I agree with Lina that a brief introductory description of key terms would be useful in order to differentiate the constructs Lina refers to (ontologies, data models etc.). This clarification of terms is needed percisely for the reasons Martin points to, namely that there is a "tendency to (ab)use the term 'ontology' for everything that has to to with data and knowledge" including terminological data. There is worrying evidence of this in projects and documents in the context I work in including museum projects.
Regards! Mika ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Synapse Computing Oy, Arabiankatu 2, 00560 Helsinki [email protected] +358-9-8569 9696 puh/tel +358-9-8569 9595 fax ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ _____ From: martin [mailto:[email protected]] To: Lina Bountouri [mailto:[email protected]] Cc: [email protected] Sent: Wed, 05 Nov 2008 13:25:37 +0200 Subject: Re: [Crm-sig] About the short document for the archive community Lina Bountouri wrote: > Dear all, > > > > I believe we should first of all make clear what a conceptual model is > (_briefly_). We have to distinguish the role of conceptual model > vis-à-vis metadata schemas, content standards etc. For example, metadata > schemas are focusing description and on covering the search and > retrieval needs of users, whilst ontologies’ target is to specify the > semantic notions that define a domain (e.g. cultural heritage, > multimedia, legal etc.) and their interrelationships providing a wider > yet explicitly specified aspect, mostly oriented to cover information > integration needs (i.e. facilitates reasoning, higher degree of > abstraction, query model that corresponds to the users’ view of the > domain, integration, inter-ontology mappings). Dear Lina, We should be careful to compare comparable things. Unfortunately, the domain is full of contradictory definitions. I recommend as a discussion base: Section 4.1.5 in: "PATEL, M., KOCH, T., DOERR, M., TSINARAKI, C., GIOLDASIS, N., GOLUB, K., AND TUDHOPE, D. 2005. Semantic Interoperability in Digital Library Systems, DELOS Network of Excellence on Digital Libraries – deliverable 5.3.1, June 2005." http://delos-wp5.ukoln.ac.uk/project-outcomes/SI-in-DLs/ To my understanding, "ontology" in general has no target other than any human conceptualization. An ontology can explicate the constructs of any metadata schema. A "metadata schema" IsA normal schema with a particular relationship, which does not affect its substance in contrast to "conceptual model". It is a normal data schema "used" to describe objects that happen to contain data. This accidental relationship has no bearing on the fundamentals of its substance. It is on the same level as "office automation schema". The CRM is an ontology particularly made to explain the underlying concepts of metadata schemata and other data schemata. In the tendency to (ab)use the term "ontology" for everything that has to to with data and knowledge, seems also to have as an effect that information science seems slowly to adopt the term as synonym of terminological data (subject terms, classification systems etc.). For CRM-SIG, I suggest to stay with Nicola Guarino's definition of ontology basically saying (in my words): a formal theory about possible states of affairs approximating a human conceptualization. Kind regards, Martin
