W3C home > Mailing lists > Public > xml-uri@w3.org > May 2000

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

From: Clark C. Evans <cce@clarkevans.com>
Date: Sat, 27 May 2000 23:28:11 -0400 (EDT)
To: xml-uri@w3.org
Message-ID: <Pine.LNX.4.10.10005272310110.17513-100000@clarkevans.com>
Clark C. Evans wrote:
> However, is this practical in our context?  I mean, is 
> there any other serious reason for retrival other than 
> for getting a schema or human description or catalog of 
> resources for the namespace?  In each of these examples 
> none of the theoritical problems occur...

I was trying to ask if there is a practical example 
where namespace-by-retrival would violate the two
concerns John Cowan raised: 

  a) two retrivals from different URLs 
     being equivalent by co-insidence

  b) two retrivals from the same URI
     differing due to a function of time.

Given a "http:\\my-schema-location" or 
some other informational text, I cannot 
think of a case where either would occur:

  a) Two different schemas would definately
     have a different text; and would therefore
     be different.

  b) If two retrivals are byte-for-byte different;
     then someone changed the schema.  In this
     case, the schema is different!  

Also, given rigourous use of expiration dates
(necessary for digital signatures and other
contracts) I don't see a case where caching
could not be used to provide a consistent
snapshot of the schema for a given process.
One would use the most recent retrival with 
an expiration date before the start of the 
process.

Clark
Received on Saturday, 27 May 2000 23:24:11 GMT

This archive was generated by hypermail 2.2.0 + w3c-0.30 : Tuesday, 12 April 2005 12:17:23 GMT