Re: [RIF] New diagram with Goals, CSFs, and Requirements and first-order logic

Hi Paula,

I've noticed that the Goal, Critical Success Factors, and Requirements 
document at 
http://www.w3.org/2005/rules/wg/wiki/Goals%2C_Critical_Success_Factors_and_Requirements 
has been modified since Tuesday to explicitly mention first-order logic. 
(There may have some explicit email to this effect which I have not seen.) 


I'm glad to see that there is now an explicit reference to the requirement 
that the Extended RIF support first-order logic.

I believe that it would be even better to have a requirement that the Core 
RIF support first-order logic. I believe that though this might seem to be 
more difficult and more work-intensive in the short term, it would 
actually be more efficient it would clarify many of the existing sources 
of vagueness in the requirements document, such as the section on "No 
surprises" and soundness. In addition, it would lead to more widespread 
adoption, since presumably there will be more implementations of core RIF 
than of Extended RIF.

Nevertheless, I'm willing to live with the compromise that Extended RIF 
will support FOL even if Core RIF doesn't.  If all the other supporters of 
the requirement that RIF support FOL  --- I would guess that Sandro, Chris 
Welty, and cmsa, among others, are in this camp --- then I will not 
further argue for the inclusion of this requirement in Core RIF. (If they 
do not agree, and would still like to argue for Core RIF supporting FOL, 
then I am still up for a good argument on this subject.)

Best regards,
Leora Morgenstern






Paula-Lavinia Patranjan <paula.patranjan@ifi.lmu.de> 
Sent by: public-rif-wg-request@w3.org
05/30/2006 11:09 AM

To
Leora Morgenstern/Watson/IBM@IBMUS
cc
public-rif-wg@w3.org, Frank McCabe <frank.mccabe@us.fujitsu.com>
Subject
Re: [RIF] New diagram with Goals, CSFs, and Requirements: Whatever 
happened to first-order logic as a requirement?






Hi Leora,

The requirement regarding FOL is under the open issues in the text at
http://www.w3.org/2005/rules/wg/wiki/Goals%2C_Critical_Success_Factors_and_Requirements


I agree that the current statement on FOL is not really clear, sorry 
about that, but I just copied and pasted this statement from Frank's 
first version on Design Constraints. I'll reformulate this in a second.

As I said in the email, the text below the diagram is under development 
that is why one might get confused at this point. I will address the 
rest of your comments together with Frank when refining the text on 
RIF's DCs.

Thank you for your comments!

Regards,
Paula
> This comment refers both to your diagram as well as to the text 
> following the diagram in the url which you reference below.
>
> Whatever happened to first-order logic as a requirement? This was a 
> requirement that Sandro had originally proposed, and which I 
> championed. Instead, what I see is something considerably weaker: a 
> notion of "no surprises" which is said to be equivalent to some kind 
> of soundness, and a statement saying that "The RIF must support a 
> substantial portion of First-Order Logic." 
>
>
> Not only is this weaker, it doesn't seem very clear. What precisely is 
> this soundness to which you refer? How do you define soundness without 
> the notions inherent in first-order logic, namely truth within a model 
> and derivability?
>
> What precisely is a "substantial portion" of first-order logic? As it 
> stands, the phrase isn't well-defined. Could one have first-order 
> logic with everything but modus ponens? One could argue that that is a 
> "substantial portion" of first-order logic, but obviously that 
> wouldn't make sense. What do you plan to leave out of first-order 
> logic? And why?
>
> Can we put this back on the table, and get first-order logic into the 
> requirements?
>
> I hope we can discuss this at our meeting today.
>
> Best regards,
> Leora Morgenstern
>
>
> *Paula-Lavinia Patranjan <paula.patranjan@ifi.lmu.de>*
> Sent by: public-rif-wg-request@w3.org
>
> 05/26/2006 07:47 AM
>
> 
> To
>                public-rif-wg@w3.org
> cc
>                Frank McCabe <frank.mccabe@us.fujitsu.com>
> Subject
>                [RIF] New diagram with Goals, CSFs, and Requirements
>
>
>
> 
>
>
>
>
>
> Hi,
>
> Frank and I have merged the requirements proposed in
> http://lists.w3.org/Archives/Public/public-rif-wg/2006May/0234.html
> and the goals, CSFs, and requirements of the diagram proposed by Frank 
in
> http://lists.w3.org/Archives/Public/public-rif-wg/2006May/0188.html
>
> The updated version of the DC diagram can be found under
> 
http://www.w3.org/2005/rules/wg/wiki/Goals%2C_Critical_Success_Factors_and_Requirements

>
> Note that the text on goals, CSFs, and requirements for RIF on the above
> given page doesn't correspond yet to the current version of the DC
> diagram; we are working on updating the text on the wiki page too.
>
> Please send comments on the current version of the DC diagram so as to
> be able to finalize the work on RIF's design constraints as soon as
> possible.
>
> Regards,
> Paula
>
>
>