open issues in RIF-BLD, RIF-RDF: metadata, built-ins

Dear all,

Besides the issues regarding identifiers in RIF mentioned in [1] and
discussed in the ensuing thread, I have uncovered a number of issues
with the RIF and the RIF-RDF compatibility which need to be resolved.

i) Metadata in RIF / referring to external data sets, RDF
===
it has long been recognized that RIF needs a mechanism for specifying
metadata.  In addition, the current proposal for referring to external
data sets and data models [2] relies on metadata for this reference.
However, the BLD language does not have it.  I propose adding adding
metadata to the language as soon as possible.
Additionally, the working group needs to agree on the mechanism and
vocabulary to use for referring to external data sets. Personally, I am
fine with following the proposal in [2]. Fortunately, this is also on
the agenda for today's telephone conference :-)

I really think these issues need to be resolved before the publication
of the next working draft.

ii) Yes/no on specifying the entailment regime for RDF graphs
===
In my original proposal for a vocabulary for referring to RDF graphs
[3], I proposed to specify the desired entailment regime explicitly.
There was some concern about this explicit specification (see [4] and
the following thread).
RDF and OWL leave it up to the user to decide which entailment regime to
use.
>From a theoretical point of view it does not matter whether the language
allows to specify the suggested entailment regime to use; explicit
specification of the entailment regime is only for practical purposes.
The working group needs to decide whether or not RIF will suggest a
vocabulary for describing entailment regimes, and whether the
specification of an entailment regime is a suggestion or a prescription.

This is not necessarily an issue for the next working draft, but should
be resolved in the not-too-distant future.

iii) Whether and if so how to relate to RDF datatype entailment
===
The RDF semantics specification additionally defines the notion of
datatype entailment, which is a non-normative extension of RDFS
entailment. The question is whether we also want to deal with this when
talking about RDF compatibility; otherwise, we simply use the RIF notion
of datatypes. Alternatively, such an extension, and/or the relationship
between the treatment of datatypes in RIF and in RDF datatype entailment
might be subject of some non-normative appendix.
I personally do not think we need to relate to datatype entailment, at
least in phase 1.

iv) Details on built-ins in BLD
===
More details are required about the built-ins which are supported by
RIF.  There is currently a list of proposed functions and operators [5];
however, it is not really clear what the status is of this list in RIF.
 Furthermore, I expect that we would need further built-ins, such as
unary predicates which corresponds to the data types supported by RIF
and RDF-specific built-ins as mentioned by Dave in [6].

It would be good if we could advance on this issue for the upcoming
working draft, but I guess it would be acceptable if it were postponed
until the next working draft.


Best, Jos

[1] http://lists.w3.org/Archives/Public/public-rif-wg/2007Sep/0001.html
[2] http://www.w3.org/2005/rules/wg/wiki/Arch/Data_Sets
[3]
http://www.w3.org/2005/rules/wg/wiki/Arch/RDF#head-cb9271092ef7391c51020efbc1900dfc64edbd08
[4] http://lists.w3.org/Archives/Public/public-rif-wg/2007Jul/0030.html
[5] http://www.w3.org/2005/rules/wg/wiki/List_of_functions_and_operators
[6] http://lists.w3.org/Archives/Public/public-rif-wg/2007Jul/0038.html
-- 
Jos de Bruijn            debruijn@inf.unibz.it
                      http://www.debruijn.net/
----------------------------------------------
As far as the laws of mathematics refer to
reality, they are not certain; and as far as
they are certain, they do not refer to
reality.
  -- Albert Einstein