Brian Ford wrote:
> On the contrary, I have strongly argued for Rubinius to be accessible
> and modifiable to the deepest levels since hearing Evan's comment
> about Rubinius kinda having evil built in. Of course it can be
> misused, just as I can load a segfaulting C extension into MRI.
...
> I understand how selector namespaces can be used. My concern is that
> they can be misused in a way that makes software harder to understand
> or that requires an even stronger antidote than monkey patching to
> work around someone's well-intentioned assumptions that may be invalid
> in a different context.

So on the one hand you support Rubinius being totally open, even if that 
openness can be misused, but on the other you don't want selector 
namespaces because they can be misused. Aren't these opposite positions?

And if Rubinius is as open as you want it to be, wouldn't it be simple 
for someone to implement selector namespaces in Rubinius? And would you 
support that or not?

So clarify for me: are you saying you support Rubinius's total openness 
and its vastly more likely potential for misuse, but you don't support 
namespaces and their relatively limited potential for misuse?

- Charlie