Re: [xacml] xpath-expression datatype

From
Seth Proctor <>
Date
2004-08-26T15:21:00+00:00
ID
Thread
Re: [xacml] xpath-expression datatype
MHonArc v2.5.0b2 -->


















xacml message






[Date Prev]
 | [Thread Prev]
 | [Thread Next]
 | [Date Next]

--

[Date Index]
 | [Thread Index]
 | [List Home]








Subject: Re: [xacml] xpath-expression datatype




From: Seth Proctor <>
To: Polar Humenn <>
Date: Thu, 26 Aug 2004 11:20:38 -0400








On Thu, 2004-08-26 at 10:53, Polar Humenn wrote:
> It is a "base type", and therefore you will not see it in anything.

What? Because string, integer, or boolean is a basetype, do I also not
see them in anything? No. I see them in functions. I see them in
normative behavior. I see many justifications for why they exist in the
specification.

> > I think it will cause a great deal of confusion to have a datatype that
> > goes unused.
> 
> Why are you confused? The fact that every time you use a specific
> construct, i.e. <Apply>, <AttributeValue>, etc. you are using
> <ExpressionType>, just because you don't see it on paper? It is a function
> of the compiler, not of the user.

I think you should go back and read my original email again. I'm not
talking about the Expression element. I'm fine with that. I'm talking
about a new datatype called xpath-expression.

> > In fact, we have three functions that use XPath
> > expressions, but they define their parameters in terms of strings, which
> > I think will further confuse people.
> 
> Where did this come from? Now we are talking about the arguments to
> expressions. XPath expressions are not in the type system of the XACML
> language. They are an element of XPATH, and defined there. We can only
> have them as strings, unless, you want to bring all of XPATH into XACML.

Again, please re-read my previous email. I think you'll find that we're
in agreement on this point.


seth