Builtin Float Epsilon? (was: Re: Does python suck or I am just stupid? )

[Irmen de Jong]

Dan Bishop wrote:
> Irmen de Jong <irmen at -NOSPAM-REMOVETHIS-xs4all.nl> wrote in message news:<3e58092a$0$49113$e4fe514c at news.xs4all.nl>...
> 
> 
>>PS. is there a language (apart from the already mentioned ABC) that 
>>doesn't use the machine floating point datatype for floats?
> 
> 
> Early versions of Turbo Pascal didn't.  But in those days, a floating
> point coprocessor wasn't standard equipment on PC's.

I wasn't talking about strictly hardware FP support, although perhaps that
wasn't clear from my message.

Usually when a platform didn't (doesn't) have hardware FP support,
some sort of FP emulation library was present, to do the FP math in
software emulation. Such a library suffers from exactly the same
problem as hardware FP because the floating point numbers are
usually represented the same in memory.

So that doesn't count.

I think that what I meant was: are there languages that don't use
"IEEE (?) floating point" numbers at all, but use some other representation
form that doesn't suffer from the problems we're talking about?


Irmen de Jong