Re: The furore over PUBLIC

Peter Flynn wrote:
> I admire your optimism. The implementors who have taken the time and
> trouble to read the spec and its underlying assumptions (8879 etc)
> will certainly do this, but others may not, and if they hit lucky,
> they can arbitrarily define behaviour regardless of the spec hooks.

In a world with no communication I would expect this to happen. In a
world where Jean Paoli talks to other people at Microsoft, Dan Connolly
talks to the Amaya developers, the first XML authoring tools are based
on SGML tools and Netscape is already going to be a year behind in XML
implementation, I suspect that some one will find the right hook and
demonstrate its proper use. 

But if we think that vendors are that stupid we could be quite explicit:

"This location is reserved for the specification of the public text name
of this object. The public identifier is a string which uniquely
identifiers a particular document in a manner that is not specified by
an address tied to an IP address or set of IP addresses. A particular
mechanism for resolving these is not currently defined."

Anyone reading that who doesn't think: "I could use this for URNs when
the specifciation for those stabilizes" isn't smart enough to be a

 Paul Prescod