[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: URIs harmful

• To: 'Paul Prescod' <paul@p...>, Joshua Allen <joshuaa@m...>, xml-dev@l...
• Subject: RE: URIs harmful
• From: "Bullard, Claude L (Len)" <clbullar@i...>
• Date: Fri, 26 Jul 2002 12:52:52 -0500

If I want a representation of a beach, please send 
the photos with the girls, not RDF or text prose. 

The problem is this:  for too many people for 
too long, http://whatever means "address of something I can get" 
and not "name which has the meaning I want to give it". 

http means "dereference" first, not secondarily.  It has 
a "preferred reading" and what is being added by claiming 
it for "names as preferred" flies in the face of common
usage.  It is always dereferenceable first, and then 
it is a name.  You are stuck with that fact.  By design.

Spilt milk.   People who want to use things which are 
clearly just names should be using urn:whatever.  The 
battle to make http prefaced strings mean whatever one 
wants them to mean was lost a long time ago.  I am 
well familiar with the "names don't have meaning, 
only people do" approach to semiotics and linguistics.  

sign -> signifier, signifieds, referents

Stand up in a crowded room and yell "FIRE!" then 
when you have made bail, we can all meet at the 
local pub and you can tell us what you really meant. 
But it has to be local since you will be under 
restraints of a judge and a bail bondsman.

Fielding is right that it has become like a word, 
and now, it has a preferred reading: GET.  So 
given namespaces, it will be a good idea to 
put something at the end of the URL being used 
to uniqueify the QName.  Or simply, polite.

len

-----Original Message-----
From: Paul Prescod [mailto:paul@p...]

So HTTP URIs have the characteristics you outline *plus* they are easy
to dereference for prose, RDF or whatever else is a good representation
for the resource.

• Prev by Date: Re: version numbers and infosets
• Next by Date: RE: DNS based URIs that don't imply access method
• Previous by thread: RE: URIs harmful
• Next by thread: RE: URIs harmful
• Index(es):
  • Date
  • Thread