[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: Vocabulary Combination

• To: <xml-dev@l...>
• Subject: Re: Vocabulary Combination
• From: Arjun Ray <aray@n...>
• Date: Thu, 29 May 2003 22:35:55 +0000
• In-reply-to: <00e701c325c0$61bc1310$4bc8a8c0@A...>
• References: <ab7bdvs3gol4j4trlefsiq6g4bfbaou70f@4...> <3ED5B096.9080102@t...> <jubbdvog5g10nv5oj0ea48qusgrlak5nit@4...> <009d01c325b7$d401da80$4bc8a8c0@A...> <30gbdvkoccrfkdoevtpvnhdfunff5vvooq@4...> <00e701c325c0$61bc1310$4bc8a8c0@A...>

"Rick Jelliffe" <ricko@a...> wrote:
| From: "Arjun Ray" <aray@n...>
 
|>| The namespaces set general semantics, 

|> The fact that a bunch of universal names share a URI "prefix" is no 
|> more than a coincidence.  They need not be parts of any coherent schema 
|> at all.  They're just a smorgasbord.
|  
| "set" is the wrong word: I should have said "The namespaces evidence the 
| original specific semantics, which we can think of as the general 
| semantics that we may have diverged from." 

But namespaces as per the rec *don't* evidence original semantics in that
sense.  Universal names are unique and isolate.  The congruence of URI
prefixes in such names is a coincidence.

The silliness in all this is the Platonist conceit that a "universal name"
in and of and by itself conveys all necessary semantics.  In the real
world, virtually all use of names is contextual and dependent on the use
of other related names.  What matters is what ties these names together
into a coherent whole.  The individual names don't matter at all.   They
can be as formal or as instrumental as one pleases as long as one knows
the role each plays in a coherent scheme.  What will we hear next?  That
BNFs are no good unless the nonterminals are universal names, pristine and
immaculate in their isolate noumenal immanence?

| By which I mean the URI indicates who was the originator of the element, 
| who *of course*  has some semantic intent in the elements (even as a 
| smorgasbord). 

Not the originator of the element, but the originator of the *use* of that
name to convey some particular meaning or set of meanings.  Markup is not
ontological.  It is annotative and/or denotative only.

| This excludes that the namespace necessarily means that the element has 
| been used correctly, or used in the same way as the originator, or use 
| in the same kind of structures as the original.

Yes, the issues here are logically prior to those of schema-validity.

| I think we need to allow for this spectrum of private divergence
| and public adherence to the originator's semantics when discussing 
| namespaces

Absolutely!  It's a first principle of the SGML/XML *formalism* that
schema (or "document type") designers be free to choose their own names,
which in turn document instantiators would be expected to use coherently. 

| W3C standards in particular are often written with the tacet assumption 
| that they describe what goes on for the public web, not what goes on 
| behind-the-scenes. 

Hence the basic problem with colonification.  It declares the problem of
vocabulary combination "solved" by vocabulary imposition.  (That's what
the html:src versus xlink:href flap was all about, for instance.)

• References:
  • Vocabulary Combination
    • From: Arjun Ray <aray@n...>
  • Re: Vocabulary Combination
    • From: Tim Bray <tbray@t...>
  • Re: Vocabulary Combination
    • From: Arjun Ray <aray@n...>
  • Re: Vocabulary Combination
    • From: "Rick Jelliffe" <ricko@a...>
  • Re: Vocabulary Combination
    • From: Arjun Ray <aray@n...>
  • Re: Vocabulary Combination
    • From: "Rick Jelliffe" <ricko@a...>

• Prev by Date: Re: Vocabulary Combination and optional namespaces
• Next by Date: Pay no attention to the URI behind the curtain
• Previous by thread: Re: Vocabulary Combination
• Next by thread: Vocabulary Combination and optional namespaces
• Index(es):
  • Date
  • Thread