- From: Hans Teijgeler <hans.teijgeler@quicknet.nl>
- Date: Fri, 2 Dec 2005 12:57:04 +0100
- To: <jeremy@miko.hk>, "'Richard Newman'" <r.newman@reading.ac.uk>, "'Danny Ayers'" <danny.ayers@gmail.com>
- Cc: "'Semantic Web Forum'" <semantic-web@w3.org>
- Message-ID: <001401c5f737$871d3f30$6c7ba8c0@hans>
Hi Jeremy, Richard, and Danny, Thanks for your responses! Yes Danny, we have no other choice than working in OWL Full, come hell or high water, as the British say. I'm struggling with the following: In ISO 15926 we have a Reference Data Library (OWLites would call it a Vocabulary). In this Vocabulary we will have tens of thousands of strongly-typed owl:Classes. We also will have triple stores filled with owl:Things, representing the lifecycle information of, for example, a particular factory. The Template instances (N-ary relations, see [1]) we use for the representation of that information refer to those Reference Data. BUT, we may not refer to what we already have in that vocabulary. Since those templates are owl:Things we MUST refer to other owl:Things, not to owl:Classes. For this unfortunate fact alone we must instantiate all Classes in that vocabulary, that's why I raised the question. Perhaps I should name those instances by suffixing them with '-T' or so, unless you have a better idea. Kind regards, Hans [1] http://smi-web.stanford.edu/people/noy/nAryRelations/n-aryRelations-2nd-WD.h tml _______________________ Hans Teijgeler ISO 15926 specialist www.InfowebML.ws hans.teijgeler@quicknet.nl phone +31-72-509 2005 -----Original Message----- From: Danny Ayers [mailto:danny.ayers@gmail.com] Sent: Friday, December 02, 2005 10:51 AM To: Richard Newman Cc: Hans Teijgeler; Semantic Web Forum Subject: Re: Class and Thing with same ID The W3C Validator [1] is handy for checking syntax. (I've no idea whether the syntax is ok ;-) Re. the model, yes it is allowed to have a a resource be both a Class and a Thing, but as Richard says, takes you from the OWL DL space into OWL Full. That may be what you want, but in my (very limited) experience the majority of OWL Full ontologies can be tweaked into being OWL DL (I suspect leading to a more accurate model in most cases). The advantage is practical - there are several sound & complete OWL DL reasoners available, it's more hit & miss with OWL Full. There's an online demo of the Pellet engine via [2], handy for quick checks of consistency and which OWL species is in use. Cheers, Danny. [1] http://www.w3.org/RDF/Validator/ [2] http://www.mindswap.org/2003/pellet/index.shtml -- http://dannyayers.com
Received on Friday, 2 December 2005 11:57:35 UTC