home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!sunic!dkuug!diku!torbenm
- From: torbenm@diku.dk (Torben AEgidius Mogensen)
- Newsgroups: comp.lang.scheme
- Subject: Re: R^4RS Authors Comments on Dylan
- Message-ID: <1992Aug17.090252.24973@odin.diku.dk>
- Date: 17 Aug 92 09:02:52 GMT
- References: <9208132109.AA16509@peanut.crl.dec.com>
- Sender: torbenm@freke.diku.dk
- Organization: Department of Computer Science, U of Copenhagen
- Lines: 20
-
- jmiller@crl.DEC.COM writes:
-
- >o Dylan's use of generic functions, derived directly from CLOS, fit
- > into the language in a particularly elegant manner. If the Scheme
- > community does decide to adopt a modest object system, we should
- > study Dylan's subset of CLOS to make an informed choice between this
- > approach and the more traditional object-dispatch mechanism.
-
- I think the traditional object-dispatch mechanism is more in the
- scheme style of using higher-order functions. When reading the Dylan
- report, it seemed to me that the generic function mechanism was a hack
- to make object oriented dispatch first order. Furthermore, method
- selection occurs at application time, where it will occur at
- class-creation time with the traditional method. This gives more
- efficient dispatch. The only advantage I can see of the generic
- function approach is that you don't need a field in the objects for
- methods. This need, however, only be a single pointer to the class
- definition.
-
- Torben Mogensen (torbenm@diku.dk)
-