On 6/11/07, Ka-Ping Yee <[EMAIL PROTECTED]> wrote: > Because the existence of these library modules does not make it > impossible to reliably read source code. We're talking about > changing the definition of the language here, which is deeper > than adding or removing things in the library.
This has already been demonstrated to be false - you already cannot visually inspect a printed python program and know what it will do. There is the risk of visually aliased identifiers, but how is that qualitatively worse than the truly conflicting identifiers you can import with a *, or have inserted by modules mucking with __builtins__? > permit alternative character sets, as long as Python offers an > option to make that choice. We can continue to discuss the > details of how that choice is expressed, but this general idea > is a solution that would give us both what we want. I can't agree with this. The predictability of needing only to duplicate dependencies (version of python, modules) to ensure a program that ran over there will run over here (and vice versa) is too important to me. When end users see a NameError or SyntaxError when they try to run a python script, they will generally assume it is the script at fault, not their environment. Michael -- Michael Urman _______________________________________________ Python-3000 mailing list [email protected] http://mail.python.org/mailman/listinfo/python-3000 Unsubscribe: http://mail.python.org/mailman/options/python-3000/archive%40mail-archive.com
