Namespace, Versioning Re: the Tool class

On Wed, 15 Jun 2005 17:26:29 +0200, Charles McCathieNevile  
<chaals@opera.com> wrote:
[snip]
> I think we should use a new URI for this Class - I suggest that we  
> simply change the suffix to ToolG so that they are not the same thing,  
> and maintain the namespace. That saves building versioning tools that  
> have to convert everything from one namespace to another, which strikes  
> me as beneficial - let's not make extra work for increasing  
> compatibility and convergence while we don't have to.

I hadn't noticed until now that at some point in the editing of the  
proposed changes the namespace was changed in the propsed new version.

I don't recall this being discussed and agreed, and I would like to note  
that I am strongly opposed to the change (just formalising it as an  
objection, in case that wasn't obvious :-). It seems to me that while we  
are developing the vocabulary we should maintain a single namespace, and  
add new versions of terms to that namespace if we change their meaning,  
rather than changing the entire vocabulary.

I realise that URIs are more or less infinite resources, but time to write  
converters that simply notice that one URI means the same as another is  
not an infinite resource. Ability to auto-process documents through OWL  
statements that two URIs are identical in meaning is also reliant on  
finite resources.

If we have to have a different namespace for the final terms (I suspect  
W3C are likely to want a "clean URI" for the final version of the  
vocabulary) then we should simply keep working in the developing namespace  
we have until then, and only at the point of final publication should we  
make the transition and force everyone to add the revision into their  
tools.

Where we decide formally to deprecate something, or version it, there are  
OWL terms that allow us to declare this in a machine readable way. However  
I do not think that we should make OWL statements about the development  
namespace while we are in development mode since thre is no way to retract  
those statements from the Web, and if (for example) we decide we want to  
revert some change we would be unable to do so without making a new term.  
So I propose that all the terms we provide should remain available until  
final publication, at which point we deprecate them all and note what  
happened to them. It might be worth considering something like the  
testing/stable terms used in some vocab projects, but I am not terribly  
concerned either way.

Cheers

Chaals

-- 
Charles McCathieNevile                              chaals@opera.com
          hablo español - je parle français - jeg lærer norsk
   Here's one we prepared earlier:   http://www.opera.com/download

Received on Thursday, 16 June 2005 18:07:34 UTC