bearophile wrote:
Andrei Alexandrescu:
Maybe community should decide in poll who's arguments are better?
That would be helpful.
Designing a language with polls is an obvious idea, but I have not suggested it in the
past because it has some drawbacks too (there's a risk of losing the coherence of the
language and to use "political" solutions that are mostly good to close the
muzzle of the majority of people).
On the other hand it has some advantages too (example: despite you are
intelligent, some names and function/templates signatures you have added to
Phobos may be not good because you aren't equal to the average person, everyone
has some personal quirks. So designing by poll helps remove such quirks).
So this time, as an experiment, we may try to see what happens using a poll :-)
Language design polls must be not anonymous, each vote must have a name beside
it.
I think that a poll done in the past may have lead to different outcomes (for example in
two recent changes to the "switch", one by you and one by Walter).
Great thoughts. One thing that maybe is not clear is that I afford to
argue in these debates exactly because I don't get to decide. Walter
never implements something that is not convinced of. It happens often
that I suggest an improvement that he doesn't approve for one reason or
another. Last time I suggested that switch should work with all types
that support comparison (it's just weird that it only works with ints
and strings), but he refused to consider it at least for the time being.
So, polls would be very useful to gauge what people want. That's very
solid information.
Andrei
