W3C home > Mailing lists > Public > www-html@w3.org > January 2000

XHTML Naming Rules - FPIs

From: Arjun Ray <aray@q2.net>
Date: Wed, 26 Jan 2000 00:22:12 -0500 (EST)
To: www-html@w3.org
Message-ID: <Pine.LNX.4.10.10001252307200.15214-100000@mail.q2.net>

On Mon, 24 Jan 2000, W. Eliot Kimber wrote:

> PUBLIC identifiers are fundamentally bogus and redundant with
> system identifiers (cf. Tim B-L's paper arguing that there is no
> fundamental difference between URLs and URNs).

IMHO, system identifiers which walk like URLs, talk like URLs - heck,
were *designed* to be no different - but still aren't "real URLs", are
just as fundamentally bogus.  The Namespace spec, for instance, has
verbiage specifically to try and stave off the natural and perhaps
inevitable problem: 404 Not Found.  

If the essential point is "global uniqueness" then non-functional URLs
with internal syntax and FPIs with internal syntax are six of one and
half a dozen of the other.  In fact, the problem with Namespace-style
URLs is the patina of functionality, thanks to the 'http:' prefix.
With a scheme prefix of something like 'nsrp:' (my invention of a
"NameSpace Resolution Protocol"), you get the right result from naive
software modules - not '404 Not Found', but 'Unrecognized Scheme'.
(How about editors and viewers that automatically highlight URLs in
text and launch a browser or something if you click on one?)  But
that's the same as formally converting a FPI to a URI as follows:

  1. Replace all spaces with underscores.
  2. Prefix 'fpi:' or 'iso9070:' or equivalent.

As such, these identifiers work like keys into system-specific hash
tables (hence, I suppose, the inclusion of the Catalog), but there's
no saying that *known* internal syntax can't or shouldn't be exploited
to rationalize the organization of such reference mechanisms.  Given
that a syntax for FPIs is indeed defined (ISO 9070), it's probably a
godd thing to attempt an explanation of it - though advisedly in an
informative rather than normative section. 

I think the FPIs should stay (in fact, I'd like to see a NOTATION FPI
added.)  Until somebody or something gets off the about Namespace
URIs, they're just as good, and besides, bridge to other systems quite
nicely.  (What if somebody *wants* to use AFs, never mind that the W3C
has kittens at the mere mention of them?)

Received on Wednesday, 26 January 2000 00:13:24 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:05:52 UTC