[
Lists Home |
Date Index |
Thread Index
]
Amelia A.Lewis scripsit:
> Urk. Right. Bob Foster points out, in another email in this thread,
> that there are ways of asking the plugin "are you sortable?". But
> clearly my proposal here is too facile.
This led me to consider what rationalized constraining facets would
look like. On the syntactic side, pattern does it all; the RNG rule that
multiple patterns are ANDed is IMHO superior to the WXS rule that they
are ORed.
On the value side, we need to know whether a type is sortable, as you
say. If so, then maxInclusive, maxExclusive, minInclusive, and minExclusive
are necessary and sufficient.
Then there needs to be some notion of whether a a type has an exact
or only an inexact representation. In the latter case, precision is
a suitable constraining facet: for example, an inexact representation of pi
like 3.1415 could be constrained to have .0001 precision.
What do you think?
--
In politics, obedience and support John Cowan <jcowan@reutershealth.com>
are the same thing. --Hannah Arendt http://www.ccil.org/~cowan
|