home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.lang.scheme
- Path: sparky!uunet!snorkelwacker.mit.edu!bloom-beacon!INTERNET!dont-send-mail-to-path-lines
- From: Alan@lcs.mit.EDU (Alan Bawden)
- Subject: Why is the exact/inexact distinction built into Scheme?
- Message-ID: <9211211945.AA01597@corn-pops>
- Sender: Alan@lcs.mit.edu
- Organization: The Internet
- References: <1992Nov17.182740.11194@cs.brown.edu>
- Date: Sat, 21 Nov 1992 19:45:28 GMT
- Lines: 16
-
- Date: 17 Nov 92 18:27:40 GMT
- From: Mark Johnson <mj@cs.brown.edu>
- ... As I sit waiting for my machine to finish a job that a
- corresponding C program would have completed 15 mins or so earlier, I
- keep wondering what it was that inspired the Scheme language designers
- to build in the inexact/exact distinction.
-
- The Scheme designers are not responsible for the fact that
- `(inexact->exact (round x))' is slow in your implementation.
-
- You don't spell out why it is that you think the inexact/exact distinction
- is responsible for the slowness -- so I have to guess that you think that
- if `round' were defined to return an exact result given an inexact
- argument, then somehow your Scheme compiler would be able to generate
- better code. Well, why doesn't your Scheme compiler generate that same
- code for `(inexact->exact (round x))'?
-