Luis M. Gonzalez wrote: > So, we can safely say that Boo is "almost" a static python > implementation. Wether you like or not, is another problem, but please, > do not insist with your reiterative anti-boo ranting.
I can't comment on Boo beyond the documentation's description of the language, but the question of what makes Python the language it is (or was) - what one would need to add or remove to stop Python being Python - remains very interesting. Boo appears to have some relatively simple type inference (compared to things like Shedskin) and has various static typing constraints presumably to fit the underlying virtual machine. Seeing through the syntax, do such things keep it sufficiently like Python? (Remember that Java looks like C++ but it has been suggested that it is closer to Smalltalk in character.) Looking not very far down the description of Boo's type inference [1] reveals some pretty fundamentally different semantics, although one could argue that they are "better" according to some criteria. I'd argue, considering other evidence, that Boo isn't enough like Python to be a kind of Python - not necessarily a criticism, though, but an observation. Meanwhile, the temptation to add similar type annotations to Python should be resisted, in my view, since the book certainly isn't closed on alternative strategies for compile-time checking and program optimisation. Paul [1] http://boo.codehaus.org/Type+Inference -- http://mail.python.org/mailman/listinfo/python-list
