Re: Namespaces aside, absolutizing is none of _X_Path's business

At 02:43 PM 6/4/00 -0400, John Cowan wrote:
>This seems a strained distinction to me, given that the base URI is an
>infoset property, and that RFC 2396 resolution is a purely syntactic
>operation.

Maybe this is wacky, but maybe the Infoset needs to include (and define how
to deal with) base URIs for namespaces and non-namespace contexts on a
per-element basis.

This would permit 'higher levels' to perform absolutization and linking
reliably and (I hope) reasonably, without having to worry about the various
difficulties (inclusions, DTD defaulting, etc.) we've identified with
absolutization.  

This isn't pretty, and it would have to be _optional_ information in the
Infoset.  The DOM (and maybe SAX, version 2.1 or 3) would need to provide
such information.

It'll take time getting this out in the world, but it might clean up this
mess.

On the other hand, I just spent my afternoon at the San Francisco Maritime
Museum (Hyde Street pier), and my mind may still be on board ship.

Simon St.Laurent
XML Elements of Style / XML: A Primer, 2nd Ed.
Building XML Applications
Inside XML DTDs: Scientific and Technical
Cookies / Sharing Bandwidth
http://www.simonstl.com

Received on Sunday, 4 June 2000 17:30:32 UTC