On Thursday 08 June 2006 01:30 pm, Andrew Lentvorski wrote: > > Unfortunately, the same experience doesn't exist for decimal floating > point. However, I can tell you that I have yet to meet even a good > programmer who can write binary floating point code that doesn't have > all manner of horrible problems. > > -a
I will go along with that. Back in the early 80's I was developing an application for the Apple II. We had data that had a wide dynamic range and low accuracy. We had lots of it and so I thought well just make up a special floatingpoint for this data and cram it all into 8 bits IIRR. It worked. But it was not nearly as easy as I thought it would be. We did it with binary floating point. I suspect it would have made more sense to invent some kind of decimal floatingpoint, but we were ignorant of all of the issues involved when we started ... as is often the case. Fun project though. BobLQ -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg
