[microformats-discuss] Re: XML & microformats (WAS: Take 3)

Lucas Gonze lucas.gonze at gmail.com
Tue Sep 20 15:20:31 PDT 2005

> On 9/18/05, Lucas Gonze <lucas.gonze at gmail.com> wrote:
> > For example, if you embed an HTML link element in another format
> > without enclosing it in a head element enclosed in an html element, it
> > may still make intuitive sense but it doesn't make formal sense.  Or,
> > there is the issue of losing the DTD of the HTML version your link
> > element is based on.

On 9/18/05, Robert Sayre <sayrer at gmail.com> wrote:
> The W3C has already done this. It's called "XHTML Modularization".
> http://www.w3.org/TR/xhtml-modularization/
> There is a discrete "Link Module"
> http://www.w3.org/TR/xhtml-modularization/dtd_module_defs.html#a_module_Link

Thanks, Robert.  This is a valuable and informative reference.

However, I didn't say that it was impossible to include portions of
other document types, I said that it would be a mess, and I don't
believe that this reference makes that untrue.  The XHTML
modularization is a thick stack of W3C arcana which will require
further work before it becomes useful.

I'll show what I mean by giving examples of arcana to be deciphered.

http://www.w3.org/TR/xhtml-modularization/conformance.html implies
that a document type which hosts XHTML modules must be validated with
DTD and no other method: "The document type must be defined using one
of the implementation methods defined by the W3C. Currently this is
limited to XML DTDs, but XML Schema will be available soon."  Since
this list is open ended, how do I find out what it includes?  And how
about host document types like RSS 2.0 and XSPF which lack a canonical
definition using DTD, XML Schema, etc?  Does Atom have a definition
which is both formal and canonical?  (The existence of a formal
validation method isn't enough to satisfy this requirement -- that
validation method must be canonical, as part of the spec.)

There are new rules for user agents which many user agents will have
to be updated for.  How many such user agents are there?  How
extensive are the changes?

There are new rules for host document types that will also require
update to the host document types, for example "XHTML Host Language
document types must adhere to strict naming conventions.  [...]   The
names for document types implemented as XML document type definitions
are defined through Formal Public Identifiers (FPIs)."  Does that mean
that every host document type has to have a DTD?  Are there host
document types which have a DTD but don't use FPIs?

Compiling a complete list of questions, then finding out and
documenting the answers is work still to be done.  I imagine this work
would be done by the usual kind of interest group, which would require
finding people with the interest, time and skills, and waiting for
them to complete the project.

This is doable, but until it is done advice to use XHTML microformats
to extend other document types is hand waving.

More information about the microformats-discuss mailing list