re: Overloaded URIs must GO!

Paul Prescod writes:

 > Furthermore, that document defines the http:-syntax URL as being
 > tied to the HTTP protocol which it clearly is not when it is abused
 > as in
 > 
 > xmlns:xsl="http://www.w3.org/XSL/Transform/1.0"
 > 
 > I haven't yet heard a defense (credible or otherwise) of this nasty
 > and non-intuitive practice. If you don't want your URI to be
 > locatable and retrievable on the Internet then *please do not use
 > the URL syntax!*

There are two good arguments in favour, though no doubt these could be
answered:

1. URNs don't really exist, or at least, last I checked, there was no
   authoritative specification of the different URN schemes (without
   which URNs are worthless).  URNs have been under development for
   most of the 1990s with few tangible results, and I'm growing
   slightly skeptical.

2. DNS alone is not sufficient, because people acquire Web space based 
   on also on protocol and directory partitioning; for example, I have
   control over the content of 
   
    http://home.sprynet.com/sprynet/dmeggins 

   but *not* of

    ftp://home.sprynet.com/sprynet/dmeggins

   or

    gopher://home.sprynet.com/sprynet/dmeggins

   To guarantee uniqueness (however temporary), I have to base the
   namespace on something that I some kind of unique access rights to, 
   and in my case, the HTTP protocol is an integral part of that.

Now, there is no reason that we couldn't embed the HTTP URL into a
URN, except that (unless something has happened recently) there is no
normative definition of URN schemes, therefore, there's nothing I can
use.  

What alternative do I have to building my namespaces on top of URLs?


All the best,


David

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

Received on Friday, 28 May 1999 11:09:43 UTC