do we dig under the fence with pseudo-public ids...

[Paul Grosso <pgrosso@arbortext.com>]

...and if we do, is that a good thing or not?

At 10:03 2001 05 07 -0700, Lauren Wood wrote:
>4) issue 21: I don't think the email we've seen covers the issues related to 
>issue 21; Norm and John had the action item to send email about what the 
>catalog resolver should actually do when handed a urn:publicid system 
>identifier.  - This issue needs more discussion in email.

I really am not sure what I think yet, but I'd like to hear
some discussion which gets into what users really need to do.

What are the practical differences here?  Which user scenarios
work/don't work under the two sides of the debate?  And what will
T.C. MITS think of all this?

I'm somewhat intrigued by slipping a public id into the system id, 
but my intuition tells me this is going to be more confusing for 
people and is going to get us into more trouble with our prefer 
system/public algorithms.  My intuition still comes down on the side 
of wanting to be clear that public ids are only those things that
are clearly public ids per the XML spec, and that anything else is
going to get us into a bigger mess.  

We can delegate system ids, we can delegate URIs, what is the advantage
of having two completely different ways to provide "-//Acme//DTD Book//EN" 
as the left hand side of a catalog lookup process?

paul