Re: [Patterns] Materialize Inferences (was Re: Triple materialization at publisher level) from Vasiliy Faronov on 2010-04-09 (public-lod@w3.org from April 2010)

From: Vasiliy Faronov <vfaronov@gmail.com>
Date: Fri, 09 Apr 2010 20:07:51 +0400
To: Dan Brickley <danbri@danbri.org>
Cc: Leigh Dodds <leigh.dodds@talis.com>, public-lod <public-lod@w3.org>
Message-Id: <1270829271.3263.37.camel@midgard>

I think Dan has spotted a very good rule of thumb for subclass
materialization with his notion of "mid-level" classes.

Here's another rule of thumb I can think of: materialize inferences that
map your data to better known, and more widely deployed, vocabularies.

Example. A consulting company could develop a custom ontology for
describing businesses. Let's say it has a class ex:BusinessEntity which
has owl:equivalentClass gr:BusinessEntity. It's likely that some LD
clients will be familiar with the GoodRelations vocabulary but unable or
unwilling to do reasoning over custom ontologies. In this case,
explicitly spelling out that every ex:BusinessEntity is also a
gr:BusinessEntity may be helpful.

-- 
Vasiliy Faronov

Received on Friday, 9 April 2010 16:08:36 UTC