xml-dev - Re: Begging the Question

[Joe English <jenglish@flightlab.com>]

Simon St.Laurent wrote:

> At 03:28 PM 12/29/00 -0800, Andrew Layman wrote:
> >Petitio Principii or Begging the Question: 'assumption of the basis'.  The
> >fallacy of founding a conclusion on a basis that as much needs to be proved
> >as the conclusion itself. [...]
>
> I would humbly suggest that it might be reasonable at this point to put
> "namespaces mean X because the namespaces spec says so" into the same
> category as "one must keep servants because all respectable people do so."

Huh?

I would think just the opposite: "namespaces mean X because
the namespaces spec says so" looks very much like a tautology
to me.


> That would suggest that because the namespaces spec has been the subject of
> so much (unfulfilling) argument, it may be considered to have virtually no
> traction whatsoever on any but the most limited points - an attribute-based
> syntax for associating identifiers with prefixes which uses element
> structures to define its scope.

But [REC-xml-names] doesn't attempt -- or claim -- to define
anything *other* than those very limited points.  That's all
it says.  That's all it means.

Most of the unfulfilling argument surrounding it springs from the
assumption that, since namespace names *look* like URLs, they should *act*
like URLs -- that is, that one should be able to to point a Web Browser
at them and retrieve something useful since they look like something one
might point a Web Browser at.  This assumption, while not unreasonable,
is explicitly disclaimed by the namespaces spec.  Namespace names are
Identifiers, not Locators.

[ I would have said "Namespace names are names, not addresses," but
  it seems the official W3C position is that this distinction is
  meaningless.  Instead -- and it took me a long time to figure
  this out -- everything is founded on a pair of circular definitions,
  namely: "A URI is anything that identifies a resource" and "A resource
  is anything that is identified by a URI".  The only way to make sense
  of most W3C specs -- RDF especially, but REC-xml-names is no exception
  -- is to take "resource" and "URI" as atomic ontological entities
  with "resource === URI" as an axiom.  But I digress. ]

Now one might argue that using (syntactic) URLs as (semantic) names
has caused a great deal of confusion and contention (with this I
would agree), and that perhaps it was a bad choice (here I'm not
so sure; it seems to work OK in practice in spite of all the confusion
and contention.)

I also agree with your conclusion, which is worth repeating:

> [REC-xml-names] may be considered to have virtually no
> traction whatsoever on any but the most limited points - an attribute-based
> syntax for associating identifiers with prefixes which uses element
> structures to define its scope.

but I would change "may" to "must" and "virtually" to "absolutely"
in the first clause, and replace the premise of your argument
with "because the namespaces spec says so" as justification :-)


--Joe English

  jenglish@flightlab.com