XHTML, XML, fancy text, and applications

Sun Apr 13 14:12:17 CEST 2003

This discussion is new on this list and sound confusing to me. It would
be nice to have a briefing on the background facts.

For example, I want to know where the *existing* and *used* transport
types application/xml text/xml fits into this, and why it shouldn't be
used for XHTML like for all other XML derivates. If the browser
recognize that this XML file belongs to a certain namespace and should
be rendered in a certain way (xhtml-wise) that is in my humble opinion
not a *transport* problem.

> I would see it as "This resource is of such-and-such type AND it is XML in
> terms of syntax."

= content type text/xml, and the first row of that XML-compliant file:

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
"http://www.w3.org/TR/2000/REC-xhtml1-20000126/DTD/xhtml1-transitional.dtd">

will tell that it is of such-and-such XML derivative type. This is thus
not a transport issue. The transport types are designed to be quite
general. A binary file has no standard for deciphering it's contents
(well with the exception of say Quicktime and such non-IETF de facto
standards) why it has to be carried in it's own transport type.

text/xml and application/xml are reserved for XML derivatives. These
additionally have clear syntax for describing their context i.e.
<!DOCTYPE ...> so they need no additional transport typing. To add it
anyway would be unnecessary duplication of information.

> I have
> to say that the pipe character ("|", vertical line, U+007C) is too devoid of
> well-known semantics.

This is a technical standard, and the pipe sign has been a part of the
Backuss-Naur form denoting alternatives since the 1960s. It is nowadays
usually used in the Extended Backus-Naur form for regular expressions.
The Backus-Naur was developed heavily inspired by Noam Chomskys language
hierarchy and formalisms for grammar introduced with his book "Syntactic
Structures" in the 1950s. Chomsky didn't have pipes, but both BNF and
EBNF has. This is the reason why it was in the DTDs, and why it is part
of most document description syntaxes. That is: it is not devoid of
well-known semantics.

This might be a totally off-topic reply in regards to this discussion,
but I could not accept that statemant as it stood.

> Right, and those should be subtypes of "text" if they can reasonably be
> treated as "text/plain", or subtypes of "application" if they contain
> non-textual markup.

application/xml and text/xml both already exist. What is wrong with
them?

-- 
Linus Walleij <triad at df.lth.se>