W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > October to December 1996

RE: Refering to versions

From: Yaron Goland <yarong@microsoft.com>
Date: Tue, 29 Oct 1996 11:53:15 -0800
Message-ID: <c=US%a=_%p=msft%l=RED-44-MSG-961029195315Z-1233@INET-01-IMC.microsoft.com>
To: "'Dale Gass'" <dale@ra.isisnet.com>
Cc: "'w3c-dist-auth@w3.org'" <w3c-dist-auth@w3.org>
It is different because the French and English versions and semantically 
equivalent. You should get the same information in the French and English 
versions. This is similar to requesting a document in MSWord or PostScript 
format. The information should be identical, they are semantically 
equivalent, even though the syntax's are radically different. However 
Version 1.2 and Version 1.3 of a document are semantically different. That 
is, generally, the reason for having different version numbers, to indicate 
semantic difference.
PS Yes, I know, you can also use versions to indicate syntactic difference 
but that is picking nits.

-----Original Message-----
From:	Dale Gass [SMTP:dale@ra.isisnet.com]
Sent:	Tuesday, October 29, 1996 2:28 AM
To:	Yaron Goland
Cc:	dale@ra.isisnet.com; w3c-dist-auth@w3.org
Subject:	Re: Refering to versions

> I think the argument is more fundamental. The feeling is that URLs point 
> a single resource. While a resource may have several representations each 
> representation is a semantically equivalent, although possibly degraded, 
> version of each other. Two different versions of a document are not
> semantically equivalent. While one could make an argument along the lines 
> of "degraded" content, I don't think the argument is very compelling. So 
> Henrik and others are arguing that we should specify version as part of 
> URL because we are referring to related but distinct resources.

But don't the Accept-*: headers allow selecting variants of a specific
URL?  I would argue that "Give me the French version" is not a request
unlike saying "Give me the Feb 14th version" or "Give me version 1.3"...

After all, a historical variant would (typically) be indicating what the
URL *once* pointed to...

Received on Tuesday, 29 October 1996 14:54:11 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:01:09 UTC