Robert Klemme
9/15/2007 3:46:00 PM
On 15.09.2007 17:39, hadley wickham wrote:
> On 9/15/07, Austin Ziegler <halostatue@gmail.com> wrote:
>> On 9/15/07, kendear <summercoolness@gmail.com> wrote:
>>> given the ease of using Ruby, I wonder whether we will have rand(m, n) in
>>> the future...
>>>
>>> right now, we need to use rand(range) + start
>>>
>>> and it can cause a bug that can otherwise be avoided if we have rand(m, n)
>>> for the clarity.
>> Why not instead suggest rand(m..n)? That would, I think, be more
>> readable than rand(m, n).
>
> But would that draw discrete or continuous random numbers?
That question actually applies to both variants. :-)
The rule could probably be, if at least one of the two numbers is non
int (i.e. float) then a "continuous" range is used. Alternatively we
can define another function that does this ("frand" or "randf").
Btw, currently if you use a float as single parameter you still get
integers back as far as I can see.
Kind regards
robert