W3C home > Mailing lists > Public > public-owl-wg@w3.org > November 2007

Re: Choosing a "short name" for our working group pubs

From: Jeremy Carroll <jjc@hpl.hp.com>
Date: Mon, 05 Nov 2007 13:53:56 +0000
Message-ID: <472F2074.20606@hpl.hp.com>
To: Alan Ruttenberg <alanruttenberg@gmail.com>
CC: public-owl-wg@w3.org

Alan Ruttenberg wrote:
> On Nov 2, 2007, at 1:20 PM, Jeremy Carroll wrote:

>> http://www.w3.org/2005/05/tr-versions
>> [[ However, one common expectation when using the major/minor version 
>> scheme is that, for a given major version number, the Recommendation 
>> with the highest minor version number supersedes all others sharing 
>> that major version number. By supersede, we mean that authors and 
>> implementers should stop using the old version and start using the new 
>> version; in effect the new version masks the old one. The status 
>> section of a minor version should state clearly that it supersedes the 
>> previous minor version. ]]
>>
>> I do not believe there will be community consensus that OWL 1.1 should 
>> mask OWL 1.0, hence OWL 1.1 seems an inappropriate name for a 
>> recommendation that evolves from the member submission.


> 2) On the question of whether we release OWL 1.1, I'm not sure I see 
> what you mean by masking. First, in the section quoted it says there is 
> an expectation, but obviously not a certainty, as it instructs that the 
> status section explicitly say what the policy is. 


> Further I don't see 
> this sort of thing happening uniformly -  just because there is http 1.1 
> doesn't mean people don't use http 1.0. 

HTTP isn't a W3C spec, so the versioning policy may differ.
Also it's nto clear which of the W3C specs have version numbers which 
respect the quoted version policy.


IMO A version masks an earlier version if there is no plausible reason 
not to upgrade.

In the minimal version of what this WG is doing, we are adding some new 
items to the OWL vocabulary, to support QCRs and sub property chains; 
and providing some new mechanisms in support of 
profiles/fragments/subsets or something.

In this minimal version, it is certainly at least arguable that there is 
no reason not to upgrade and OWL 1.0 to OWL 1.1 would be the correct 
numbering on this policy.

The reason non-vocab-extensions issues such as punning, or the mapping 
rules, or whatever are, at least for HP, more contentious - is that the 
balance between DL and Full is fragile, and changes other than simple 
vocab extensions threaten that balance, and it may be plausible that 
because of that there would be good reasons to stay with OWL 1.0 and not 
upgrade. In particular, such reasons may persuade HP's Jena team not to 
upgrade.

> Finally, our charter says we are 
> aiming for backwards compatibility. 

Yes - if we achieve this then maybe the number 1.1 will be appropriate.
But that means taking backward compatibility with OWL 1.0 as more 
important than honouring the intent of the member submission document. 
While this is my strong preference, I get the feeling that many WG 
participants have a greater allegiance to the member submission docs, 
than to the current recommendation.


Jeremy
Received on Monday, 5 November 2007 13:54:56 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:13:27 GMT