On Feb 2, 2005, at 10:31 PM, [EMAIL PROTECTED] wrote:
After I got all convinced that Python was not the final word in rapid prototyping languages there you go and drop a quote in my lap that has Norvig and others trying to convince me to stay in Python and dump LISP!?
No, I have always said to use the language that *fits*.
If I'm heavy doing numeric work, I pull out my Fortran manuals and blow the dust off. The lack of aliasing is a big win over *every other computer language*.
Sysadmin one-liners get treated to sh or Perl.
If I were writing software which communicated to enterprise server machines, I'd probably use Java.
I use Python for day-to-day programming activities still. Being able to just grab a library that "just works" is damn useful. Having hash tables with very flexible keys is *extremely* useful. I *destroyed* a software interview question by writing it in Python due to that particular feature.
I have been using Lisp for more tasks with my VLSI stuff. That stuff seems to require higher algorithmic abstraction. I am hoping that Lisp will improve my productivity a little (I would be happy with even 10%). OpenMCL also interacts with OS threads better than Python and compiles to faster code in most cases. I'm doing operations on a million or a billion objects. I need speed in places. Lisp seems to be getting me that without going back to C++.
I will go use C++ in those places where I *must* *have* *more* *speed*. In fact, one person has an experience that echoes mine with C++. Correctly used C++/STL is a fine programming language. See:
http://userpages.umbc.edu/~bcorfm1/C++-vs-Lisp.html
I am an engineer; I use what works.
-a
--
KPLUG-List mailing list [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
