RIF: "lingua franca" or "esperanto"?

Dear All,

Frank McCabe wrote:

>Notwithstanding some of the technical issues raised, I think that  
>hoping (or even requiring) yet another lingua franca is a hopeless  
>endeavor. For a number of reasons:
>1. There is already too much content out there that would need  
>'translating' from what would be 'legacy' notations to the lingua  
>franca. That translation effort is probably not going to happen, and  
>to base the RIF on requiring it is silly.
>  
>
The "lingua franca" analogy was not aimed at suggesting:

1. RIF should introduce its own, novel notations where widespread 
notations are available.

2. RIF shpould match its own notations to those of other, already 
existing languages.

In fact, the "lingua franca" analogy means exactly the opposite of 1: If 
a word/notation is widesrpead, then it has good chances to belongs to 
the lingua franca, but must not.

>2. Each system that has adopted a particular notation/semantics/meta  
>model has, presumably, good reasons for making that choice. But,  
>there is no single semantics that will cover all possible choices; so  
>you will disenfranchise a lot of people - no matter what your choice.
>  
>
I do not think so. I beleive that RIF only has two ways:

- Either to define a full-fledge novel language -- including a 
procedural semantics. My objections to this are: (1) it requires much 
more time than the RIF WG has scheduled, (2) even if the language 
eventually derfined is great, there is no guaranty that it becomes a 
standard for interchange.

>3. Ownership of information is paramount in the Internet. This  
>includes the semantics as well as the axiom set (i.e., some people  
>may want to live inside that 'blast radius' of fuzzy logic:)
>  
>
Ownership of information is paramount in the Internet. Making information is also paramount in the Internet. Users are free to use/re-interpret information availalbe as they like. 

>4. Creating a new lingua franca (or even lingua über toutes to mix  
>languages) sets us up for another fight when someone decides that yet  
>another language is going to be the LF.
>  
>
The analogy "lingua franca" does not mean "lingua über toutes" (over all other languages). Over is wrong. "lingua franca" means exactly an "interchange" language, a language with a simplified grammer in addition to the mother tongues of its users. Understand the analogy as follows: rule languge withouut a full-fledge procedural semantics in addition to the several different rule languages and engines that are in use. 

>I believe that laying the foundations for rules interoperability -  
>even in the face of semantic heterogeneity - is a more appropriate task.
>Off the cuff, this suggests that
>1. the original semantics of a rule set be exposed in any encoding of  
>the rule set in RIF
>2. a technique for 'communicating' between different logics is a key  
>technology for the RIF
>  
>
Objections:

1. Achieving this requires more time than the RIF WG has.
2. The appropriate analogy for a RIF of that kind would be "esperanto".

There is a very simple reason why RIF should be a "lingua franca" but no 
"esperanto": a "lingua franca" grows up from practice and is widerly 
used. "esperanto" might be very well designed but in spite of its 
quality only a very small group of afficionados use it.

I think Frank's viewpoint has considerably helped to precise the issue: 
Should RIF be a lingua franca or an esperanto.

>3. a technique for preserving the provenance of information during  
>inference would also be required.
>  
>
Agreed.

Regards,

Francois

Received on Friday, 27 January 2006 10:15:47 UTC