Re: Metadata in CORE (was Re: [Admin] Agenda for RIF telecon May 8)

Gary Hallmark wrote:
> 
> -1 for RDF.  My rules product supports Java data and XML data but not 
> RDF data.  In fact, we don't have any RDF capability in the "tech stack" 
> for Oracle Business Rules  (although there are other Oracle products 
> that do).  Mandating RDF in Core will raise the implementation cost of 
> RIF translators for existing rule systems that otherwise do not use RDF 
> to unacceptably high levels.

We should separate the question of using the RDF data model as the 
structure for the metadata from the question of serialization syntax.

To me the minimal requirement is to use the RDF data model. Adopting a 
simplified serialization syntax to remove barriers to uptake would be 
reasonable.

To use the RDF data model at a minimum we just need:
    - to structure metadata assertions as triples
    - URIs for the subjects of the metadata assertions (i.e. rules and
      rules sets, we've already agreed that)
    - URIs for the metadata predicates (I hope this is uncontroversial
      and allows us to reuse existing vocabulary like rdfs:label,
      rdfs:comment, dc:creator, dc:date etc)
    - plain strings, xsd-typed literals and URIs for objects of
      metadata assertions (for example, we Could define URIs to denote
      the official RIF dialects)

My thumbnail sketch of "Data Sets" [1] also made use of bNodes and I 
would argue for supporting those.

The question of how we attach that metadata to RIF files is separate. It 
is true that my original suggestion from 8 months ago proposed using the 
RDF/XML syntax but if that is an adoption problem then there are other 
alternatives:

  o Encode the metadata in a separate section of the RIF file using RIF 
syntax as Sandro suggests.

  o Encode the metadata in a separate section of the RIF file using a 
simplified (XS-friendly) XML encoding with a well defined mapping to the 
RDF data model.

  o Attach the metadata to relevant RIF XML elements (Ruleset, Rule, 
Variable etc) using a simplified (XS-friendly) XML encoding with a well 
defined mapping to the RDF data model.

For the latter two the mapping to the RDF model could be part of our 
specs or could be explicitly encoded in the RIF XML files using the new 
GRDDL spec [2].

  I very much favor having meta-data.  We
> should have a small set of standard properties like ruleset name, rule 
> name, modification/translation history, etc. and allow additional 
> (non-standard) properties as well.  This should be specified using the 
> same abstract syntax notation used in the Core and should end up as both 
> a human readable syntax and more importantly as part of the Core XML 
> schema.

Dave
-- 
Hewlett-Packard Limited
Registered Office: Cain Road, Bracknell, Berks RG12 1HN
Registered No: 690597 England

[1] http://www.w3.org/2005/rules/wg/wiki/Arch/Data_Sets
[2] http://www.w3.org/2001/sw/grddl-wg/

> 
> Dave Reynolds wrote:
> 
>>
>> Christian de Sainte Marie wrote:
>>
>>> - Meta-data: how to satisfy the "meta-data" requirement in RIF Core?
>>>   * Open meta-data field? Each dialect defines a closed set of 
>>> meta-data fields? RDF? Others?
>>
>>
>> Just a suggestion, but perhaps when new topics like this are going to 
>> be tabled it might be worth having a few days notice so people can 
>> prepare.
>>
>> By way of background I proposed use of RDF for this (now there's a 
>> surprise :-)) some time ago:
>>   http://lists.w3.org/Archives/Public/public-rif-wg/2006Sep/0077.html
>>
>> Then used that suggestion as the syntax in the worked example for Use 
>> Case 8:
>>   http://www.w3.org/2005/rules/wg/wiki/UC8_Worked_Example
>>
>> Then showed how this could be applied to relevant, open ended, 
>> metadata annotations:
>>    http://www.w3.org/2005/rules/wg/wiki/Arch/Data_Sets
>>
>> Dave
> 
>

Received on Thursday, 10 May 2007 08:32:48 UTC