Related articles 

[5 earlier articles] 
Re: Q: division vs multiplication davidm@Rational.COM (19950403) 
Re: Q: division vs multiplication brandis@inf.ethz.ch (19950404) 
Re: Q: division vs multiplication Terje.Mathisen@hda.hydro.com (19950406) 
Re: Q: division vs multiplication meissner@cygnus.com (Mike Meissner) (19950416) 
Re: Q: division vs multiplication martens@cis.ohiostate.edu (19950416) 
Re: Q: division vs multiplication jmccarty@spdmail.spd.dsccc.com (19950418) 
Re: Q: division vs multiplication leichter@zodiac.rutgers.edu (19950411) 
Re: Q: division vs multiplication kptben@aol.com (19950417) 
Re: Q: division vs multiplication pcg@aber.ac.uk (19950417) 
Re: Q: division vs multiplication gsc@magna.com.au (19950418) 
Re: Q: division vs multiplication jbuck@Synopsys.COM (19950428) 
Re: Q: division vs multiplication davidm@flora.Rational.com (19950428) 
Re: Q: division vs multiplication Roger@natron.demon.co.uk (Roger Barnett) (19950428) 
[1 later articles] 
Newsgroups:  comp.compilers 
From:  leichter@zodiac.rutgers.edu 
Keywords:  arithmetic 
Organization:  Rutgers University Department of Computer Science 
References:  9504003 9504059 
Date:  Tue, 11 Apr 1995 16:52:28 GMT 
davidm@Rational.COM (David Moore) writes:
 [ASIDE]. A year or two ago, there were a couple of papers that
 appeared together somewhere, and which I know I have, but which I
 have been unable to locate. One was entitled something like "How to
 write floating point numbers" and the other "How to read floating
 point numbers".

 Can anyone give me the reference?
A bit more than a year or two ago  time does fly, doesn't it? See:
Proceedings of the SIGPLAN'90 Conference on Programming Language Design and
Implementation; appeared as SIGPLAN Notices 25(6). The paper on printing
floatingpoint numbers is by Steele and White; the one on reading them by
Clinger. The actual titles are: "How to [readprint] floatingpoint numbers
accurately".
I've always wondered whether the authors planned to submit their papers
together, or whether they coincidentally submitted to the same conference
and adjusted the titles later.
A related algorithm, due I believe to Lyle Ramshaw: Given a floatingpoint
specification equivalent to C's %f with no explicit precision, how much digits
of precision should you actually use? Ramshaw's answer is the only thought
through one that I've seen: "%f" prints the decimal representation R of a
floating point value F with two properties: If R is read back by the corres
ponding input conversion on a machine with the same internal representation,
it gets back exactly F; and no decimal representation R' that also, when read,
produces F, is shorter than R. I'm sure the algorithm  which is efficient
enough to be used in the standard Modula3 libraries  has been published, but
I don't have a reference. (A reference almost certainly appears in the
Modula3 library documentation.)
 Jerry

Return to the
comp.compilers page.
Search the
comp.compilers archives again.