[Edu-sig] My new CP4E Mega Theory

[Arthur Siegel]

Group -

Devoting time to some redesign of PyGeo - trying to emphasis
the concept of the sourcecode as an integral part of the  application -
using the app and attempting to gain an understanding how it is doing what
it is doing are intended to be two sides of one coin.

Where this is leading me:

1) Simplification - less attempts at bells and whistles. Killed much of the
GUI stuff, e.g

2) Source code annotation - the source is the text. Doing it is hard and
boring at the same time. I'll see how long my enthusiasm lasts and how far I
get.

But most interestingly (here's the Mega Theory part) -

3) Using OOP to layer complexity. The top-level classes accessed directly by
scripts are simplest, which inherit from classes which reveal more of the
basic analytics, which inherit from classes which wrap Numeric in friendlier
terms and handle more of the advanced analytics and matrix algebra, which
inherit from Numeric, which extends C, where I get lost. Younger or less
motivated observers will stop at top-level classes and hopefully learn a
good deal about programming and geometry. Other folks might trace the path
beyond where I choose to stop.

But the basic concept of using inheritance in this way has general
significance in a CP4E sense, I believe.

And in the CP4E sense, is a better alternative - I believe - to relegating
Python to the role of simple scripting of compiled code (e.g, C or C++),
which (even if opensource) is beyond the reach of most of us anxious to take
a peek behind the Wizard's curtain - the languages being different and the
shift in levels of complexity (not necessarily because of the languages
themselves, but more because of a shift in focus by the creator) too abrupt.

ART