- From: Henri Sivonen <hsivonen@iki.fi>
- Date: Mon, 12 Feb 2007 13:19:16 +0200
On Feb 12, 2007, at 12:57, Benjamin Hawkes-Lewis wrote: > W3C explicitly doesn't specify how UAs should treat XHTML1 served > as text/html http://lists.w3.org/Archives/Public/www-html/2000Sep/0024.html > W3C says all XHTML1 documents must include a DOCTYPE declaration; The HTML WG of the time seemed to think that validity in the XML 1.0 sense is somehow very important and needed to make conformance tractable. (One could argue, that doctypes in XML actually make conformance less tractable.) The doctype is mostly required as a talisman (to use HTML5 terminology). > But UAs could certainly DOCTYPE sniff in this way. The right way to go about this is to invent new names for the incompatible elements/attributes or to abstain from specifying incompatible stuff in the first place. > (For an argument against the desirability of doing so, see http:// > hsivonen.iki.fi/doctype/, but I still can't see there's > any practical alternative when elements with different semantics > share the same namespace.) The practical alternative is to refuse to implement specs that require legacy-incompatible changes to processing models. > Better question: if XHTML2 ends up using the same namespace as > XHTML1, will UAs be able to treat XHTML2 documents in a way that > conforms to both of W3C's own XHTML1 and XHTML2 specs? That's indeed a very good question. > (I think yes, but again only with DOCTYPE sniffing.) "Only with doctype sniffing" should be read as "No". -- Henri Sivonen hsivonen at iki.fi http://hsivonen.iki.fi/
Received on Monday, 12 February 2007 03:19:16 UTC