- From: Jonny Axelsson <jonny@metastasis.net>
- Date: Tue, 22 Feb 2000 13:59:10 +0100
- To: www-html@w3.org
EXPEL CODE, VAR, KBD, SAMP Don't get me wrong, these are clearly defined and very useful elements in their context. But they are legacy elements from the early days of HTML, where it was used to a large extent for coding. If there were to be elements with predefined meaning today, elements like ABSTRACT, BYLINE, CAPTION (for pictures), HYPERBOLE, PULLQUOTE... would make much more sense. Programmers belong to a small special interest group, catering to them especially in XHTML Basic would be counterproductive. I think it is safe to say that the vast majority of HTML producers will never use this module the way it was intended. Most will prudently ignore these elements, some will see these "free" elements (that they would never use normally) as experimental ones to play around with. There is a particular danger. If CODE is available, and TT is not, and either styles are not available or the designer is ignorant of them, some of the clever ignorant designers will use CODE to encode TT. Even if that is a minority of a minority, that minority will still outnumber the ones using CODE properly. PURITY OF ESSENCE In my opinion, the cleanest option is to strip off all semantic elements from the kernel of XHTML. Having elements with universal meaning were a lot easier when the Universe was the University. One approach would be to make special interest modules (that is XHTML elements with predefined meaning), science and programming, academia, press, office, manufacturing... That would be back to the pre-XML days, so leaving this to XML seems a more reasonable course. These elements should be in XHTML Basic Structural elements: body, head, html, title, div, span (these are basic both for documents and data) Fundamental text elements: h1-h6, p; also blockquote, q, abbr, em (only relevant for documents) Special: br and pre (these can also be defined in style sheets) These should be in XHTML 1.1, but not in XHTML Basic Interoperable text elements: b, i, sub, sup, tt? "Semantic" elements: address, cite, dfn (must pass test of universality and clarity) Legacy programming module: code, var, kbd and samp ADDRESS is losing out. The traditional 1994 web page is far less common, and ADDRESS really doesn't have the power and flexibility needed for its purpose. CITE may be ambigous. My association to "citation" is scientific citation (true to its roots), the examples I have seen tends to legal citations (similar, but different, still has "citation-ness" to it), I suspect the use of this element varies widely. Incidentally, why doesn't CITE have a CITE attribute? DFN may be the safest of these. Most producers would agree what a definition is. I think. These are my opinions based on my samples of documents. I don't know if there is any research on actual usage of the semantic tags (and HTML coding in general) based on a large and representative sample. A State of the Web report would be very useful.
Received on Tuesday, 22 February 2000 08:00:02 UTC