Re: Resource Gloss (Human Readable)

[Paul Grosso <pgrosso@arbortext.com>]

At 11:58 2001 01 05 -0500, Charles Reitzel wrote:
>At 10:07 AM 1/5/01 -0500, John Cowan wrote:
>>Charles Reitzel wrote:
>>
>> > My understanding is that John Cowan's XML Catalog 
>> > was done as part of this OASIS effort.
>>
>>Actually, XML Catalogs came first, but I am part of the OASIS TC.
>
>Sorry, I should have said "has been incorporated into".  Has it?
>
>> > The difference, I think, is that it 
>> > treats human readable readable documents 
>> > as "just another associated resource".
>>
>>The catalogs that OASIS is developing, like 
>>their SGML predecessors, are meant to leverage 
>>system and public identifiers in XML entity
>>and notation declarations, not random URIs 
>>(even namespace URIs) in XML content.

Correct, though the in-development xmlcat will very likely have
a mapping entry type to map namespace names into a URI.  Note,
this would apply only to namespace names that are recognizable
as namespace names per the Namespace spec, that is, something
that is the value of xmlns="whatever" or xmlns:prefix="whatever".
John is correct that some string in XML content that you might
think is a namespace URI isn't covered, since it isn't recognizable
as a namespace name by the XML processor.

>. . .
>
>OTOH, RDDL probably isn't done until it can map PUBLIC
>IDs to SYSTEM IDs like XML Catalog (as others have suggested in this
>thread).  Two 3/4 overlapping specs would be, well, redundant.  

I completely disagree.  As I've discussed before in this discussion,
xmlcat mapping is completely orthogonal to resource discovery directories,
and the latter, while a useful thing for a namespace URI to resolve to,
is not only tied to namespaces.

>As an aside, I think the difference between PUBLIC IDs and NS URIs is not
>important.  

I'm not sure what you mean by this.  I agree that the syntactic
differences are not that important, but you argue the opposite below 
which I find confusing.  And the semantic differences are crucial.

>I checked the grammar for the NS "URI" and it is a garden
>variety XML attribute value, not even a URI, URN, let alone a URL.  I look
>at the production for PublicId (in XML, not XML Catalog) and I see it has
>character set limitations (roughly a-z,A-Z + comic book cursing). This
>limitation is, after all, getting dated. 

The character set for public ids was strictly limited in SGML 1986.  
XML allows letters, digits (which you don't mention, but they were
allowed in SGML too), and most punctuation.  Admittedly, this only
covers most of 7bit ASCII.  On the other hand, the namespace name 
is a URI reference, so it does have some limitations too (see RFC 2396).

>To answer my own question to Tim
>B., perhaps the benefit of RDDL is that it isn't weighed down with SGML
>compatibility issues.
>
>To my mind, the discussion is all about a) deciding if a NS "URI" is an
>entity or not and b), if it is, what kind.  Personally, I think b) is a
>minor point until you decide a).

I don't really understand this.  A URI isn't an entity in either the 
mime sense of the word entity or the XML sense; perhaps you are using 
the term in yet another sense.

paul