detailed, in the description part you can for example mention the techniques and failures Wilco: Is EARL good model for us? Annika: yes, it is a good model Hanno: are these tests repeated for each appareance of a selector in a page? That are massive result sets for all success criteria. Annika: yes Hanno: with QUAIL we currently get around 1500 results for each page. We plan to keep the failures, but only the amount of passes and not other oass details. Annika: yes, we did that in our program as well Hanno: shouldn't we give feedback with the WCAG2 techniques and WCAG2 failures? Wilco: WCAG2 failures and techniques are more for developers, and always in development Wilco: is it possible to have multiple pointers in the same result? Annika: needs research Hanno: With QUAIL we are implementing the outcomes of EARL: 'Not applicable' is used when the success criterion is not applicable for that page. In the example I only see passed or failed. Important to define the applicability Wilco: we plan to only allow Passed or Failed as that is that the new direction in WCAG-EM Annika: to give a pass, we have a problem with with selectors, because if you don't get a result at all if a selector doesn't find something, your selector doesnt match, so you won't get any result. Hanno: not having ' not applicable' is a radical change, not sure about the consequences. --- TEST FORMAT Wilco: I proposed a format to describe the tests. This could be the format: http://www.w3.org/TR/html5/tabular-data.html#table-model-error Wilco: is this a good format? Camiar: completely clear to implement, will be happy with the format Annika: this is a lot of writing, but makes sense. Hanno: this format is useful for correct implementations but is this also useful for detection when no semantic correct technique is implemented? For example when a table, or list items should have been used? Wilco: yes, then you should write down all situations. TEST for SC 1.4.1 Camiar: we should check one for one the use of color https://www.w3.org/community/auto-wcag/wiki/SC141-a#Step_1_.28tool.29 HL: 'position': for QUAIL we developed a 'unique selector script' for an HTML element, so that we can traceback the position of the code: You can find it here: https://github.com/quailjs/quail/issues/125 Other issues: Wilco: Hanno could you share script on specific tests implemented in QUAIL as an example? Hanno: will do for next meeting Next meeting Thursday 16:00 19/6 no meeting
Received on Tuesday, 10 June 2014 12:08:28 UTC