From mossakow at iws.cs.uni-magdeburg.de  Sat Jan 23 11:32:20 2016
From: mossakow at iws.cs.uni-magdeburg.de (Till Mossakowski)
Date: Sat, 23 Jan 2016 11:32:20 +0100
Subject: [ontoiop-forum] DOL constructs
In-Reply-To: <56A2686C.8000703@mie.utoronto.ca>
References: <56A2686C.8000703@mie.utoronto.ca>
Message-ID: <56A356B4.4020206@iws.cs.uni-magdeburg.de>

Dear Michael,

I cc the OntoIOp list because this may be of general interest.

Am 22.01.2016 um 18:35 schrieb Michael Gruninger:
> Hello Till,
> one of my students has some questions about DOL,
> and I was not sure about the answer.
>
Indeed, the differences between the different DOL constructs for getting 
rid of a part of a theory/OMS are a bit subtle, so your questions are a 
good chance to illustrate them.
> Is it correct to say that the extract construct in DOL only
> extracts a module from the ontology, where a module
> is a subtheory that is conservatively extended by the original ontology?
yes
> Furthermore, is the module a subtheory or is it only a subset of the
> axioms of the original ontology?
>
it is a subset of the axioms. If you want a subtheory, use "forget" or 
"keep". This gives you a uniform interpolant. However, it is 
computationally much harder to actually compute it.
> If the above is correct, how can one extract some subtheory from
> an ontology such that the subset is not a module?
> In other words, the extracted subtheory is nonconservatively extended
> by the original ontology.
>
You can use "select"/"reject" to throw out some axioms from a theory. 
With "select", you select what should stay, with "reject", you select 
what is thrown out.
You have to explicitly give the list of axioms that is thrown out (or, 
in the case of "select", is kept), because you do not have a criterion 
like conservativity.

Best, Till
> thank you,
> michael