[
Lists Home |
Date Index |
Thread Index
]
> Indeed, if I correctly understand the literature I've read on
> the subject
> (and I may not, as much of it is too mathematical for me),
I get that feeling when I look at my old college math books. I can only
hope those brain cells were put to good use elsewhere.
> this demand is
> worse than hard: it's impossible in principle.
In the specific case I gave, it surely isn't "impossible in principle".
I'd accept that it's impossible in the general case. How pathological do
the general cases have to be, though?
From the specification standpoint, it's accepted that one shouldn't make
requirements of a validator that can't be guaranteed of being performed in
reasonable and timely fashion. How can that be taken that to mean that such
an engine is absolved from making at least a token effort to ensure that key
fields are required by the content model?
|
