Boost logo

Boost :

From: David B. Held (dheld_at_[hidden])
Date: 2002-08-28 19:49:18


"Dave Gomboc" <dave_at_[hidden]> wrote in message
news:Pine.LNX.4.44.0208281646430.21018-100000_at_boyne.cs.ualberta.ca...
> [...]
> I'm sure the originator of the proposed name (not myself) didn't pick
> the term out of left field. Would you not agree that the union of all of
> the possible values that can be assigned to the structure (whatever
> its name) is the power set of the individual items represented by the
> bits?
> [...]

No, I would not agree, though I may be totally off-base for doing so.
I would expect a "power set" to include all the subsets of a set, and
offer some way to access them. I don't see the proposed library as
offering this capability. To take a simple example:

Set S:
{0, 1, 2}

Power set P(S):
{{}, {0}, {1}, {2}, {0, 1}, {0, 2}, {1, 2}, {0, 1, 2}}

Now, while the set library in question can represent *each* of the
elements of the power set, as you say, I don't see it as *being* the
power set. I suppose I would expect an actual "power set library"
to perhaps offer, say, iterators for all the 0-th order elements, all
the 1st order elements, etc. I think "domain" is more mathematically
accurate, though perhaps not as obvious to programmers as is
desired. There should be a way to characterize the various set types
being proposed and in existence, and probably name them
according to the major discriminating features.

Dave


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk