RE: integer math

[Luiz Henrique de Figueiredo <lhf@...>]

>The solution to this problem is education.

Thanks for a nice summary.

>I think the essential 
>integer-backwards-compatible behavior of double floating point should be 
>emphasized more in the documentation, the website, the faq, the wiki, etc. 

That's a good suggestion.

>There are users of older or smaller or embedded CPUs who have no floating 
>point, or no 64-bit floating point, or CPU emulated floating point, or slow 
>hardware floating point. E.g. 80386, 68040, ARM, etc. I guess the config 
>feature mostly fills their needs.

Precisely. Plus using float instead of double to save memory. If you have very
large graphical databases, as we do in TeCGraf, then sometimes a float version
of Lua is needed. Of course, using float gives you *less* precision for
representing ints (23 bits instead of 32).
--lhf