Re: Toward the self-describing web [was: Irony heaped on irony]

-----Original Message-----
From: sam th <sam@bur-jud-118-039.rh.uchicago.edu>
To: John Cowan <jcowan@reutershealth.com>
Cc: Matt Sergeant <matt@sergeant.org>; xml-uri@w3.org <xml-uri@w3.org>
Date: Sunday, May 21, 2000 9:08 AM
Subject: Re: Toward the self-describing web [was: Irony heaped on irony]


>On Fri, 19 May 2000, John Cowan wrote:
>
>> Matt Sergeant wrote:
>>
>> > No. Namespaces are names. Unique identifiers. Nothing more, nothing
>> > less. They don't need to be resolved in any way shape or form. How are
you
>> > going to resolve this valid namespace uri:
>> >
>> >         xmlns:ms="mailto:matt@sergeant.org?subject=XPathScript"
>>
>> Red herring.  No one is calling for any change to this URI, since it is
>> already absolute.  Indeed, no "mailto:" URI can ever be anything but
>> absolute.
>
>Sadly, this is not true.  As a test, I just got mail to be delivered to an
>address of the form "name".  It was automatically interpreted as meaning
>"name@localhost.localdomain".  While this is unlikely to be useful
>(esp in the current context) it does work.


Ah. That behaviour was nothing to do with absolutization of the URI, which
only affects
schemes that use "/".   What happened was that the absolutizaion would have
left mailto:name as mailto:name.  (The mailto: spec may of course forbid -
this
- I would have to look it up. That spec of course just references the RFC
and I
bet the RFC allows a name without host.)

Give a mail addressed to "name" mail systems will tend to use defaults. So
an SMTP client getting a mail for name may try to deliver it locally.

But  -- it is not a relative URI in that the context is not the URI of the
document making the reference.

If you were (perish the thought) to use it as a namespace URI then the
absolute URI you would be using would be mailto:name. There are no relative
mailto: URIs.


>      sam th


Tim BL

Received on Sunday, 21 May 2000 17:55:55 UTC