Re: [xacml-comment] C003 and matching in targets and conditions

[Polar Humenn <polar@syr.edu>]

On Mon, 25 Nov 2002, John Merrells wrote:

>
> Yeah, I understand A12 now. Explicitly stating that the Match element
> means that the match function and match arguments should be rewritten
> as (any-of match-function primitive bag<primitive>) would help clarify
> this section.
>
> I seem to be failing to persaude you that the similarity of the MatchId
> and FunctionId attributes, but with subtlely different semantics, is going
> to confuse people, so I'll give up for now and wait to see if anyone else
> trips over the same problem.

I think we are leaning toward changing the schema so that the attribute
designator/selector comes as the second element of a <*Match> element.
This would sort of make <*Match MatchId="f"><Value/><*Designator/> equal
to (any-of f value designator).

> BTW: I was curious to find Haskell expressions in the spec. I haven't
> seen it in the wild since ~1990... from Edinburgh isn't it?

Yep, now maintained by Yale. Still pretty active.

> Out of curiosity was OCL (the Object Constraint Language) considered for
> writing the formal notation?

We considered OCL for the combining algorithms, i.e. the firt run at
desiring some "code" to sure up the semantics. However, some wanted
something non-vendor specific (i.e. Java, VB) nor non-platform specific
(i.e. C,C++).  As for OCL, I think that some just generally didn't like
it. So, we ended up doing some psuedo code without formal semantics for
the combining algorithms.

For the other functions, of which I took to writing, I used Haskell,
because I'm familiar with it, and it does a declarative specification
Also, its type system is pretty robust to handle the higher order stuff.

Cheers,
-Polar

> John
>
>
>
>
> ----------------------------------------------------------------
> To subscribe or unsubscribe from this elist use the subscription
> manager: <http://lists.oasis-open.org/ob/adm.pl>
>