Open retog opened 11 years ago
We should probably first consider using a single namespace for the ontologies and vocabs. It looks like ecs is at http://fusepool.eu/ontologies/ecs
. The current vocab is at http://fusepool.info/vocab#
. If you know of other ones, can you list them in issue #4 ? This is also related to issue #3 .
Frankly I think Fusepool has too many TLDs and there is some overlap on their use. In any case, we can probably stick to .eu for this. The "cost" of that is to also make sure that something will let it dereference there and return an RDF serialization.
I don't actually see a relationship between this issue (making your ontology easy to use in java code) with your comment (consistent naming). I agree naming should be consistent. May I suggest you change your ontology to the match the naming of the already existing one or raise an issue against ecs justifying why anothe naming should be chosen.
The vocabulary that's used for the datasets predates the ECS ontology. So, I would rather not decide based on what came first. Having said that, I don't mind going with the eu
TLD.
Regarding using a maven project structure - thats no problem, I think the engineers can take care of that if they need to use the vocabulary. Otherwise, the point of vocab.ttl is to simply have it available from a place like http://fusepool.eu/vocab
or whatever. So, it needs to be either a static file on the server or in the RDF store (or both), and be dereferencable.
I'm marking this issue as an enhancement because these type of things go around the vocabulary and its development (e.g., file path, additional metadata for maven) is not tied to a piece of software.
By creating a project like https://github.com/fusepool/fusepool-ecs/tree/master/ontologies an OSGi bundle proving java classes with constants for the defined properties and classes is generated. This is of great value when genearting java code using this ontologies.