Re: RIF must cover RDF triples as data --> RDF as top level rule syntax?

Paul Vincent wrote:
> Dave - assuming this was a qu to the whole group...

Hi Paul, yes. Though I should have asked that question in a separate 
email, I've now gone and tangled up the two issues and had been trying 
to keep them separate :-)

> Given one of the classification of rule communities in RIF are semantic
> web and "everything else", what would be the rationale for a non-SW user
> translating a rule to and from RDF tuples? 

That's not quite what I had suggested. The suggestion is in:
http://lists.w3.org/Archives/Public/public-rif-wg/2006Sep/0077.html

I'm not actually suggesting using RDF for the rule expressions but for 
rule metadata and for packaging rules into rulesets. I tried to put the 
rationale in that message but let me expand on the key point.

We seem to want to have rules with metadata annotations (to indicate 
semantics, default behaviours, documentation, etc). That metdata scheme 
also needs to be extensible since RIF itself is extensible. The charter 
calls for us to do this and suggests we consider using RDF for such 
metadata.

I was mostly trying to show that we could use RDF for rule metadata 
quite simply without having to have all the the details of rule 
expressions encoded in find grain triples. To invent our own extensible 
metadata scheme for annotating rules and *not* use RDF would be odd 
unless we had really good reasons and I'm trying to indicate there are 
no such "really good reasons". In fact I'd go further and suggest that 
for a SemanticWeb track working group to invent an extensible metadata 
scheme and not use RDF would be so perverse that we would be unlikely to 
get it past the review process. People may not like RDF as a primitive 
knowledge representation language but the one thing it is really good at 
is extensible metadata.

One other point. In that sketch proposal I retained the head/body 
separation at the RDF level rather than having a single XML structure 
for a rule. One could do it the other way. However, keeping them 
separate seems to me to be one way to allow us to use RDF's 
extensibility to meet our language extensibility goals. For example, for 
production rules one might want to use the same condition language for 
bodies that has already been proposed but have a completely different 
notion of rule heads. The rule heads would become actions (which might 
include "assert" as one of the available actions but plenty more 
besides). With the RDF approach then we can just add a new class of rule 
head but still be able to reuse the rule body piece (and annotate the 
ruleset with a pointer to the production rule semantics). Furthermore 
someone with new sorts of rule head actions could easily extend the 
language to add them. Achieving that same level of extensibility and 
forward compatibility with XML Schema is, I believe, much more work.

> (Indeed I have a follow-up qu: why would the Semantic Web want to
> represent rules in RDF? 

Personally I don't think it does. It could be useful to embed or carry 
rules in RDF so that you could fetch fetch a set of data plus the set of 
rule that apply to it one document but it is not critical. Indeed none 
of the current RDF rule languages really encode their rules in RDF.

I do think it is useful to reuse the Semantic Web representations for 
things they are good at - metadata annotation, classification, 
extensibility.  I was trying to show how to do that.

> I can understand this for "simple" relationship
> rules which need to be shared as semantic information alongside RDF
> "data", but for more complex rules what is the advantage of
> representation in RDF over say XML? As the rule components will only
> make sense in the context of each other, splitting up a complex rule
> into RDF statements surely will not "buy" anything? Or is this to avoid
> mixing RDF with other (eg XML rules such as RuleML) information on the
> web?).

Hopefully the above has answered this. I'm not proposing splitting up a 
complex rule into RDF statements. At most I'm proposing a head/body 
split but (a) as I've tried to indicate I think that does buy us some 
very useful extensibility and (b) if that doesn't fly you could keep 
individual rule as monoliths but still use RDF for the annotation and 
ruleset structure.

Does that all make sense?

Dave

> Apologies if I am missing something in my Semantic Web education, and
> pointers to where I can further educate myself on this matter would be
> appreciated! 
> 
> Cheers
> 
> Paul Vincent, 
> TIBCO - ETG/Business Rules 
> 
>  -----Original Message-----
> From: public-rif-wg-request@w3.org [mailto:public-rif-wg-request@w3.org]
> On Behalf Of Dave Reynolds
> Sent: 11 October 2006 10:38
> To: Christian de Sainte Marie
> Cc: RIF WG
> Subject: Re: RIF must cover RDF triples as data
> 
> 
> Christian de Sainte Marie wrote:
>> Dave Reynolds wrote:
>>> [[[
>>> *RDF Data*
>>>
>>> RIF must cover RDF triples as data where compatible with Phase 1 
>>> semantics.
>>>
>>> Specifically:
>>>
>>> o RIF MUST provide a mapping from bNode-free RDF statements to RIF 
>>> facts so that RIF rules can be applied to (data derived from) RDF
> fact 
>>> bases. This MAY be a "natural" mapping in the sense that RDF facts
> are 
>>> mapped to ground instances of unary and binary RIF predicates or MAY 
>>> be based on a "holds" predicate.
>>>
>>> This mapping will support the derivation of new RDF statements (i.e. 
>>> bNode-free RDF statements can be made in the head of rules).
>> I think that this is what confuses me. Granted, since some rules 
>> interchanged in RIF will apply to RDF data, such rules may contains 
>> statements that RDF triple XYZ holds, as a condition or as a conlusion
> 
>> of a rule; and thus, such statements need be representable in RIF 
>> (covered). Granted, they could be mapped in various way (btw, why
> isn't 
>> embedding RDF/XML an option?).
> 
> If you mean "embed RDF/XML as a means to convey facts in a RIF document"
> 
> then not only is it an option but I've already proposed a start in that 
> direction by suggesting we use RDF/XML as the top level rule syntax.
> [Any other reactions to that suggestion BTW?]
> 
> [PV>] ...
> 
> 


-- 
Hewlett-Packard Laboratories    | Phone: +44-117-3128165
Filton Road, Stoke Gifford      | FAX:   +44-117-3128924
Bristol BS34 8QZ, UK            | dave.reynolds@hp.com

Received on Thursday, 12 October 2006 08:32:10 UTC