On 5 Mai, 23:15, rjf <[email protected]> wrote:
> I disagree with much of the above sentiments.

Not that precise... (or is the following meant to be a complete
enumeration?)

> If you are using a number which might be complex and your intention is
> to drop the imaginary part if it is very small, then you can do so, by
> taking its real part.
>
> The producer of the complex number with zero imaginary part could have
> dropped that imaginary part if it intended for it to be equal to a
> real.
>
> If the imaginary part is the integer 0, then there is probably no
> confusion, and
> the imaginary part might just as well be dropped.
>
> I think that associating the number of bits of precision with the
> number of decimal digits typed in, is a mistake, whether it is
> 1.000000000 or 0.000000000.

Yes. As I mentioned earlier in another thread, that's the reason for
having hexadecimal float literals in e.g. C99.
If the number of *decimal* digits matters, one should also do base-10
arithmetic (or keep track of it in computations).

On the other hand, one could simply *define* decimal numbers with n
digits to the right of the point to be converted to binary numbers
with m bits precision/mantissa. There are of course corner cases when
allowing arbitrary bases.

> There is a substantial difference between a real number and a complex
> number
> whose imaginary part is very small.

How small? This obviously depends on the application, but one should
never *implicitly* discard the imaginary part.

-Leif

-- 
To post to this group, send an email to [email protected]
To unsubscribe from this group, send an email to 
[email protected]
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Reply via email to