From ...
From: Erik Naggum <erik@naggum.no>
Subject: Re: best CL type for fixed decimal values
Date: 2000/03/03
Message-ID: <3161081269457609@naggum.no>#1/1
X-Deja-AN: 592685699
References: <841E799D01879E00.B3B5EAAA6E7974E3.74A6C500761082CD@lp.airnews.net> <m2aekgfb3w.fsf@sophia.axel.nom>
mail-copies-to: never
Content-Type: text/plain; charset=us-ascii
X-Complaints-To: newsmaster@eunet.no
X-Trace: oslo-nntp.eunet.no 952094477 23977 195.0.192.66 (3 Mar 2000 14:41:17 GMT)
Organization: Naggum Software; vox: +47 8800 8879; fax: +47 8800 8601; http://www.naggum.no
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.5
Mime-Version: 1.0
NNTP-Posting-Date: 3 Mar 2000 14:41:17 GMT
Newsgroups: comp.lang.lisp

* Marc Battyani
| What CL types are best suited for computing with fixed decimal values with
| exactly n places after the decimal point?
| 
| I can use
|     floats with a special rounding after each operation

* Johan Kullstam <kullstam@ne.mediaone.net>
| floats are out since finite length decimal point numbers often have
| non-finite binary expansions.  0.1 for example cannot be represented
| exactly by floating point.

  but if he knows he has N places after the decimal point, that translates
  directly to a factor (expt 10 N), which means he could do well with the
  extra precision and speed of a double-float value relative to bignums.

#:Erik