Re: A bit on the formal semantics

On Sep 15, 2006, at 3:43 PM, Anne Anderson wrote:

> Just a caution on making formal semantics normative: The OASIS  
> XACML 2.0 specification used Haskell to specify the semantics of  
> our higher order functions.  When the specification was submitted  
> to the ITU for cross-standardization, there was no approved  
> standard for Haskell that we could reference.  In order to satisfy  
> ITU requirements, we had to make the Haskell description non- 
> normative and the English text description the normative version.

First, the WG voted against doing this so it's a bit moot.

Second, I would do it directly, in Englihs. Formal doesn't mean not  
in English :)

Third, if I were going to do it by reference to another computer  
language I would do it by translation to OWL. But I wouldn't (and  
didn't) propose this because I'm not sure future extensions could all  
be translated into OWL.

So, I don't think that would have been a problem :) If the informal  
English is good enough, then the formalizing English should be good  
enough.

Cheers,
Bijan.