2014-07-29 13:10 GMT-03:00 Stefan Karpinski <[email protected]>:

> Integer size is system-dependent while floating-point size is not. The
> 8087 <https://en.wikipedia.org/wiki/Intel_8087> already had 64-bit floats
> and every x86 since the i486 <https://en.wikipedia.org/wiki/Intel_80486>
> has too, so there's no reason to change the size of floating-point numbers
> on different systems. When you write 1.0 in Julia, it is always a Float64.
>

Nice.


> Real is an abstract type that has many non-floating-point subtypes,
> including all integer types and rationals. Being a subtype of Real means,
> as the name suggests, that a type represents values on the real number
> line. There is a FloatingPoint abstract type, which is closer to what
> you're thinking of, but it also includes BigFloat.
>

That is what I thought. But since Float64 is ubiquitous, why not the
cleaner alias Float instead? It would improve readability by a little yet
considerable amount, IMHO. The "64" in the symbol is boiler plate for most
use cases.

Júlio.

P.S.: Why Julia users do not hang on IRC? Such trivial questions would be
addressed instantly. I guess this mailing list is being misused as an IRC
channel. (~54 emails per day).

Reply via email to