🚀 go-pugleaf

Andrew Gierth wrote:

> The following message is a courtesy copy of an article
> that has been posted to comp.unix.programmer as well.
>
> [bogus groups comp.gcc.help,comp.math.num-analysis dropped]
>
> >>>>> "Fred" == Fred Ma <fma@doe.carleton.ca> writes:
>
>  Fred> Hello, I'm using gcc's erand48.  On solaris8, the man page
>  Fred> header says SunOS5.8, while on cygwin, it says NEWLIB.  Both
>  Fred> pages say that the range of random numbers is the closed range
>  Fred> [0.0,1.0].  However, "many people" have said on various web
>  Fred> pages that the values lie in the range [0.0,1.0) i.e. doesn't
>  Fred> include 1.0.
>
> The standards say [0.0,1.0) is the correct range.

Hi, Andrew,

Just wondering if you could clarify the standard you're
referring to.  Reason I ask is because I'm told that
erand48 is not part of the C standard.  That's why I'm not
posting to the C newsgroup.

>
>
>  Fred> So I wonder if anyone knows for sure whether the random number
>  Fred> spans the closed range, or semiopen (latter case).  I mean know
>  Fred> for sure, because suspicions abound (I have my own suspicions
>  Fred> that the semiopen range is correct because the integer
>  Fred> counterparts in this family of random generators always goes up
>  Fred> to +/-( 2**31 - 1 ), whereas I suspect the floating point
>  Fred> versions just normalize the integer version by 2**31.
>
> not quite. The '48' in the name of the function may give you a clue
> here; all 48 bits of the generated value are used as the mantissa of
> the floating-point result (after normalization). (This can't round up
> to 1.0 with IEEE format, which has more than 48 bits of precision; the
> lower order bits are left as 0.)
>
> --
> Andrew.
>
> comp.unix.programmer FAQ: see <URL: http://www.erlenstar.demon.co.uk/unix/>

--
Fred Ma, fma@doe.carleton.ca
Carleton University, Dept. of Electronics
1125 Colonel By Drive, Ottawa, Ontario
Canada, K1S 5B6