- From: Jason White <jasonw@ariel.ucs.unimelb.edu.au>
- Date: Tue, 6 Jul 2004 16:37:07 +1000
- To: Web Content Guidelines <w3c-wai-gl@w3.org>
John M Slatin writes: > > > > From: w3c-wai-gl-request@w3.org [mailto:w3c-wai-gl-request@w3.org] On > Behalf Of Yvette P. Hoitink > > [requiring noun phrases instead of verb phrases in link text] > > I don't think we can require noun phrases instead of verb phrases. > Sometimes the action is what makes the distinction between two links. > For example, consider these links: > > "View item A" > "Edit item A" > "Delete item A" > > If you ban verb phrases, what should the noun phrases be for these > links? Constructs like "deletion of item A" only make the text more > complex and _less_ accessible. > > > John: > Great point! Thanks, Yvette. which takes us back to my comment at last week's meeting that no one has refined this into a testable proposal. I am also concerned that the emphasis is on links, not on user interface options more generally. Clearly, whatever the success criteria are, they apply to all sorts of user interface components, not just links. My underlying concern here is that the more the guidelines are written with HTML implicitly in mind, the less applicable they will be to other technologies, exactly the problem that prompted the development of WCAG 2.0 in the first place. Thus I suggest stepping back and considering what the desirable outcome is, in this case, perhaps that text characterizing the function to be performed, or the information to be retrieved, by every user interface action is programmatically associated with the user interface control. This would apply to all formats where there is a direct correlation between the content itself and the kind of interface that te user experiences, i.e., to those situations in which the author is designing the user interface. Then there are non-normative suggestions such as: The text associated with the user interface control should be as brief as possible. It should be possible to understand when read out of context. I don't think these last two are testable. Does a proposal along these lines seem more appropriate? Are there other alternatives?
Received on Tuesday, 6 July 2004 02:37:12 UTC