On Sunday, May 31, 2015 at 10:28:39 AM UTC+5:30, Chris Angelico wrote: > On Sun, May 31, 2015 at 2:46 PM, Rustom Mody wrote: > > More to the point no language matches perfectly¹ everything that a learner > > needs to learn. > > Can you write a kernel module in python? (Or Haskell?) > > Can you see details of machine state and transitions in python? > > Can you client-script a browser in python? > > C is the best fit for the first > > Assembly for the second > > Javascript for the third. > > And if I want to build a web site, I won't write assembly code using > Python idioms. I will use Python. Why fight against a language?
I thought I answered. I try again... Because its the law. Why does iron rust? Why do cars/rockets not go faster than light? Why do we age/shrivel/wither/die? Because thats the law. The law here is Gödel's second theorem -- no single formalism can encompass all formalisms. Any real solution to a significant problem will use some (set of) languages which will fit good here and not so good there. Square pegs in round holes are not ideal Yet every so often there's no choice but to use a wrench to hammer a screw through a brick. The C/assembly/javascript were extreme black-n-white examples. Most real cases are more fuzzy. Linus Torvalds thinks C++ (both the language and programmers) consititute an asinine choice for writing kernel-code. I am ready to bet that a democratic election on the subject will leave Torvalds trailing. ie the world's C++ programmers likely disagree with him. Who is right? Ive no idea. Do you? The main point lost in this discussion is "What is the point?" If you ask a beginner to write a function to give fibonacci numbers, are fibonacci numbers the point? I dont believe that python is 'the point' much more than fibonacci numbers. -- https://mail.python.org/mailman/listinfo/python-list
