- From: Al Gilman <asgilman@access.digex.net>
- Date: Fri, 7 Nov 1997 11:38:58 -0500 (EST)
- To: w3c-wai-hc@w3.org (HC team)
to follow up on what Pawson, David said: > Surely these are attributes? They are defined by their use in headers and footers. They are attributes, but not SGML attributes. They are defined in the page context in CSS. They are not attributes in the document language model for HTML because HTML declines to define a page context model. For details consult the discussion in and around 12.3.6 Marking elements for the running headers & footers http://www.w3.org/TR/WD-CSS2/page.html#h-12.3.6 > All managed very well by standard SGML techniques. Creating > unique anchor namess from node references permits easy use of > cross references, but the overriding need is for document tree > traversal, something more in the realms of XML/DSSSL than > CSS. I think that a lot of functionality will wind up deferred to the XML generation of formats. But right now the HC job is to try to capture what the desired functionality is before we discuss it with the developers of CSS. Following that discussion, we may have to come back in the next generation to get certain "nice to have" upgrades. My position is that the scope of a "document" i.e. an HTML file is an artifact of the communications protocol and dominant medium. To come up with a universal WWW we need to base the content model on a continuous web of information which is segmented into HTML files much as print styles segment a long document into page units. The reason the Web is a winner is that the content is a better fit to a web than to a tree. As TimBL says, we don't want to lose that baby with the bathwater as we add formatting niceties. -- Al
Received on Friday, 7 November 1997 11:39:14 UTC