Re: RIF vs Rule Language

On 9 Dec 2005, at 03:12, Jim Hendler wrote:
> At 23:49 +0100 12/8/05, Enrico Franconi wrote:
>> The behaviour of a reasoner has to be constrained by the semantics  
>> of the language it processes, otherwise its behoaviour will be  
>> just arbitrary: that's why we have semantics :-)
>
> well, that's true - but let's not forget there are a number of  
> different kinds of semantics - I don't know of any programming  
> language in major use that has anything other than operational  
> semantics, and they seem to work pretty good (at least this email  
> seems to be getting to you through a whole bunch of computers and  
> routers and programs, and I suspect few if any have a model theory) --
>   this is not to advocate not doing a formal semantics, but just to  
> remind people that an option is to do basically an operational  
> semantics for the RIF in phase 1, and then nail down the details in  
> phase 2 (While people are starting to implement the relatively  
> stable stuff) -- this might be a way to get a lot more out the door  
> sooner -- I wish I had considered that option when I was chairing  
> the WebOnt group - I suspect we'd of gotten to CR sooner if we'd  
> been able to move things this way (i.e. if we'd had the reference  
> manual in LC or close thereto while we were finalizing the details  
> of the semantics)

A may agree that this is a viable way to go for the group: first  
understanding and agreeing on the different behaviours we may find  
relevant, and then studying and standardising the different semantics  
for each of them. I am not sure that phase one may end without any  
semantics though: I think we have to have at least an understanding  
of the different semantics options *before* fixing the final  
deliverable of phase one; otherwise we may have taken some non well  
informed decision which may be painful to fix afterwards.

cheers
--e.

Received on Friday, 9 December 2005 10:33:44 UTC