RE: [xacml] concrete proposal of condition reference (#7)

>I do feel strongly about it. You may indeed invent extension higher
order
>functions that may take two <Function> elements at arbitrary argument
>positions. Your HigherOrderApply assumes that the higher-order function
>applies exactly one function to its other arguments in a certain way.


Since we have this (awfully unneeded and cumbersome) construct in our
language, I have to reluctantly agree with Polar.  As we accepted this
as a data type we need to address it with expressions in a generic way.

But if I remember correctly, when we argued about higher order function,
an explicit argument was given that their usage will not "spread" and be
confined to the collection of standard functions.   I did not buy that,
and now I have a justification why. :)


Daniel.