Re: [SWC] comments on frozen draft (19oct07)

Igor,

Thank you very much for your comments.  Find replies in line.


> 1) Sec.1 improved my overall understanding a lot.
> However, I still find that many things are missing
> (mostly motivation, overall picture) to make a link
> between the title (How to use RIF+RDF?) and very technical
> details about the semantics.

The main technical thing which is missing is how RDF graphs or documents
are referred to from RIF rule sets/documents.  This is, however, still
an open issue.

But apart from that, I will try to improve the motivation.  It would be
helpful if you could mention specific text which you feel is missing.

> 
> 2) Aparently we are dealing with two cases:
> a) there is a RDF-aware rule system
> b) the rule system is not RDF-aware
> The main text deals with the first case (combination), and the
> appendix deals with the second (embedding).
> It would be useful to reference some RDF-aware rule systems.
> Why to treat the two cases differently? Should both be "normative"?

We are only dealing with one case: the exchange of RIF rule sets which
talk about/refer to RDF graphs.
The appendix can be seen as an implementation guide.
There is a note on the wiki about the status of this appendix: it is
currently not clear whether it should be part of this document or not.
I see that the note did not make it into the editors draft; I will move
it to a suitable location.

> 
> 3) How does Sec.2 help a user of a RDF-aware rule system to exchange
> rule via RIF? I'm not sure if I got it.
> Apparently, it defines the conditions that the rule system
> must satisfy to be compliant. How does a user of a specific system
> go to check this?

This section defines the semantics of combinations of RIF and RDF. A
user can check whether or how he can use such combinations by comparing
his semantics with the semantics of his system.
This is analogous to the exchange of rules using RIF BLD; the user of a
rule system will have to check whether the semantics of his system is
compatible with the semantics of RIF BLD.

> 
> 4) Sec.2.1.
> To improve the readability and for didactic reasons, I would suggest
> to reverse the order:
> - start with the last paragraph

OK

> - define the three RIF-RDF names mappings (in boxes)
> - and then deal with all the exceptions and special cases

I think I can tweak the order little bit.  However, much of the text is
required for defining ill-typed literals, which is in turn required for
defining correspondence between literals in RDF and symbols in RIF.

In any case, I will add some textual description to the beginning of the
section, to explain the reader what is going to happen, before the
formal definitions.

> 
> 5) Sec.2.2.
> Do we really need to deal with combinations for the 4 kinds of
> RDF interpretation? Wouldn't the D-interpretation suffice?

It depends.  It has not yet been decided whether or how to support
multiple entailment regimes.  I think D-entailment would probably
suffice, but there might be disagreement about this in the working group.

> Again, for didactic reasons, show the essence (the meat) of one,
> and the rest as special cases...

OK

> 
> 6) Sec.4.1.
> Embeding symbols seems closely related to RIF-RDF mappings from
> Sec.2.1. I am missing the large picture here.

The mappings between symbols in section 2.1 are rather for illustrative
purposes, and should probably be removed; at least, it should be
explained in a different way.
Note that the mappings in section 2.1 are about establishing
correspondence between symbols in RDF and symbols in RIF; note that this
correspondence cannot be established for all symbols (e.g. ill-typed
literals). The embedding in section 4.1 is about the transformation of
RDF symbols to RIF symbols; this embedding is defined for all symbols.



Best, Jos

> 
> Regards,
> Igor
> 

-- 
Jos de Bruijn            debruijn@inf.unibz.it
+390471016224         http://www.debruijn.net/
----------------------------------------------
If you live to be one hundred, you've got it
made. Very few people die past that age.
  - George Burns