Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!seismo!sundc!pitstop!sun!amdcad!ames!ucbcad!zen!ucbvax!ECNCDC.BITNET!MSRS002
From: MSRS002@ECNCDC.BITNET ("THE DOCTOR.")
Newsgroups: comp.lang.modula2
Subject: Re:Re: Sets.
Message-ID: <INFO-M2%87102620254979@UCF1VM>
Date: Mon, 26-Oct-87 17:04:00 EST
Article-I.D.: UCF1VM.INFO-M2%87102620254979
Posted: Mon Oct 26 17:04:00 1987
Date-Received: Thu, 29-Oct-87 04:58:45 EST
Sender: daemon@ucbvax.BERKELEY.EDU
Reply-To: Info-Modula2 Distribution List <INFO-M2%UCF1VM.bitnet@jade.berkeley.edu>
Organization: The ARPA Internet
Lines: 15

Hmmm, perhaps the BitSet should be defined to be a word size, and uses
exclusively for bit twiddeling and implementation specific things.  Then,
I think we would rather need a more abstract set construct.  I still think
the compiler should allocate as many words as necessary to hold the range
which should be expected in the set ( boy, that made a lot of sense ), and
the limit should be infinity, or at least a rather large approximation of
infinity.  The maximum cardinal value supported by the implementation might
be a reasonable approximation to infinity.  If a cardinal is 16 bits, then
such a set, SET OF CARDINAL would only take up 8k bytes, which isn't too
extreme, not even for my old 8 bit system.

I hope the forthcomming standard says something reasonable about sets.

                                                    Tom Ruby
                                                 MSRS002@ECNCDC