From: utzoo!decvax!cca!mike@Rand-Unix@sri-unix
Newsgroups: net.unix-wizards
Title: Re: Horrid float code in 4.1 C compiler?
Article-I.D.: sri-unix.3008
Posted: Mon Aug 30 21:08:27 1982
Received: Wed Sep  8 02:00:12 1982

Date: Wednesday, 25 Aug 1982 12:20-PDT

Yes!  It the code is horrible.  The C compiler does not do good things with
single precision numbers: it converts them to double, does the math,
and then converts back.

One solution is to just use double precision.

Another solution is to use single precision and run slower thann 
double.

The right solution is for someone to fix the C compiler to handle this
very useful primitive data type correctly.