Re: On constraining/validating datatypes

In message <3.0.32.19970523185748.00c71da8@pop.intergate.bc.ca> Tim Bray writes:
[...]
> 
> One point - count me as one vote *against* wiring this into either
> XML-lang or WG8, and *for* doing it in a separate doc, using
> vanilla attribute/notation mechanisms, for now.  That way everybody
> gets to use it now, and people who don't care don't have to
> build the machinery into the parser... let's try and defend
> that grad-student-week, it's one of our proudest achievements. -Tim

I agree completely with this.  Don't forget that XML-lang *as at present*
is an extremely powerful tool that 99% of potential users have never seen
and it will take time to get used to.  Except for a few very minor tweaks as
are occasionally reported it works.  People are developing DTDs using it.

XML-TYPE is functionally completely independent of XML-lang and can be
managed by a post-parser module.  I may have missed it in the draft, but 
I'd suggest adding the following phrase in 3.3 or appropriate place:

"the use of attribute names beginning "XML-" in any way other that those
described in this specification and [reference to other XML specs] is
[some form of error]"

And to advise as an annexe that the following attribute names are currently
reserved:
	XML-TYPE
	XML-LINK
	XML-SPACE
	XML-STYLE

It might be appropriate to add something like "some of these attributes
may be processed and/or validated in dedicated modules"

For the SD5 debate, I was only asking for a convenient character that 
current parser technology could pass through unaltered.  If necessary
I can always use the colonic approach with a string like 'CML___...MOL' 
(work that one out :-)

For the XML-LINK discussions, I think it's important to decide whether the 
spec is implementable by a CS grad student *without previous HyTime 
experience* and purely from the spec.  I think it should be.  I also think 
those implementations should have a communality of function [I understand, 
but I'm worried by, simply defining the terms and leaving the rest up to 
the implementation.  Implementers need more guidance than that.]

	P.


-- 
Peter Murray-Rust, domestic net connection
Virtual School of Molecular Sciences
http://www.vsms.nottingham.ac.uk/

Received on Saturday, 24 May 1997 11:41:50 UTC