Re: a personal conclusion.....

At 04:54 PM 6/2/00 -0400, Michael Mealling wrote:
>Rev XML to 2.0, make it axiomatic that anything that remotely looks like a 
>name or an address be an "absoluteURI" according to the 2396 ABNF. Don't
make 
>it required that it be resolved in order to be parsed but make it so that 
>NOTHING (sans things like not being connected to the network, not having
paid 
>for it, etc) can make it so that the URI can't be resolved if something
above 
>it so desires.

That'll require a revision to 2396, I think, as XML documents may come from
situations where there is no base URI whatsoever.  Maybe XML 2.0 could
define a canonical base for such cases without revising 2396, but I'm not
(personally) sure the axioms hold up or make sense in the context of
document processing anyway.

I'm also not sure XML 2.0 is worth considering until the world's had more
time with 1.0, and I don't like the solution presented above, but I
certainly hope that these issues will get a full hearing in any further
development of XML.

(If XML 2.0 is any more complicated than XML 1.0, and I expect it will be,
don't be surprised if it gets the same treatment SGML got from XML, and
subsets rule.)

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 Friday, 2 June 2000 17:11:43 UTC