Re: XHTML with Internet Explorer

"Jim Ley" <jim@jibbering.com> writes:

> What is the compatibility subset?   Could you please document it - remember
> Appendix C is informative and 5.1 and RFC 2854 don't make it any more than
> that we're repeatedly told.

I look at it this way: RFC 2854 and the various specifications for
HTML <= 4.01 notwithstanding, "text/html" has been declared de facto
the arena for "tag soup".  The idea is that an XHTML 1 document is
likely to do well when construed as tag soup if the rules of the
non-normative Appendix C are followed.  This appears to have been
based simply on observed behavior of historic tag soup user agents.
It does indeed appear to be true that some historic tag soup agents
were never given the capability of dealing with XHTML correctly
(although some of those provide visual handling of XHTML based on CSS
when such a document is served as general XML).  There are also
historic user agents capable of handling XHTML documents that will not
process them correctly if they are served as "text/html"; for these
the reason seems to be based not on obstinacy but on processing
design.

I suspect that Connolly and Masinter, the authors of RFC 2854,
intended to leave room for _possible_ future forms of XHTML under the
"text/html" umbrella.  Note that RFC 2854 assigns future responsibility
for the definition of the content type "text/html" to W3C.

Another wrinkle: for XHTML documents that include MathML, e.g.,
XHTML 1.1 + MathML 2.0, because of the existence of XML+CSS+XSLT
capable user agents that are not XHTML agents the use of
"application/xhtml+xml" is currently deprecated by the W3C Math group
in deference to "text/xml".  The trick here is to use XSLT to feed
different user agents according to the idiosyncracies of each.  See
http://www.w3.org/Math/XSL/ .

It's not a perfect world, and there's more than one imperfect player
in it.

                                    -- Bill

Received on Wednesday, 5 November 2003 12:20:50 UTC