RE: TB16 Re: Comments on arch doc draft

Catalogs work fine.  We do understand 
how to operate in a disconnected mode, how 
to disseminate to a data warehouse for 
complete separation, and that some kinds 
of information simply don't get URLs ever. 
One can use XML without namespaces. Or 
we'll use SGML and that will improve 
other aspects.

This one comes down to it *not* being the 
responsbility of the TAG or the W3C to 
write the rules for SHOULDN'T.  That 
has been done by other agencies before 
and will be done by other agencies now.  
Those that don't will have the problems 
of the California utilities.  How 
was it Steve Newcomb put it:  "Don't 
put on the web anything you don't want 
to see in the New York Times."

Good advice.

len

-----Original Message-----
From: Paul Prescod [mailto:paul@prescod.net]

"Bullard, Claude L (Len)" wrote:
> 
> It worked well enough for markup until
> namespaces were introduced.  

XML has many references to web architecture embedded in it. Most are
much more inconvenient for offline use than namespaces. Nobody seems to
mind. We SGML-ers swore that requiring system identifiers would make XML
unusable off the Web. We were wrong.

> ... Maybe the
> problem is namespaces and then perhaps
> the only problem is insisting that to
> be a namespace it must be dereferenceable.

There is essentially no circumstance where it is difficult to make
namespaces dereferencable and if there were such a circumstance it would
be handled by the distinction between SHOULD and MUST. There is a reason
that there is a distinction.

 Paul Prescod

Received on Tuesday, 2 July 2002 18:53:46 UTC