W3C home > Mailing lists > Public > www-tag@w3.org > October 2002

Re: Enhance XLink: infer xlink:type

From: Ann Navarro <ann@webgeek.com>
Date: Fri, 04 Oct 2002 17:46:38 -0400
Message-Id: <4.3.2.7.2.20021004173823.01f6cd20@geekserv1>
To: Tim Bray <tbray@textuality.com>, WWW-Tag <www-tag@w3.org>

At 08:46 AM 10/4/2002 -0700, Tim Bray wrote:

>Obviously people feel XLink has too much syntax, and I'm pretty well 
>convinced that they're right.  How about the following:
>
>Make xlink:type inferrable from xlink:href as follows:
>
>If an element has xlink:href=, then
>- If its parent element has
>   xlink:type="extended", infer xlink:type="locator"
>- Otherwise infer xlink:type="simple"
>
>Hey-presto, minimal simple xlinks are now
><foo xlink:href="bar">foo</foo> - no defaulting required!  Of course you 
>could still put in xlink:title and other good stuff if you wanted.

This is good, provided you mean this would be addressed as a formal part 
of  XLink 1.1 (or 2.0, or whatever). Currently the XLink specs suggests 
that defaulting is available, however it has to defaulted in the DTD. 
Requiring UAs to fetch and process that DTD to manage a document written in 
the minimalized form would be a huge cost. Therefore, improvements to 
defaulting would need to be in the XLink Recs that processors are based upon.

Some other threads have brought up the idea of "generic XML", and I'm a bit 
surprised to hear that people don't have any idea of what that may be.

"Generic XML", or more specifically as "generic XML processor" to me is 
something that is:

1. Namespace aware
2. XLink aware
3. can process or cooperate with the processing of style information (XSLT, 
CSS, or both)

Given these constraints, "generic" XML is something that could be tossed at 
such a processor and be displayed according to the style information provided.

A requirement of landing at the end of our fabled bridge between HTML and 
XML, is the ability of an XHTML document to be processed by such a generic 
XML processor. That means that no arcane knowledge of special linking 
semantics (as is required today with <a>, etc), or other special 
button-holing can occur.

IF XLink were to be updated to incorporate the defaulting described above, 
we'd be very far along the path toward linking in XHTML being both simple 
and generic.

It should be noted that nothing in the quest to be 'generic XML' should 
result in an inference that browsers such as IE or Netscape may choose to 
do things in a less-than-generic fashion. However, any 'generic XML 
processor' should be able to handle a well-formed XHTML document as readily 
as it would handle any other well-formed XML document that has linking and 
is styled with XSLT/CSS.

Ann



-----
Ann Navarro, WebGeek, Inc.
http://www.webgeek.com

say what? http://www.snorf.net/blog
Received on Friday, 4 October 2002 17:48:32 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:12 GMT