- From: David MacDonald <david100@sympatico.ca>
- Date: Mon, 13 Feb 2012 12:10:55 -0500
- To: "'James Nurthen'" <james.nurthen@oracle.com>, "'W3C WAI Protocols & Formats'" <w3c-wai-pf@w3.org>
- CC: "'WCAG'" <w3c-wai-gl@w3.org>, <wai-xtech@w3.org>, "'Cynthia Shelly'" <cyns@microsoft.com>
- Message-ID: <BLU0-SMTP84957722F6FD5D7298AC5DFE7F0@phx.gbl>
Aria landmark has been completed in response to the comments. Sorry I had sent them to the WCAG editors last week... http://www.w3.org/WAI/GL/wiki/Using_ARIA_landmarks comments found here: https://www.w3.org/2002/09/wbs/35422/20120126misc/results Detlev Fischer Accept this with the following changes Two issues: (1) Is this technique earmarked as sufficient technique for 2.4.1? There are no user agent and assistive technology support notes that would detail the level of support across the istalled base. Should the current recommendation be to use landmarks as additional technique to meet SC 2.4.1, but not as sole means? That might make it an advisory technique at the time being. (2) I remember some tests have shown problems with redundant use of ARIA landmarks and structural elements in HTML5 (some screen readers announcing role twice in particular UA/AT combos). Should this be covered? Not sure what the best practice would be right now. Added support notes Reject comment #2... not worried about redundancy for such a short phrase and only 5 or 6 elements on a page... More important to have structure Christophe Strobbe Accept this with the following changes Regarding applicability: Perhaps we should scope the applicability of the technique to technologies where these landmarks have accessibility support. (As far as I know, you can also use WAI-ARIA in SVG Tiny 1.2 <http://www.w3.org/TR/SVGTiny12/struct.html#RoleAttribute> but I doubt that there is any accessibility support.) Regarding the HTML5 examples: Section 7.5 of the WAI-ARIA spec says: "WAI-ARIA roles, states, and properties are intended to add semantic information when native host language elements with these semantics are not available, and are generally used on elements that have no native semantics of their own." <http://www.w3.org/TR/wai-aria/complete.html#host_general_conflict> If duplicating native HTML5 semantics with WAI-ARIA roles leads to duplicate announcements in AT, wouldn't it be better to take out e.g. the nav example and add a warning about duplicating semantics? Test procedure: The above would also affect the test procedure. We could add for example: 3. Check that the landmark attribute does not duplicate the element's native semantics. (If duplicate announcements in AT are considered an AT bug, this point may be moot, but I think that duplication of role information just leads to maintenance overhead. As software developers say: DRY: "Don't repeat yourself".) Changed applicability...Limited it to HTML support... -Regarding the HTML5 question... I think that should be extended to a wider question. Do we want to separate out HTML5... Currently the semantics of HTML5 are not exposed to most accessibility API's and so these Landmarks are necessary. The only redundant one in those examples currently would be Firefox 10. Also, not too worried about a redundancy of one word from an accommodation perspective. David Todd Accept this with the following changes Should an author be required to uniquely label multiple landmarks of the same type? Regarding the HTML5 examples, don't the semantics of the elements themselves communicate the necessary accessibility information? I don't think unique labels are an issue with landmarks, except <DIV>s of course have to have unique ids... which is already covered. Regarding HTML5, see above responses Frederick Boland Accept this with the following changes need to document degree of widespread support for implementing this technique as well as any known issues Mostly done... feel free to add anything else known... Michael Cooper Accept this with the following changes Pretty good start. The description seems to assume you know about the role attribute. Michael Cooper Accept this with the following changes Pretty good start. The description seems to assume you know about the role attribute. I suggest a paragraph saying that landmarks are put in with the role attribute on the element, whose value is the name of the landmark. Also in the list the landmark names start with a cap, but the role value has to be all lower case, so they should be shown that way in the list. Also in the list the landmark names start with a cap, but the role value has to be all lower case, so they should be shown that way in the list. Done! Marc Johlic Accept this with the following changes Agree with comments from Christophe, David, and Michael. Also, consider adding a statement around using the "region" landmark to avoid orphaned content that doesn't fit properly into the others listed. Check Resources link - should be: http://www.w3.org/TR/wai-aria-practices/ Done! Loretta Guarino Reid Accept this with the following changes This is an interesting draft, and may make a good proposal for the Task Force to start with. I think the test procedure is a little vague, and we will need to decide how much of the spec gets repeated, or how we reference parts of the spec. I would also like to see the description focus less on selling how easy this is and more on telling someone who is not familiar with ARIA what this technique is, that is, how to use the markup. I agree that we have to make a decision re: HTML5. A compelling reason to leave HTML5 in this technique is the issue that HTML5 sections are not supported well, but ARIA is, which is in fact it's purpose, to bridge unsupported features of technology. I took out the sales line about how easy it is. Tried to simplify and be more specific in the test. Tried to be further explain what a Landmark is to a lay person. Cheers David MacDonald ... access empowers ... ... barriers disable ... www.eramp.com Cheers David MacDonald ... access empowers ... ... barriers disable ... www.eramp.com -----Original Message----- From: James Nurthen [mailto:james.nurthen@oracle.com] Sent: February-11-12 11:15 PM To: W3C WAI Protocols & Formats Cc: WCAG; wai-xtech@w3.org; Cynthia Shelly Subject: [html-techs] Feb 13 2012 meeting cancelled. The meeting scheduled for Feb 13 has been cancelled. As we don't have any techniques to review it would be better to spend the meeting time working on writing technique(s) so we have something to review next week. Sorry for the late notice but I was hoping we would have something to review this week. Regards, James
Received on Monday, 13 February 2012 17:11:36 UTC