Re: Philosphy 101

At 09:05 AM 2000-06-05 -0700, Larry Masinter wrote:
>
>Are there any practical considerations left besides deciding whether
>we should 'disallow', 'deprecate' or 'define' the use of relative
>URIs in namespace names?  Once you strip away all of the philosophical?
>

You left out 'undefine' which is more to the point.

If we just make it clear that 

  "Relative URIs are prima_facie suspect as far as being globally defined
are concerned.  Where a relative URI-reference appears in the ns-attr, you
may need to know more than what is provided in this Recommendation to
process it correctly." 

...we will have provided the technical substantiation for a warning and
correctly described the extent of the misfit between the de_jure status quo
and the de_facto status quo.

This sounds like something minor enough it could be published as an erratum
and any revision of the base document be left to follow a schedule geared
to the natural workings of the technology lifecycle.  That's as far as the
Namespaces Rec. is concerned.  We still need to have an answer for InfoSet
and XPath.

Then we could be able to get on to larger issues (i.e. of more practical
importance) such as what pattern of practice will work well for importing
more than just the names from a language module into the current document
in a sufficiently informative, open and interoperable way.


>Larry
>-- 
>http://larry.masinter.net
> 

Received on Monday, 5 June 2000 17:13:36 UTC