- From: Alexander Savenkov <w3@hotbox.ru>
- Date: Thu, 8 Aug 2002 19:30:50 +0400
- To: www-html@w3.org
Hello again, A few questions and proposals concerning attribute collections of XHTML 2.0, a few possibly editorial notes and a couple of proposals. 5.5. Attribute Types -------------------- Datatype: Datetime. If backwards compatibility isn't that important, I believe it's time to provide a link to Date and Time Formats Note [1]. Datatype: Coordinates. How come 'Length's are here? Was it meant that the 'coords' attribute can now accept percentage values? 7.3. The title element ---------------------- How come the 'title' element has all the common attributes now? Let's say Core collection and I18N collection are quite ok (i. e. the 'id' and 'xml:lang' for title could be used), but what's the purpose of 'href' attribute on a document's 'title'? I though 'title' is not a part of the flow, so is it really meant to be used as a link? As far as I know, with current CSS mechanisms it's not possible to set such a special value of 'display' property for the 'title' element, that's why it'd be interesting to know the means of making the 'title' a link, giving the focus to it, etc. 7.2. The head element --------------------- Anyway, the table at the beginning of the page assigns to the 'title' I18N collection only (which seems to be correct), while assigning Common collection to 'head' (which is denied later). Seems to be a pure editorial error (sorry for not sending it to www-html-editor as I'm not totally sure it's editorial). Any comments? 6.3. Hypertext Attribute Collection ----------------------------------- It's a brilliant idea to let most of the elements represent a link, though I don't quite get the idea of excluding the 'charset', the 'hreflang', the 'type', the 'rel', and the 'rev' attributes from the HyperText collection. What's wrong with them? If, say, a list item is a link, shouldn't it have a hreflang attribute if it points to a document in foreign language? 11.1. The bdo element --------------------- Is it correct that the 'bdo' element's only collection is Core? What were the reasons? The inability of a run of Hebrew text, for instance, to be a link is obscure and not having an 'xml:lang' attribute on it is at least strange. 12.1. The area element ---------------------- The WG needs to standardize the base for x and y coordinates for the 'coords' attribute not to confuse document authors and implementors (is the top left corner (0,0) or (1,1)?). Described in the HTML 4.01 errata [2]. 15.1. The meta element ---------------------- There's no 'scheme' attribute on 'meta' anymore. And if someone decides to specify a single 'meta' element which is not using the value of 'profile' on 'html' (still some other profile) then he or she has to rely on external resources. Pretty wise, one would say. Though seems illogical to me. If the WG decided to remove this functionality here it has to be replaced at some other point, or rather, at the specific point (LinkTypes datatype). 16. XHTML Object Module ----------------------- No examples are provided for how to specify a long description of an image (object). Since the 'img' element is removed, either the 'longdesc' attribute could be added to the 'object' element or the spec could contain some examples of how to provide a means to include a longer description, e. g.: <object data="http://www.w3.org/WAI/wcag1AAA.png" type="image/png"> <object data="http://www.w3.org/WAI/WCAG1AAA-Conformance" type="text/html"> Level Triple-A conformance icon, W3C-WAI Web Content Accessibility Guidelines 1.0 </object </object> The evident disadvantage of this method is that a user has no access to the long description page unless his or her agent is unable to display PNG images. What would the WAI fellows say? 20.1.3. Specifying external style sheets ---------------------------------------- Isn't it better to deprecate the use of the 'link' element with 'stylesheet' as one of the values of the 'rel' attribute in favor of xml-stylesheet pi? In any case a couple of sentences describing this mechanism cannot hurt. 21.3. The table element ----------------------- I'm not sure whether the table element needs the 'longdesc' attribute (or similar) or not. I suppose there were reasons for the introduction of the 'longdesc' attribute for images in HTML4 [3] to provide a *longer* description, that's why the lack of such attributes for often more complicated tables always worried me. No doubt having a separate (X)HTML file for the explanation improves accessibility. A large table could then be split into a series of smaller tables or an alternative graphical presentation could be provided which obviously cannot be achieved by means of a simple attribute. Links: [1] http://www.w3.org/TR/NOTE-datetime [2] http://www.w3.org/MarkUp/html4-updates/errata [3] http://www.w3.org/TR/1999/REC-html401-19991224 Best regards, --- Alexander "Croll" Savenkov http://www.thecroll.com/ w3@hotbox.ru http://croll.da.ru/
Received on Thursday, 8 August 2002 11:34:07 UTC