- From: Detlev Fischer <detlev.fischer@testkreis.de>
- Date: Wed, 16 Aug 2017 14:59:28 +0200 (CEST)
- To: public-wcag-act@w3.org, shadi@w3.org
- Cc: w3c-wai-gl@w3.org
Some comments on the aria-describedby rule example, mostly editorial but maybe with implications regarding the format. ---- Assumptions: It is not checked for references DF: "It is not checked" was hard to parse for me. "This test does not include a check ..."? ---- Assumptions: "If the aria-describedby attribute is provided, a long description was intended..." DF: aria-describedby may often be used to provide contextual help for controls, so you may want to reflect that possibility? ---- Step 1: Check if at least one of the aria-describedby attribute values is a valid identifier. DF: As a reader, I wondered wether this means that something like the non-exisitence of one of the values referenced (say, due to a misspelt ID) would be acceptable if the result passes later steps. I also wonder whether the question for "valid identifier" implies that it uses correct synax (e.g. does not have characters that would not be allowed) or if validity already means that there is a matching element ID. ---- Step 2: Check if at least one of the elements referenced by the valid aria-describedby attribute values exists. DF: again, if one of the referenced elements does not exist, would that not be flagged here? For authors, this would be relevant info? ---- Step 3: "Concatenate the results of Text Alternative Computation Algorithm run on the element itself and assign it to variable T1 and on all elements referenced by the aria-describedby attribute and assign it to variable T2." DF: I find this hard to parse. Would this not be eaier to understand as two bullet points? Like: * Concatenate the results of Text Alternative Computation Algorithm run on the element itself and assign it to variable T1 * Concatenate the results of Text Alternative Computation Algorithm run on all elements referenced by the aria-describedby attribute and assign it to variable T2 It is longer, but easier to grasp. ---- User Input Question: DF: I am not sure what "User input question" (singular?) is - what input? Is the answer my input? What follows is a list of property - value pairs, not a question. Therefore, the question at the end "If yes..." is difficult to understand. The one element it seems to map onto, 'question', is kind of buried in that list. Maybe its just that the question, form a user perspective, does not naturally appear as a property (which is technically may be) but a decision point, the core check of the entire rule. This is probably all quite obvious and natural to you but it puzzled me as a potential end user of this rule. ---- Outcome The selected element DF: I am not quite sure what "The selected element" refers to - I imagine it is the instance of aria-describedby under test. ---- step2-fail DF: Would there be a fail outcome if the referenced element contains links or other markup that cannot be communicated by AT? Put differwently, are there any formal constraints on what qualifies as valid? What about referenced content hidden by display:none? So much for now - best, Detlev -- Detlev Fischer testkreis c/o feld.wald.wiese Thedestr. 2, 22767 Hamburg Mobil +49 (0)157 57 57 57 45 Fax +49 (0)40 439 10 68-5 http://www.testkreis.de Beratung, Tests und Schulungen für barrierefreie Websites Shadi Abou-Zahra schrieb am 16.08.2017 11:22: > Hello, > > Minutes of the WCAG Accessibility Conformance Testing (ACT) Task Force > teleconference meeting on Monday 14 August 2017: > - https://www.w3.org/2017/08/14-wcag-act-minutes > > Next meeting: Monday 21 August 2017. > > Regards, > Shadi > > -- > Shadi Abou-Zahra - http://www.w3.org/People/shadi/ > W3C Web Accessibility Initiative (WAI) > >
Received on Wednesday, 16 August 2017 13:00:02 UTC