an unfilled need

[David Megginson]

Ann Navarro writes:

 [on Namespaces]

 > That need is: a formal deterministic means of discovery. 

Discovery of what?  Structural rules are easy enough, but they buy you
almost nothing -- after all, if you give me an arbitrary piece of XML
with a DTD, my software can tell you whether it's valid or not, but
not what it means.

 [snip]

 > The "rest of the world", is looking for a means of discovering what
 > belongs with that name, it's associated definitions, semantics, and
 > other data that may be necessary to complete their operations.

I cannot imagine how we could accomplish this, without simply mapping
to an artificially restricted set of a-priori meanings.

For example, for displaying HTML in a browser all the meaning you need
is a collection of flow objects ("block text", "link", "graphic",
etc.), but what kind of schema language or other discovery mechanism
could allow software automatically to determine the meaning of

  <irony>Have a nice day!</irony>

or

  <task>
   <tools-required>
    <tool>spanner</tool>
    <tool>needle-nosed pliers</tool>
   <tools-required>
   <steps>
    <step>loosen the main bolt with the spanner</step>
    <step>pull out the blue wire with the needle-nosed pliers</step>
   </step>
  </task>

?

My point is that it's a pointless task, at least without access to
much more advanced artificial intelligence than is currently
available.  

In the mean time, though, it is entirely possible to write software
that *does* know what it's looking for, such as a search engine that
knows that HTML <cite> can hold a book title, or a Newsroom system
that knows that XMLNews <subject-name> holds a description of one of
the subjects of a News story.  Namespaces are valuable precisely
because they allow software to find markup that it knows about and to
do something useful with it.

While the Utopian vision of software that can figure out how to deal
with what it *doesn't* know about is enticing, we're so far away from
any practical implementation right now (except, again, for
artificially-restricted subsets) that standardizing would be a waste
of time.  

That's not to say that there's no point in standardizing some of those
tiny subsets by providing facilities for linking to structural schemas
and stylesheets, but it does mean that structural schemas and
stylesheets provide don't solve the problem of what markup means -- so
far, and for the forseeable future, that knowledge is hard-coded in
the software itself at some level.


All the best,


David

-- 
David Megginson                 david at megginson.com
           http://www.megginson.com/

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)