Re: 8.5 beta and decimal precision handling

spam@xxxxxxxxxxxx wrote:
On Fri, 28 Sep 2007, keithv wrote:

There are very good reasons for the change (shimmering
can no longer influence the value), but as Michael
suggests, to mitigate the above effect, you should
use [format] (and perhaps traces).

Yuck, that seems like a step backwards.

I learned from my FORTRAN days programming large statistical software packages on mainframes and minis to have great and abiding respect for floating point arithmetic, and moreover floating point comparison, and never to take things for granted. After all, as Leopold Kronecker succinctly stated "God created the integer, all else is the work of man".

Cheers,
Rob Sciuk

Please repost the question next week when the appropriate people are reading c.l.t. -- they have been at the Tcl/Tk Conference and I do not think reading c.l.t..

Also note, in 8.5 you have bignum as a built in.

--
+--------------------------------+---------------------------------------+
| Gerald W. Lester |
|"The man who fights for his ideals is the man who is alive." - Cervantes|
+------------------------------------------------------------------------+
.

Relevant Pages

  • Re: Rounding error
    ... that you are reading...you may not be reading correctly...if this is ... The format does *NOT* affect the accuracy of input. ... The OP's problem is undoubtedly the finite precision of floating point ...
    (comp.lang.fortran)
  • Re: frustrated by fscanf and sscanf
    ... there is no threads talking about reading a series of numbers. ... <snip description of input file structure> ... Your description of the file format makes no sense at all. ... along with a sample of your input file or a clearer ...
    (comp.lang.c)
  • Re: converting float to ascii w/o printf
    ... Currently I am using sprintfto format this output into ... the rule here is never use floating point at all if it can be avoided. ... If you know what the range of values and required accuracy is, scaled integer arithmetic will be faster and more efficient, especially if you can arrange the data to allow shifts, rather than multiply and divide. ... Floating point on small machines is slow, inefficient, adds complexity and rounding errors everywhere it's used and you can usually get better control and predictable accuracy using other methods. ...
    (comp.arch.embedded)
  • Re: converting float to ascii w/o printf
    ... Currently I am using sprintfto format this output into ... this is using simple multiplies and ... check if you have the ftoa function available. ... Do you really need to store the information in floating point format? ...
    (comp.arch.embedded)
  • Re: [Q] Text vs Binary Files
    ... floating point and other more exotic formats). ... format in which a binary file is stored. ... have no text editor but do have a hex editor, ... you have a copy of the relevant file format standard, ...
    (comp.programming)