RE: purpose/goals for observations ontologies

Hmmm. 
For me 'modular' and 'self contained' are somewhat orthogonal concerns. 
But there really *are* authoritative sources for some base concepts - 
- SI for units of measure
- EPSG for coordinate systems
- the International Commission for Stratigraphy for the geologic timescale
- ISO 19107 for georeferenced geometry
Any ontology which wishes to position itself as authoritative should make
reference to these, so that it can be easily related to any other ontology
which also wishes to be authoritative. 
If each ontology stays self-contained, then relationships between them will
have to be based on 'mappings', stored in ... yes, another ontology. Why not
leverage what exists?


Simon Cox

European Commission, Joint Research Centre, 
Institute for Environment and Sustainability, 
Spatial Data Infrastructures Unit, TP 262
Via E. Fermi, 2749, I-21027 Ispra (VA), Italy
Tel: +39 0332 78 3652
Fax: +39 0332 78 6325
mailto:simon.cox@jrc.ec.europa.eu 
http://ies.jrc.ec.europa.eu/simon-cox 

SDI Unit: http://sdi.jrc.ec.europa.eu/ 
IES Institute: http://ies.jrc.ec.europa.eu/
JRC: http://www.jrc.ec.europa.eu/

-----Original Message-----
From: John Graybeal [mailto:graybeal@mbari.org] 
Sent: Wednesday, 12 August 2009 06:55
To: Simon Cox
Cc: Laurent.Lefort@csiro.au; public-xg-ssn@w3.org
Subject: Re: purpose/goals for observations ontologies

I believe I have seen in many places that if you want your ontology to be
re-used, it is best that it be "modular", which I understood to mean
self-contained.  If it has many references to the rest of the web, then its
complexity is higher, and its likely usability becomes lower (because
somewhere in those references will be a statement that conflicts with
statements in the user's world -- whether or not they "need" to use those
statements).

If my (admittedly naive) understanding is correct, this is essentially a
case of conflicting design architectures -- one has lots of little atomic
plug-in pieces, and another has deep and rich connections to components with
sophisticated capabilities or relationships. Both have value, but for most
applied uses I would suspect the former architecture is easier to reuse.

I welcome enlightenment.

John


On Aug 10, 2009, at 5:54 AM, Simon Cox wrote:

> What I find curious about this discussion, in the context of the 
> 'semantic web', is that O&M and HollowWorld are quite honest about the 
> dependencies on external resources.
> Inevitably, this can lead to a rather large graph, if you chase them 
> all down.
> But the point of the 'web' is deferred resolution of references, so 
> you only need look at a sub-graph at any one time, with interface 
> classes/ resources represented as URIs.
>
> OTOH many of the ontologies I see coming out of the semantic web 
> community have 'self-contained' as part of their design criteria.
> This seems to miss out on the 'web' part of 'semantic web'.
>




--------------
NOTE NEW EMAIL ADDRESS
--------------
John Graybeal   <mailto:jbgraybeal@mindspring.com>
Marine Metadata Interoperability Project: http://marinemetadata.org

Received on Wednesday, 12 August 2009 07:02:37 UTC