Re: Strange result with exponents...

[Douglas Handy <dhandy1@xxxxxxxxxxxxx>]

Bob,

>This is one of the reasons why you don't presume that EVAL will yield
>"correct answers". IBM is using floating point on EVAL, that is a well known
>fact. Why did they decide to do that?  

Whoa.  I don't think that is necessarily true.  It used float in this instance
because of the use of ** for expotentiation.  If you are doing * or / operations
and the two operands do not include a float field, I think you have the exact
same operation occuring as with MULT/DIV *except* that it will not suppress the
overflow error.

With more complex expressions, the rules for what is used for intermediate
results are important to understand.  But simply changing MULT or DIV to an eval
of two operands will NOT cause float to be used.  In fact, I'd argue that it is
EVAL which has the "correct" answer and that:

     C           MyMDY     MULT 10000.01  MyYMD

is the *incorrect* answer.

Doug