W3C home > Mailing lists > Public > www-tag@w3.org > May 2003

Re: denotation of http://www.vrc.iastate.edu/magritte.gif

From: Reto Bachmann-Gmuer <reto@gmuer.ch>
Date: Wed, 28 May 2003 18:24:19 -0400
Cc: www-tag@w3.org
To: Tim Berners-Lee <timbl@w3.org>, Bjoern Hoehrmann <derhoermi@gmx.net>
Message-Id: <1FAC4C7E-915B-11D7-9D30-003065CDBE5C@gmuer.ch>

Hash: SHA1

Bjoern Hoehrmann wrote:

> * Reto Bachmann-Gmuer wrote:
>> I think a HTTP-URI should identify a document as the abstract entity
>> that all of the possible result of a http-request represent, that is
>> the document independently of encoding and language.
> How would you query a webserver for a list of available representations
> of the abstract entity? What if you need to reference a specific
> representation?
I don't think that to determine the meaning of the URI  it is necessary 
to actually have all the results of possible  http-request. If  a 
webserver returns the image of a pipe when requesting the URI with the 
accept header set to image/gif and returns the image of an apple when 
the accept header is set image/jpeg the webserver is not behaving as 
expected and one of the two responses misses the actual meaning of the 
URI. If you agree that the webserver  misbehaves you agree that the 
HTTP-URI identifies the document a user sees in a browser when 
accessing the URL independently of the content-negotiation between 
client and server.

Tim Berners-Lee wrote:
> That is a good way of running a server.  that is what
> is represented on our server by the URI without a ".gif".
> But it is also very useful to have a URI specifically for the GIF 
> version.
> So the best is to have both. But use the generic one wherever you 
> can...
I think there may be good reasons to have a URI to identify the actual 
bytes returned by a webserver on a specific request, however I don't 
think the HTTP-URI is a reasonable choice. There is no way to verify 
that the sentence "http://w3.org/foo.gif  has 3457 bytes" is true while 
"http://w3.org/foo has 3457 bytes" is untrue (or of indeterminable 
boolean value) even if the server may return a 3457 byte image on a 
particular request. There should be an RDF-vocabulary to describe an 
http-request so that you could specify something like 
"urn:urn-5:jhk2345sd8a9dfa is what you get requesting 
http://w3.org/foo.gif  sending an http-request with accept-header set 
to image/gif" in this case you could make the verifiable assertion 
"urn:urn-5:jhk2345sd8a9dfa has 3457 bytes".


Version: GnuPG v1.0.7 (Darwin)

Received on Wednesday, 28 May 2003 18:24:42 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:55:59 UTC