Hi All,

I have a question about floating point exponential representations
using 'e' vs. 'E'.  I'm in the middle of changing a lot of Blue code
and noticed that using doubles, simple printing of the double results
in big E usage, i.e. .0001  would be 1E-4 and not 1e-4. It looks like
for Java it uses big E for both floats and doubles.

Csound's lexer currently supports small 'e', but not big 'E'. It's a
simple change to use [eE] in the lexer, but one thing I'm not sure is
if there is a difference between usage of 'e' and 'E'.  My initial
look at Java says no, that e and E are somewhat interchangeable, but I
am unsure whether that is universally true. (i.e., is one used for
32-bit and another for 64-bit floating point?)

Any thoughts on this would be appreciated!

In C89 standard e and E are identical for constats (section 3.1.3.1) and 
in fprintf e or E are contrlled by e or E format.

So in my world they are te same.  I can add E as acceptable -- little more 
work than e -> [eE] thoug.  Perhaps I will do it anyway

On Tue, 27 Sep 2016, Steven Yi wrote:

> Hi All,
>
> I have a question about floating point exponential representations
> using 'e' vs. 'E'.  I'm in the middle of changing a lot of Blue code
> and noticed that using doubles, simple printing of the double results
> in big E usage, i.e. .0001  would be 1E-4 and not 1e-4. It looks like
> for Java it uses big E for both floats and doubles.
>
> Csound's lexer currently supports small 'e', but not big 'E'. It's a
> simple change to use [eE] in the lexer, but one thing I'm not sure is
> if there is a difference between usage of 'e' and 'E'.  My initial
> look at Java says no, that e and E are somewhat interchangeable, but I
> am unsure whether that is universally true. (i.e., is one used for
> 32-bit and another for 64-bit floating point?)
>
> Any thoughts on this would be appreciated!
> steven