Python 3 Is Not Turing Complete
ahaha what the fuck, he made some sense before (ASCII > Unicode, obviously) the logic behind this point is completely nonsensical.
in general, his rant boils down to two problems: inconsistencies with string handling and incompatibility with python2. the first one I can get behind, especially given how py2 did a fairly good job with strings (introducing a unicode string type with a support of unicode literal would be a superior solution to what they did) - but I don't think strings are
that important, this isn't perl after all. the second one is missing the point - py3 is backwards incompatible because it doesn't want to be limited by the design choices of py2, and py2 is still maintained so that legacy code still works.
this of course doesn't mean that py3 doesn't have questionable design choices - it's just that they're different from what he complains about. my pet peeve is lazy evaluation of
map() and
filter() - this makes sense with immutable haskall data structures but becomes unpredictable with FIOC lists that keep getting
append()ed and
extend()ed and having elements
remove()d from them. not to mention that
map() with side effects won't have those side effects applied by default. because of mutability of everything, those things should work the way they work in lithp, not the way the work in haskall - and that's how it used to be in py2.