2014-06-26 13:17 GMT+02:00 Yuriy Tymchuk <yuriy.tymchuk@me.com>:
I think that while working with float arithmetic it will be nice to provide something like #~= instead of #closeTo:, and also add #~<, #~<= and so on. But maybe Iâm wrong :)
Uko
Yes, it's wrong because there is no such thing as a universal tolerance. We need at least two parameters.
On 26 Jun 2014, at 13:07, Christophe Demarey <Christophe.Demarey@inria.fr> wrote:
Le 26 juin 2014 à 12:46, Nicolas Cellier a écrit :
$! is not classified as a binary character right now, is it? Of course it's possible but I'd just say why this particular symbol? It does not look mathematical either. Or do you love C/C++ so much?
Of course not. It is just because it is something widely used but other names could do the stuff like <> or /=
Note that ~= has same meaning in Matlab.
There are other possibilities like /= or <> (Fortran 95, ADA), even the long form =/=.
But frankly, I don't buy this kind of change, for the sake of looking like some {} language, making such a deep change is overkill.
Indeed it is a deep change but maybe at least we could have an alias and use it for all new code. At a time, it will be less painful to do this change. As expressed in another thread, Smalltalk code convention for method naming strategy is *intention revealing*. That said, I think ~= is not intention revealing and it will be good to fix or propose a good alternative.