- From: Henri Sivonen <hsivonen@iki.fi>
- Date: Tue, 14 Oct 2003 07:51:41 +0300
- To: David Woolley <david@djwhome.demon.co.uk>
On Tuesday, Oct 14, 2003, at 00:09 Europe/Helsinki, David Woolley wrote: >> XML processors to be validating parsers, we cannot rely on information >> within the DTD for selector matching, and in order to have >> interoperability, we must therefore require that all UAs _ignore_ such >> information. Otherwise stylesheets could result in radically different > > It seems to me that you are missing a third category of user agent, > which > is actually the dominant class, namely those with specific application > knowledge about the language in use, but which do not read this from > a DTD, e.g. the typical HTML web browser. Typical HTML browsers use a tag soup processor--not an XML processor. > If, for example, one denies the existence of entities defined for HTML > (not the specific issue here) you violate the "least astonishment" > principle. Either the full set of HTML entities should have been included as predefined in the XML spec or dropped out of the XHTML DTDs. The astonishment isn't caused by browsers but follows from the specs. OTOH, if you wished to avoid the astonishment by having the full set of HTML entities defined without reading the DTD, the parser wouldn't be a conforming XML processor and it would be inappropriate to use such a parser to parse an XML content type. -- Henri Sivonen hsivonen@iki.fi http://www.iki.fi/hsivonen/
Received on Tuesday, 14 October 2003 00:54:21 UTC