Re: Comments on Motion 42 (not 41): Decoration system, revised text
John Pryce replied to my comment:
>> Question at the end of 8.8 (last page):
>>
>> It seems to me that the natural way to define ANY function or operation
>> on compressed intervals that are a decoration dx is to perform it as if
>> it were applied to normalInterval(dx).
>
> AFAICS that looks OK for arithmetic operations. There seems to be a snag
> with intersection and convexHull, in view of the default being that these
> produce a bare result.
I don't see a snag. Good decorations behave like Entire (propagate for
ConvexHull and has no effect for Intersection); bad decorations behave
like Empty and have no effect on ConvexHull (this may annoy some people
but be ok for others), and propagate for Intersection.
If one wants a bad decoration to propagate through ConvexHull one should
use ConvexHullDec. This may be an argument for requiring the two forms
(bare and decorated) for set operations, at least in the "set" flavor.
Michel.
---Sent: 2012-12-23 01:45:49 UTC