MO clean-up
Bob Ferris
since I recently looked a bit more in detail on the Music Ontology
specification, I came to the conclusion that MO might need a clean-up.
What do you think about this? It is especially useful to deprecate some
relations and concepts, since they are bit out-of-scope, to narrow or
covered by recently created (outsourced) ontologies.
Here is my proposal to deprecate the following terms:
1. Trading relations. I think they should be covered by terms of the
GoodRelations vocabulary [1] (I haven't checked them).
- mo:exchange_item
- mo:possess_item
- mo:sell_item
- mo:want_item
- mo:mailorder
- mo:freedownload
- mo:paiddownload
2. Concrete information service relations. Albeit, they seem to be quite
useful one a first view, they aren't designed to cope with the fast
evolution of the Web (music information services that are not covered,
e.g., Last.fm, SoundCloud, Echo Nest, AllMusic). To address (web)
information services and be able to describe them more in detail, I
designed Info Service Ontology [2] (some time ago). Please also have a
look at the small collection of proof-of-concept examples [3]. One can
relate to each website etc. its information service, see an example at [4].
- mo:myspace
- mo:musicbrainz
- mo:discogs
- mo:imdb
- mo:musicmoz
- olga
- onlinecommunity
- wikipedia
3. misc properties:
- mo:license (-> dc:license)
- mo:origin (-> something from the Bio Vocabulary (?))
- mo:publishing_location (deprecate to the introduction of mo:ReleaseEvent)
4. All mo:Medium sub classes, since kurtjx created for this concern the
Media Format Ontology [5], i.e., all these sub classes are already
outsourced to this extension ontology (or?).
=> result a clean up of 10 classes and 18 properties. Cool, eh? ;)
Cheers,
Bob
[1] http://purl.org/goodrelations/v1
[2] http://purl.org/ontology/is/core#
[3] http://purl.org/ontology/is/inst/
[4]
http://infoserviceonto.sourceforge.net/is/spec/infoservice.html#sec-example
[5] http://kakapo.dcs.qmul.ac.uk/ontology/mfo/0.1/mfo.html
Bob Ferris
I checked the GoodRelations vocabulary [1] re. the mentioned business
items (more below).
On 4/6/2011 7:46 PM, Bob Ferris wrote:
> Hi,
>
> since I recently looked a bit more in detail on the Music Ontology
> specification, I came to the conclusion that MO might need a clean-up.
> What do you think about this? It is especially useful to deprecate some
> relations and concepts, since they are bit out-of-scope, to narrow or
> covered by recently created (outsourced) ontologies.
> Here is my proposal to deprecate the following terms:
>
> 1. Trading relations. I think they should be covered by terms of the
> GoodRelations vocabulary [1] (I haven't checked them).
>
> - mo:exchange_item
-> gr:offers [2], or a sub property of gr:offers
> - mo:possess_item
-> gr:owns [3]
> - mo:sell_item
-> gr:offers [2]
> - mo:want_item
-> gr:seeks [4]
> - mo:mailorder
-> I guess, this is a bit out-of-date, or?
> - mo:freedownload
-> maybe a description that make use of the
gr:DeliveryChargeSpecification concept [5]
> - mo:paiddownload
-> maybe a description that make use of the
gr:DeliveryChargeSpecification concept [5]
What do you think about the proposed mapping/alignment/usage suggestions?
Cheers,
Bob
[1] http://purl.org/goodrelations/v1
[2] http://www.heppnetz.de/ontologies/goodrelations/v1#offers
[3] http://www.heppnetz.de/ontologies/goodrelations/v1#owns
[4] http://www.heppnetz.de/ontologies/goodrelations/v1#seeks
[5]
http://www.heppnetz.de/ontologies/goodrelations/v1#DeliveryChargeSpecification
Dan Brickley
> Hi,
>
> since I recently looked a bit more in detail on the Music Ontology specification, I came to the conclusion that MO might need a clean-up.
There's an underlying assumption here that isn't universally shared.
You seem to believe that each ontology ought to purge itself of terms
that are available in other RDF vocabularies.
If we take that practice to an extreme, we risk making our instance
data rather complex, with terms from many different vocabularies
alongside each other. The potential to do this is of course one of the
strengths of RDF, ... but it is not always the only reasonable
approach. Ontology designers are also free to try to be relatively
self-contained, to simply things (including documentation and instance
syntax) for their users. So perhaps some of your purge/cleanup could
be re-characterised as a set of mappings between vocabs, rather than a
switch to new descriptive idioms?
cheers,
Dan
Bob Ferris
may I have to clarify some of my thoughts a bit (see below)
Yes, of course, I didn't exclude a mapping task at all ("What do you
think about the proposed mapping/alignment/usage suggestions?") ;)
However, separation of concerns and application of probably more
established and appropriated terms is important to easier establish a
"shared understanding", or?
Cheers,
Bob