Some comments on the defintions in
Section 4.
In general, I think the definitions would read better if they
started with an indefinite article ("a" or "the", as appropriate).
In my experience, this is always done in English definitions. The
extra meaning from the choice between "a" and "the" can be
significant. E.g.
OMS
a set of expressions ...
signature
the set of all non-logical symbols of an OMS
Some particulars
Section 4.1
1. The definitions of OMS and OMS language are slightly circular.
Does any language with a formal, declarative, logic-based
semantics (plus non-logical annotations) qualify as an OMS? Is
there something extra in the requirement (paraphrased) that it be
"used in the formal specification of sets of expressions in
itself"?
Also, in the list of Examples, is OWL really a single OMS
language? There are two semantics (Direct and RDF-based).
2. non-logical symbol: this definition depends on a certain
interpretation of "atomic" (as opposed to its use in "an atomic
sentence
" http://en.wikipedia.org/wiki/Atomic_sentence). All uses
of "atom" through the document appear to be consistent with this
usage, in the sense of indivisible. It is perhaps worth a footnote
to
clarify.
3. the definition of "signature" does not account for partitioning
of the non-logical symbols. Elsewhere, I see signatures being
defined as a structure consisting of several sets. There is a
certain use of signature (as in "a signature-free syntax", when
used to describe syntaxes like CL ) that depends on this narrower
sense of "signature". This is particularly relevant if one is to
say "The sentence A conforms to the signature S." When a sentence
is conformant to a signature, is it also necessary that the
structure of the signature be respected?
This definition of signature is pertinent to the phrase "signature
of O1", where O1 is an OMS. Elsewhere I have seen the usage "the
signature of L1" where L1 is an OMS language (e.g.
http://en.wikipedia.org/wiki/Signature_%28logic%29). In this
document, I do not see any usage of signature in this sense. This
is perhaps worth a footnote to clarify.
4. In model, certain phrases (model of an OMS) are left to be
defined in the Note. Is "model of a signature" is assumed to be
the primary usage?
Section 4.5
1. union: aggregation
of several OMSs
to a new OMS where (only)
identically-named non-logical symbols of the involved OMSs are identi�ed
"Identified" is being used in a non-typical sense here, and not in
the sense used in the definition of "resource" in section 4.2. If
I am reading this correctly, this is saying that the signatures of
the new OSMs is the union of the involved OSMs, without any name
mapping.
2. Similar concern for "combination"
Section 4.6
profile:
sublanguage of an OMS language that targets speci�c applications
or reasoning methods
I'm not sure this definition is compatible with the example of
OWL. Is OWL a single OMS language (there is a different
semantics for OWL 2 Full and the other
profiles, I believe).
Section 4.8
1. heterogeneous OMS. OMS whose parts are supported by different
logics.
Note: At first glance, this seems to contradict the definition of
OMS (which is a set of expressions in a given OMS language).
However, in looking at Annex C, I see the graph where various
profiles of OWL are supported by different logics.
Tara
On 3/10/14 11:35 AM, Till Mossakowski wrote: