Re: Namespace names: a modification of a semi-serious proposal

-----Original Message-----
From: Clark C. Evans <cce@clarkevans.com>
To: David Carlisle <david@dcarlisle.demon.co.uk>
Cc: cce@clarkevans.com <cce@clarkevans.com>; xml-uri@w3.org <xml-uri@w3.org>
Date: Sunday, May 28, 2000 11:31 AM
Subject: Re: Namespace names: a modification of a semi-serious proposal


>On Sun, 28 May 2000, David Carlisle wrote:
>> > What use would "mid:" or "cid:" have as a namespace name?
>>
>> The very purpose for which namespaces are designed.
>> They provide globally unique identifiers without the need to wait
>> for iso to give you an FPI prefix.
>
>IMHO,  "data:com.clarkevans.timesheet.submit"
>[or     "http:\\clarkevans.com\timesheet\submit.xml"]
>are just as unique, and far more readable.

But that means the resource whose only and definitive *representation* is
"com.clarkvans.timesheet.submit".

Someone said you can't dereference a data: URI. Wrong - you can verey
easily.
When you dereference "data:com.clarkevans.timesheet.submit" you get
"com.clarkevans.timesheet.submit"

The meaning of to "dereference" is "to aquire a digital representation of
that identified by".
If you make a system which allows dereferencing but renders soemthing random
then
it is not what people would expect.

timbl

>
>
>

Received on Thursday, 1 June 2000 17:45:19 UTC