FPIs as locators in XML-links ?

Tim Bray tbray at textuality.com
Fri Aug 8 19:18:37 BST 1997


At 01:08 PM 08/08/97 -0400, Paul Prescod wrote:
>> Two reasons, really.  XML-Link is designed specifically for use in
>> the context of the Web, and on the the Web, things exist if they
>> can be addressed by URI's, otherwise not.
>> 
>> Secondly, whereas PUBLIC identifiers are very interesting and useful,
>> it is not the case that virtually every server and desktop in the
>> world comes with excellent free machinery to use them across the network,
>> which is the case with URLs.  -Tim
>
>1. Don't these arguments apply equally to XML-Lang? 

No.  Links on the Web are based on URI's; that's a fact of life.  If 
you want another kind of link that isn't, go ahead and build it, but
our mandate was to build a Web-oriented hyperlinking facility.  There is
no Web machinery that knows anything about FPI's.

>2. You've argued why PUBLIC identifiers will sometimes not be useful.
>You haven't argued why they will *never* be useful. They were put into
>XML Lang because some argued that they will sometimes need them. That
>applies to XML-Link equally.

Yes, you've said this many times.  So far, the WG membership is
unconvinced.

>3. What about entities declared through system identifiers? Why can't I
>link to them through their entity names?

Because that's not how things are done on the Web.  Of course, you in
XML you *can* say 
<a href="http://&host;/&path;/&base-file;#&xpointer;">

> What is the point of "binary"
>entities if the "standard" linking and transclusion mechanism can't use
>them? Or to go the other way, why wouldn't the standard linking and
>transclusion mechanism be able to use the standard mechanism for mapping
>external resources into document names?

The key point is the use of the word "standard".  The use of entities
and PUBLIC identifiers is standard only in the world of SGML.  For
interoperation with the universe of Web documents, the only standard
way to do things is the URI mechanism.

To summarize, we were not trying to extend the SGML entity mechanism
to do network hypertext; we were trying to extend the existing Web
hypertext mechanism to be usable in XML.

Anyhow, this argument is over.  Sorry. -T.

xml-dev: A list for W3C XML Developers
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To unsubscribe, send to majordomo at ic.ac.uk the following message;
unsubscribe xml-dev
List coordinator, Henry Rzepa (rzepa at ic.ac.uk)




More information about the Xml-dev mailing list