Re: Proposal: New Anchor attributes

Ian S. Graham (igraham@alchemy.chem.utoronto.ca)
Wed, 29 May 96 14:20:29 EDT


From: "Ian S. Graham" <igraham@alchemy.chem.utoronto.ca>
Message-Id: <199605291820.OAA18753@alchemy.chem.utoronto.ca>
Subject: Re: Proposal: New Anchor attributes 
To: connolly@w3.org (Daniel W. Connolly)
Date: Wed, 29 May 96 14:20:29 EDT
Cc: marc@pele.ckm.ucsf.edu, www-html@w3.org
In-Reply-To: <199605290346.XAA23152@anansi.w3.org>; from "Daniel W. Connolly" at May 28, 96 11:46 pm


Dan Connolly wrote:
> 
> In message <199605241546.IAA24083@pele.ckm.ucsf.edu.UCSF-LIBRARY>, Marc Salomon.......
> 
> >4.  Overloading ALT is problematic.  I had suggested that a convention of
........
> That appeals to me as a good idea in some ways, but I wonder if
........
> 
> In fact, I brought this up at a lunch discussion at the Danvers
> IETF with TimBL and some other folks at the table. Tim shot this
........
> 	http://www.w3.org/pub/WWW/MarkUp/Resource/Specification
.......
> 	<resource href="http://www.foo.com/foo.tar.gz">
> 		<meta name="content-length" content="1235234">
> 		<meta name="content-md5" content="23l4kj23l4kj23">
> 		<link rel=replica href="http://www.foo.com/foo.tar.gz">
> 		<link rel=replica href="http://www.bar.com/mirror/foo.tar.gz">
> 		<link rel=replica href="http://www.baz.com/net-stuff/foo1.9.tar.gz">
> 	</resource>
> 

This reminds me of the "pointer" link type proposed by Murray Maloney in 
his REL/REV draft of a few months back, and on some stuff I wrote up about 
document-querying forms (http://www.hprc.utoronto.ca/Ian/metaform.html).

In a real sense, 'resource' is specifying part of the interface from 
the document to elsewhere, but in a way that binds the information
to the location of the link. This is ok, but it would be nice to 
be able to reuse this information, and/or store it separate from the
document.

For example, suppose you used a head-level structure like

   <mlink rev="linkdata" id="idref" href="primary_url" >
     <meta name="content-length" content="1235234">
     <meta name="content-md5" content="23l4kj23l4kj23">
     <link rel=replica href="http://www.foo.com/foo.tar.gz">
     <link rel=replica href="http://www.bar.com/mirror/foo.tar.gz">
   </mlink>

(No prize for html here -- 'm' for 'multilink'....) then this interface
could be referenced by multiple anchors  using the "id" name:

     <a href="#idref"> ....</a>

For back compatibility, you could instead add an extra attribute to A, 
and code this as something like

     <a href="usual_url" relref="#idref">

In principle, you could then have a secondary document containing nothing 
but link relationships, and reference this through pointer  such as:

     <a href="url-to-catalog#idref"> ....</a>

This would simplify markup within the document and store the interface
details separately (in the head or in a separate document), which would
seem sensible.

Of course this is treading a bit on link catalogs, but perhaps the
time is ripe.

> Typed links are cool.
>
> Dan

I agree completely ;-)
 
Ian
--
Ian Graham ....................................... ian.graham@utoronto.ca
Information Commons                                University of Toronto