W3C home > Mailing lists > Public > www-html-editor@w3.org > April to June 1999

Re: Liaison statement on fragment identifiers from Linking WG

From: Shane P. McCarron <ahby@themacs.com>
Date: Thu, 13 May 1999 14:50:29 -0500
Message-ID: <373B2D05.3BCB5EF4@themacs.com>
To: Bill Smith <bill.smith@sun.com>
CC: Steven Pemberton <Steven.Pemberton@cwi.nl>, w3c-xml-cg@w3.org, w3c-html-wg@w3.org, www-html-editor@w3.org, w3c-xml-linking-wg@w3.org
As in your original message, my comments are personal and may not be
shared by the rest of the working group.

The HTML working group continues to discuss the general issue of making
XHTML a better XML application, in particular in the areas of fragment
identification/extraction. We fully recognize that HTML is the lingua
franca of the Internet, and that an XML version of HTML can only help to
improve the general XML-development environment.

At issue here is the working group's (I believe legitimate) concern
about our responsibility to the existing community of HTML writers and
their wealth of existing HTML documents. The working group, in its first
deliverable, is attempting to provide a migration strategy that helps to
transition those documents from ad hoc, quasi SGML to well formed, valid
XML. In the long run, fixing these documents and teaching the users how
to write such documents is good for the entire XML-using community, and
great for the Internet.

Admittedly, HTML 4.0 has many semantics and commonly used syntactic
conventions that make it not XML. However, the working group has
determined that it is relatively easy to develop documents that are XML
conforming and backward compatible.  This is the scope of XHTML 1.0. You
may argue about whether such documents will ever be sensibly processed
as media type text/xml in generic XML user agents.  However, such
processing was never our goal in this first Recommendation.

Going forward, we fully intend to effectively force the issue. In the
XHTML Modularization specification, the working group is developing a
collection of abstract modules that will be used in the development of
future document types. While some of these modules still contain many
elements and attributes that have semantics inexpressible in XML/XSL, we
are attempting to isolate those aspects of XHTML so that document
developers wanting to have maximally portable XML documents can do so
while remaining within the XHTML family.

Specifically, the working group has decided to remove the name attribute
from the "a" element. The working group has also recently discussed
that, in a future iteration of XHTML, it should fundamentally change the
way it recommends defining fragment identifiers. This change would
effectively eliminate the use of historical anchors in favor of elements
that wrap their content and have identifiers. We believe that this is
consistent with the goals of your working group.

Regarding your specific concern about the "name" attribute, I think it
is important to remember that the XHTML 1.0 document is ONLY targeted at
backward compatibility. We are definitely not advocating that the name
attribute on the "a" element be processed as an ID in future user
agents. However, I can see from a reading of the specification that a
new user agent developer might determine that the best way forward is to
develop an agent that recognizes "id" and, if that is not present, then
looks for "name". I personally agree that this would be a disservice,
and could set a trend that would end up reserving the "name" attribute
in certain contexts.

Perhaps a better solution is one that has different conformance
requirements on user agents depending upon the document's media type:

If the resource is served as text/html and claims to be an XHTML
document, examine XML-preferred attributes but permit fallback to
historical equivalents (id/name, xml:lang/lang)

If the resource is served as text/xml and claims to be an XHTML
document, only examine XML-preferred attributes.  If necessary, we could
even accomplish this through a different DTD - although I think that
might confuse content developers.  Rather, we would set a conformance
requirement that certain attributes are to be ignored in conforming
XHTML user agents when processing text/xml streams.

Would something like this help to alleviate your concerns while still
helping us with our backward-compatibility goals?
--
Shane P. McCarron                              phone: +1 612 434-4431
Testing Research Manager                         fax: +1 612 434-4318
                                              mobile: +1 612 799-6942
                                              e-mail: shane@themacs.com

OSF/1, Motif, UNIX and the "X" device are registered trademarks in
the US and other countries, and IT DialTone and The Open Group are
trademarks of The Open Group.
Received on Thursday, 13 May 1999 15:50:42 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:16:44 GMT