I think there is a serious misconception going on among some Python
users. There are some of us -- me included -- who looked very
seriously at Python, and found we couldn't use it for what we wanted
to do. In my case, the indentation mechanism was the most serious
(though not only) problem.
I'm not interested in making Python into C++. I have, and use, C++
for that. What I want out of a language similar to Python is a
language in which I can effectively write higher-order functions in my
programs. I want a language where it's easy to reify ideas I want to
try reifying. In short -- I want a language where it's easy to
manipulate programs in programs. The tools I like best for this so
far are languages in the lambda-calculus inspired family (various
lambda calculi, to Lisp, to ML and friends). In other words, I agree
that things shouldn't be fixed if they aren't broken. But Python is
broken, for my needs.
(In practice, these days, I generally use a hybrid Tcl/C++ system in
my work -- not because I think it's a perfect solution, but because it
got the job well enough at the right time. The use, by the way, is to
configure an object-oriented simulator system -- essentially we use
C++ interfaces for the object interface and Tcl for the metaobject
interfaces. To put bigger words to it than it deserves :-)
When one writes a system and makes it publicly available, it ceases to
be completely one's own system. Other people look at it and evaluate
it for their needs. Python is no exception to this, and there are
those of us whose needs are legitimately not being served by the
current system. We can choose not to use it, as I have done, or we
can try to suggest what it is about the system that could be changed
to make it suitable for our needs. (Personally, I value users who
come up with solutions much more than ones that only complain about my
systems.) I'm not completely keen on the current proposal, but at
least it's a concrete proposal that can be discussed. I hope people
will keep this in mind, so that _discussion_ rather than _flamage_ can
be the order of the day.
-richard (golding@hpl.hp.com, formerly @cs.vu.nl)