RE: I-D ACTION:draft-daigle-uri-std-00.txt

At 12:30 PM 9/7/00 -0700, Henrik Frystyk Nielsen wrote:
>> I've had more than enough
>> epistemology and language theory to appreciate the
>> distinction between a
>> name and a thing and the perilous connections between them.
>
>Apparently not - you are mixing up the name and the description of the
>thing the name identifies.

I don't think I'm the one mixing up the name and the description, or the
name and the thing - I'm complaining that URI specification fails to make
that distinction clear.

I have a definite appreciation for the difference between signified and
signifier, but I'm not willing to accept that the signified exists solely
in relation to its signifier.

>> Does http://www.w3.org/1999/xhtml refer to that Web page
>> alone?  I don't think so.
>
>The Web page is one manifestation of that resource - there may be others.
>You are likely to be told in the HTTP response whether there are others
>and how they differ.

I'm not sure that holds as a general rule - nor do I know how I'd describe
the page returned by a GET in some manner that didn't also seem to describe
the namespace.

>> >> >For "elements of the common syntax", the equality operation
>> >> is defined by
>> >> >RFC 2396. For everything else, you use case-sensitive matching.
>> >>
>> >> That's not specified anywhere I've seen, except in the
>> >> Namespaces in XML
>> >> Rec that got us into these problems in the first place.
>> >
>> >No, the XML-NS doesn't make it clear that the first part of
>> the sentence
>> >is true.
>>
>> It makes it clear that the second sentence is true.
>
>which is not enough.

We've been on this terrain before.  I think it's clear that I'm not
sympathetic to scheme-by-scheme comparison, especially if case folding is
involved.

>> >RFC 2396 specifies the equality rules for each common syntax element.
>>
>> But it doesn't provide equality rules for URIs that don't use
>> the common
>> syntax elements - and acknowledges the possibility of such cases.
>
>Yes it does - it is the "opaque_part" BNF construction. By default,
>elements are case-insensitive, the exceptions are listed in section 6.

If that's the case, Section 6 is very poorly written - that's hardly an
obvious interpretation.

To say it again, RFC 2396 needs a thorough rewrite expounding on all the
parts that are unspecified but mysteriously assumed if we want to use it as
a foundation for application-building.

Simon St.Laurent
XML Elements of Style / XML: A Primer, 2nd Ed.
XHTML: Migrating Toward XML
http://www.simonstl.com - XML essays and books

Received on Thursday, 7 September 2000 17:11:13 UTC