Re: Biological Taxonomy Vocabulary 0.1

On 11 May 2008, at 10:39, Dan Brickley wrote:
>>>> Suffice to say, the new namespace URI is:
>>>>
>>>>   http://purl.org/NET/biol/ns#
>>>>
>>>> I can't imagine many people have had time to implement it yet,  
>>>> given that
>>>> I only posted about it about 16 hours ago, but for what it's  
>>>> worth, I'll
>>>> keep the old documents at the old URI for the next week or so, to  
>>>> allow
>>>> any early adopters to catch up.
>>> Consider using /additional/ versioned URIs, e.g.
>>>
>>>  <http://purl.org/NET/biol/20080510/ns#>
>>>  <http://purl.org/NET/biol/20080921/ns#>
>>>
>>> from which people can access the previous versions of your  
>>> ontology. The
>>> version-free URI, which you have chosen now, will always refer to  
>>> the
>>> latest of these versioned ontologies.
>>> The URI of the ontology, which is used within the ontology header,  
>>> will
>>> always be the version-free one. But you may, as a hint to people,  
>>> add an
>>> 'owl:versionInfo' annotation property to the ontology header, which
>>> points to the respective versioned URI:
>>>  <owl:Ontology rdf:about="http://purl.org/NET/biol/ns">
>>>    <owl:versionInfo rdf:resource="http://purl.org/NET/biol/20080510/ns 
>>> " />
>>>  </owl:Ontology>
>> I think this is probably a good approach, but I'd refine it a  
>> little bit
>> and say that the "main" namespace should not so much dereference to  
>> the
>> "latest" ontology, as the "latest release".  There can (and probably
>> should) be alpha- and beta-test versions which are "later" but not  
>> (yet)
>> suitable replacements.
>> Make sense?

This would mean that there could be “draft” versions of proposed  
changes, so people can review them before they get published in the  
main version of the vocabulary. This is probably overkill for most  
vocabularies, but sounds like a good technique if broad community  
review of new versions is desired.

>> (I've never done this with ontologies -- there could be a
>> problem with it -- but i think it's right.)
>
> Makes sense to me (although 'latest ontology' to me doesn't  
> necessarily include any experimental versions). Maybe we can borrow  
> terminology from Subversion etc and have a 'trunk' URL for works-in- 
> progress?

How about this: Work-in-progress happens in branches, and gets merged  
back into the published trunk. This means that experimental versions  
should be named something like <20080511/draft#>, and if the changes  
get accepted after a review period, they go into <20080601/ns#>.

Richard




>
>
> Dan
>
> --
> http://danbri.org/

Received on Sunday, 11 May 2008 10:10:37 UTC