[sc34wg3] CXTM CD posted

[Kal Ahmed]

On Tue, 2003-12-30 at 15:02, Lars Marius Garshol wrote:
> * Kal Ahmed
> | 
> | I have updated the CXTM specification in accordance with comments
> | received both on this list and at the meeting in Philadelphia. 
> 
> Actually, there's one thing you appear to have overlooked, and that's
> the comment I made on section 4.10 in this email:
> 
> <URL: http://isotopicmaps.org/pipermail/sc34wg3/2003-December/002079.html >
> 
> 
> I see that source locators are not canonicalized, but we still have a
> problem with URIs that were relative in the input file and also with
> subject identifiers used for reification. We can solve the latter by
> modifying 4.4 (and probably should...), but the former isn't solvable,
> and I think it should be possible to also test files that contain
> relative URIs.
> 

I agree that it certainly should be possible to canonicalise files
containing relative URIs. The question is whether to require
relativisation of the URIs or to resolve the URIs against some
externally specified base (so the input to the process then becomes a
TMDM instance and a base URI). I think I actually prefer the latter
rather than the former.

> This was just a quick glance at the one crucial issue I thought had to
> be solved. There may well be more, but I'll leave those for when I
> have time to look more carefully at the spec.
> 

Sorry about that - I think it must have got lost from the list of issues
that I took to the Philadelphia meeting.

> BTW: An informative annex with a RELAX-NG grammar for CXTM in compact
> syntax would be nice...

http://www.isotopicmaps.org/sam/cxtm/cxtm.rnc

I guess it should be an annex too.

OK I'm off to write two points down somewhere where they won't get
forgotten.

Cheers,

Kal
-- 
Kal Ahmed, Techquila
Standards-based Information Management
e: kal@techquila.com
w: www.techquila.com
p: +44 7968 529531