OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

 


 

   Re: [xml-dev] Vocabulary Combination and optional namespaces

[ Lists Home | Date Index | Thread Index ]

"Simon St.Laurent" <simonstl@simonstl.com> wrote:

| It seems like namespaces aimed at diambiguation with the "let's create 
| big long names" approach, and then had to abbreviate the big long names 
| to keep the language usable.

Basically, yes.  The inveterate metaphysical babbling about "global
uniqueness" or whatever is just smoke to make up for the lack of a fire.

At the root of the disambiguation problem, as conventionally conceived, is
the entirely unnecessary premise that all vocabulary-specific names (i.e.
names of externally fixed provenance) must be syntactically visible, as
garden-variety generic identifiers and attribute names of the much simpler
world where only one externally fixed vocabulary applied to a document and
at that, by default.

This is what leads to idiotic flaps such as html:src *versus* xlink:href.
The same value makes sense in two vocabularies simultaneously, but
everyone gets hung up on the fact that syntactically an attribute can have
only one name.  Well then, let that circumstantial name be arbitrary, and
have a syntactic means to say that x:a and y:b and z:c and whoever else
wants to come to the party have each as their value that which is merely
*tagged* by the arbitrary name.  You could save on having to invent foos
and bars for this by using a or b or c or whatever if the same syntactic
means said that one of the name mappings was in fact an identity mapping.

The same delusion about all names having to be syntactically visible leads
to stuff like 

   <h:td><xdc:author>Simon St.Laurent</xdc:author></h:td>

in Tim's example.  The same value, "Simon St.Laurent" is being ascribed to
two separate taxonomies simultaneously.  That by no means implies that we
somehow have to plunk two generic identifiers in the document, and at that
introduce factitious extra element "structure" in order to do so woodenly.

Note that, from the point of view of *authorial intent*, this markup could
have been:

  <xdc:author><h:td>Simon St.Laurent</h:td></xdc:author>

So, if it makes absolutely zero difference to all semantic intent, then
there should not be a difference in parsing one or the other.  But the
extra element structure *does* create a problem of whether the factitious
containment is "real" or not.  Entirely unncessarily.

The general problem here is that a specific vocabulary may apply to only a
partition of the element structure of a document, not the entirety.  There
is no presumption that all such partitions must be disjoint: they could
overlap (as in xlink:ref + html:src or h:td + xdc:author), and in many
real world cases, often do.  (Reflecting the fact that the same piece of
information is apprehensible in more than one taxonomy simultaneously).

If you have a means of determining such a partition from the instance
markup alone, then you have solved the vocabulary combination problem,
because the markup *tells* you where and how each vocabulary applies.

All you have to give up is the idea that all such names must occur
directly in syntactically visible positions.

The "problem of rcognition and collision", to quote a spec, is one
monumental and stupid illusion. 





 

News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS