Re: Personal view

On Sun, Jun 18, 2000 at 12:58:52PM +0100, David Carlisle wrote:
> On the other hand I do accept that the other half of your proposal
> 
> >    the 'base' for
> >    a namespace declaration is the pseudo-scheme 'xmlns-base:/'.
> >    In no case is the base used for namespace absolutization to be
> >    taken from the containing document.
> 
> is a viable alternative to the  "literal" and "forbid" options (which
> are the other two viable proposals on the table). This is the
> "fixed base" option with another different specific proposal
> for the base.


Even with "forbid" one should be able to optionally associate a retrieval
URI Reference for namespace (this one being possibly relative)
independantly of the definition of the namespace. This addition doesn't
have to go immediately in a revision of the namespace specification but
could be added later.
It allows to do a design where:
   - the namespace name is always absolute, so that no document will
     ever break at the XML processor level when moving them, and allowing
     cheap and intuitive testing of namespace values.
   - the metadata associated to the namespace could be packaged relative
     to the base document, could be defined like a:xmlhref="../a", and
     allowed only on the same elements as xmlns:a="..."

 I have seen this proposed a couple of time, and it could be the easiest
way to allow retrieval from the base of the resource when needed. I
can't remember if it was suggested for Schemas, I would rather like to
see it in the Namespace specification itself. It seems to me it has the
following good features:
   - doesn't break in principle xml-name 1.0 since there is no reference
     in the REC about dereferencing. 
   - doesn't need to change radically deployed code, unless that code
     tries to dereference namespaces name values where it could break
     in some cases.
   - it would still be possible to reuse "common" namespaces like XHTML
     or SVG, and their associated metadata in tools expected to be
     "embedded" or ran without connectivity.

 Possible drawback, is how it would affect Schemas processors. Maybe
a cleaner packaging mechanism would be a better place to put efforts
since it would solve the problem in a more generic way.

Daniel
 
-- 
Daniel.Veillard@w3.org | W3C, INRIA Rhone-Alpes  | Today's Bookmarks :
Tel : +33 476 615 257  | 655, avenue de l'Europe | Linux XML libxml WWW
Fax : +33 476 615 207  | 38330 Montbonnot FRANCE | Gnome rpm2html rpmfind
 http://www.w3.org/People/all#veillard%40w3.org  | RPM badminton Kaffe

Received on Sunday, 18 June 2000 12:07:16 UTC