Boost logo

Boost :

From: David Bergman (davidb_at_[hidden])
Date: 2002-08-30 12:17:04

The only real discriminating factors of the Spamjunk Set Library, w.r.t
structures such as std::set, are

1. Elements are coded (see Goedel's Coding Theorem...) as integers,
being enums or
   other integerizable elements
2. There is an explicit universal set for each set application, to which
all sets are

These discriminators should affect the naming, not arbitrary "cool, and
not used in Boost" terms.

If this seems reasonable, lexemes such as "pascal", "scalar" or "binary"
should not occur anywhere in the name. I mean, in what way are the sets
more "scalar" than other sets (such as std::set sets)? Or, more
"binary"? And, although Pascal influenced the creation of a very good
language, to which a lot of us adhered at one point, why should we bring
him into this set discussion...

What we have to play with, according to (1) and (2) above are lexemes
dealing with "selection of a universal set" and "coded as integers". No
more, no less. We can have different tastes of the exact names, such as
me choosing "constraint" and "domain" for the terms in (1) and just
skipping (2) altogether, since the Finite Domain Constraints do deal
with integers, normally...

Anyhow, suggestions such as "selection_of", "domain" (with the Finite
Domain semantics, rather than the algebraic semantics),
"constrained_set" are all at least logical. Alternatives such as
"finite_set" are not, since they do not reveal the discriminating nature
of this structure (these sets are no more finite than many other sets,
as pointed out: 2^32 is not extremely finite in terms of modern
computers, unfortunately; I wish it was).


-----Original Message-----
From: boost-bounces_at_[hidden]
[mailto:boost-bounces_at_[hidden]] On Behalf Of Bohdan
Sent: Friday, August 30, 2002 11:20 AM
To: boost_at_[hidden]
Subject: [boost] Re: set class

I think that old pascal_set name brings more information to user than
But if there is some names competition: slot_set, scalar_set,


<spamjunk_at_[hidden]> wrote in message
> I guess I'll just keep picking names, until the protests stop. :)
> This
time I'm
> going with "finite_set."
> Anyone have any problems with the code? Do you see any portibility
> Thanks,
> Rich Herrick
> _______________________________________________
> Unsubscribe & other changes:

Unsubscribe & other changes:

Boost list run by bdawes at, gregod at, cpdaniel at, john at