Meta-level requirements for RIF

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Meta-level requirements for RIF

Francois Bry

Dear All,

This is a try to clarify issues discussed today as well as again and
again in the past. In my opinion, we rurgently need an agreement on the
following:

1. Data (with their semantics) RIF is to access. In my opinion: RDF, XML
and OWL -- and I would stronlgy suggest Topic Maps as well.

2. What means "rule" for RIF. In my opinion, "deduction rules",
"normative ruyles" (aka integrity constraints), and "reactive rules" (or
ECA rules = Event-Condition-Action rules).
ECA rule include but are not limited to production rules. Events are
needed for an exchange of reactive behaviour between nodes in a
distributed context like the Web.

3. Semantics requirement for RIF: In my opinion, a declarative semantics
is needed in the first place but no procedural semantics. This would
leave interesting questions like termination (in certain cases) open. In
my opinion, many of these questions could be resolved without specifying
a full-fledged procedural semantics. And if some questions are left
open, the RIF would still be very usefull in practice.

Fran├žois



Reply | Threaded
Open this post in threaded view
|

Re: Meta-level requirements for RIF

Sandro Hawke


I think you meant to send this to [hidden email], Francois.

   -- Sandro

> Dear All,
>
> This is a try to clarify issues discussed today as well as again and
> again in the past. In my opinion, we rurgently need an agreement on the
> following:
>
> 1. Data (with their semantics) RIF is to access. In my opinion: RDF, XML
> and OWL -- and I would stronlgy suggest Topic Maps as well.
>
> 2. What means "rule" for RIF. In my opinion, "deduction rules",
> "normative ruyles" (aka integrity constraints), and "reactive rules" (or
> ECA rules = Event-Condition-Action rules).
> ECA rule include but are not limited to production rules. Events are
> needed for an exchange of reactive behaviour between nodes in a
> distributed context like the Web.
>
> 3. Semantics requirement for RIF: In my opinion, a declarative semantics
> is needed in the first place but no procedural semantics. This would
> leave interesting questions like termination (in certain cases) open. In
> my opinion, many of these questions could be resolved without specifying
> a full-fledged procedural semantics. And if some questions are left
> open, the RIF would still be very usefull in practice.
>
> Fran├žois
>
>

Reply | Threaded
Open this post in threaded view
|

Re: Meta-level requirements for RIF

Adrian Walker
In reply to this post by Francois Bry

Francois --

At 06:43 PM 4/18/2006 +0200, you wrote:
>Semantics requirement for RIF: In my opinion, a declarative semantics
>is needed

Agreed.

However, there is a problem.  W3C folks seem to lean strongly towards open
world, whereas database (and most other) folks mostly use closed world and
negation as failure.

In OW, if Adrian is not in the list of people who work for W3C, you cannot
conclude that Adrian does not work for W3C.  In CW you can.

                                  Cheers,



Internet Business Logic (R)
Executable open vocabulary English
Online at www.reengineeringllc.com
Shared use is free

Adrian Walker
Reengineering
PO Box 1412
Bristol
CT 06011-1412 USA

Phone: USA 860 583 9677
Cell:    USA  860 830 2085
Fax:    USA  860 314 1029




Reply | Threaded
Open this post in threaded view
|

Re: Meta-level requirements for RIF

Francois Bry

Adrian Walker wrote:

> Francois --
>
> At 06:43 PM 4/18/2006 +0200, you wrote:
>> Semantics requirement for RIF: In my opinion, a declarative semantics
>> is needed
>
> Agreed.
>
> However, there is a problem.  W3C folks seem to lean strongly towards
> open world, whereas database (and most other) folks mostly use closed
> world and negation as failure.
>
> In OW, if Adrian is not in the list of people who work for W3C, you
> cannot conclude that Adrian does not work for W3C.  In CW you can.
>

THis refers to nonmonotonic negation which is for the second phase.

If RIF is to be used with and as an extension to RDF, then "closed world
assumption" and "nonmonotonic negation" will be necessary -- whatebver
W3C folks lean strongly towards.

Francois