Re: Best Practice for Renaming OWL Vocabulary Elements

Hi Martin,


On 4/21/2011 3:28 PM, Martin Hepp wrote:

> In my opinion, explicit versioning of ontologies should be avoided in general; rather, one should evolve deployed ontology in a backwards-compatible fashion only. While there is is quite some academic work on ontology versioning, fact is that in practice, if you use fine-grained versioning, the Semantic Graph will break more frequently and you will need a lot more inferred triples.
>
> So in GoodRelations, we use only one version of the ontology (http://purl.org/goodrelations/v1) and all modifications are as much backwards-compatible as possible. From time to time, we have widened the domain or range of a property from a single class to the union of multiple classes or other minor changes, but the side-effects of that are practically irrelevant, in particular if you judge them in the light of real-world data quality of WWW data.
>

Yes, of course, one outcome of the conversation on SemanticOverflow [1] 
was that one should always use the generic ontology URI*. This URI 
should preferably point to the latest (stable) version of the ontology. 
However, to be able to retrace some modellings or inferences of datasets 
that made use of an older version of an ontology, one should assign 
version IRIs to each ontology version as well. Furthermore, an applied 
ontology version in a dataset can be assigned with help of 
dcterm:conformsTo [2] or owl:imports.
Generally, if you do some changes in your ontology specification, you 
will always get a new version. Whether you express this modification 
explicitly or not.

> It's only now that I would like to use shorter labels for 2 - 3 conceptual elements that are already in use, without forcing anybody to

Yes, if you only like to introduce some new labels, then you can easily 
on change the value of the related rdfs:label relation. However, you 
proposal includes the introduction of new identifiers for concepts and 
relations. The naming of identifier should be well thought-out**.
Btw, DC relates versioning information to each concept or relation 
definition by utilising dcterms:hasVersion [3].

> From my perspective, the popularity / usage of an existing element is more important than the indicated degree of stability.

Well, this is a metric which is not measurable, because some terms are 
due its nature more often applied as other ones. For example, a label 
property, such as dc:title, has a higher usage than a very specific 
(/narrow) property of a specialised purpose ontology, e.g., 
void:sparqlEndpoint of the VoID vocabulary.
That is why, information about a design status of a term of an ontology 
specification might sometimes be useful, e.g., properties that are 
marked as 'experimental' should rather be avoided in a dataset of a 
production.

Cheers,


Bob


*) for example it might be better to utilise, e.g., 
http://purl.org/goodrelations/, as generic URI of GoodRelations ;)
**) from my POV an example of a rather bad identifier naming in 
GoodRelations is gr:ProductOrService

[1] 
http://answers.semanticweb.com/questions/2815/how-do-i-knowmodel-the-applied-version-of-an-ontology-specification
[2] http://dublincore.org/documents/dcmi-terms/#terms-conformsTo
[3] http://dublincore.org/documents/dcmi-terms/#terms-hasVersion

Received on Thursday, 21 April 2011 16:00:48 UTC