Not sure I'd counterpose "array-based versus OOP" as it's quite possible to represent and manipulate a multidimensional array *as an object* (like a list, but deeper).
That being said, true enough that J presents a strongly functional right to left pipeline, very different from Python. More like REBOL perhaps? (Jason Cunliffe has been our resident REBOL booster). So here's where I insert my "tour the languages" approach: rather than focus on NumPy, just focus on J for awhile. Learn what that's like. Maybe your Python'll be better as a result. Instead of charming our snake into doing stupid pet trix, use it to show off the power of OO, and *other languages* to show of the power of other-than-OO. Touring may be fairly painless, some high rez screencasts and such. No one says you have to become a master. My CS0 at Princeton featured APL, FORTRAN, PL/1, SNOBOL and some simulated Assembler as I recall, among other languages. Why not some xBase while we're at it, with embedded SQL? http://worldgame.blogspot.com/2006/08/toontown-revisited.html http://worldgame.blogspot.com/2006/07/taking-refuge-in-j.html Kirby _______________________________________________ Edu-sig mailing list [email protected] http://mail.python.org/mailman/listinfo/edu-sig
