home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!uknet!edcastle!dcs.ed.ac.uk!pdc
- From: pdc@dcs.ed.ac.uk (Paul Crowley)
- Newsgroups: comp.object
- Subject: Re: Object hidden state and side effects
- Message-ID: <BzCv32.Mp3@dcs.ed.ac.uk>
- Date: 16 Dec 92 14:04:14 GMT
- References: <1992Dec11.163449.18439@crd.ge.com> <1992Dec13.152136.16852@hasler.ascom.ch> <1992Dec14.112222.13987@kei.is.s.u-tokyo.ac.jp>
- Sender: cnews@dcs.ed.ac.uk (UseNet News Admin)
- Reply-To: pdc@dcs.ed.ac.uk (Paul Crowley)
- Organization: Edinburgh University
- Lines: 13
-
- Are people arguing that whichever set of integers or floats happen to
- be supported by the hardware should be treated as a special case, or
- that any immutable object should be? If the former, I find this level
- of hardware specificity in the name of language purity curious. If the
- latter, then to the extent that it's sensible to do you can do it in
- Eiffel: use objects that support functions but not procedures. It would
- seem sensible, for example, to support bignums in this way. I think the
- Eiffel way of allowing functions to change internal state if they
- promise not to change external behaviour is the right one; asking the
- language to enforce guarantees would be very limiting.
- __ _____
- \/ o\ Paul Crowley pdc@dcs.ed.ac.uk \\ //
- /\__/ "I'm the boy without a soul." \X/
-