On Tue, 29 Mar 2011 19:34:27 +0200, Wojciech Mu?a
<w...@p...null.onet.pl.invalid> wrote:

>A.L wrote:
>> jeden z najlepszych amerykanskich uniwersytetow, Carnegie-Mellon
>> oglosil ze pzrestaje uczyc programwoania obiektowego na wstepnych
>> kursach programwoania
>
>I to wg Ciebie zle, czy dobrze?
>
>w.

Dobrze

Yes. There was some discussion in CACM (Communications of the ACM -
journal) regarding teaching/not teaching OO. Majority of opinions
were that OO is JUST ONE programming paradigm, and for some
applications hardly usable (such as embedded systems, control,
avionics), and teaching OO as first paradigm is equally bad as
teaching BASIC with GOTO.

In principle, I believe that they are right, in the sense that
bringing too much OO too early can create more harm than good.
Imperative paradigm is still doing well and students must know how to
use it.

Moreover, in the link above mentioned was the issue of modularization
and teaching students how to modularize programs. Actually,
modularization is totally forgotten in OO community, despite that it
was pointed out many times (Parnas, Wirth, Szyperski) that
modularization and OO are orthogonal concepts. I agree that
modularization should be presented as "stand alone", not within OO
context

This doesn't change my (negative) opinion about MIT's decision to
replace Scheme by Python. This was obviously marketing/business
decision. As they openly said, world today is not the same as 20 years
ago, and now students (even at MIT) have such bad preparation that
book SICP + Scheme is too hard for majority of them.

A.L.