A Sourceforge project for newbies
Ron Stephens
rdsteph at earthlink.net
Wed Jul 3 16:34:36 EDT 2002
Hmmm, announcing a new Sourceforge project called Decision Analysis at
http://sourceforge.net/projects/decision/
The code available for downloading is decide.py, a much updated
(improved?) version of my very simple (yet functional, as in working,
code) program which some folks on this list helped me with a year ago,
when I was a total newbie to Python.
The idea is this: newbies can use decide.py, and the Decision Analysis
Sourceforge project, to learn new programming skills, and practice the
Python techniques learned from books and tutorials. No one is too new,
or too clueless, to contribute.
The ideas behind decide.py are very simple, and the code is well
documented, if (certainly) not elegant. But there is much that can be
added, and I am betting that almost any newbie can find ideas in their
own head to add to decide.py, without hurting their nogin too much . ;-))
The general version of decide.py is a generalized decision maker. The
same code can be easily altered to create new modules, of a specific
nature, in any areas in which the individual coder is an "expert" Hey,
we're all expert in somethings! For instance, one of the areas in which
I claim vast expertise, and therefore I created a module for, is
deciding where to go and what to eat for lunch on any given day. Others
may be experts in nuclear physcis or world peace, and decide.py can just
as easily be extended to incorporate their expertise as well!
Beyond adding new modules of exxpertise, which probably represent the
easiest additions coders can make to decide.py, there are many other
areas for coding, a short list of which follows:
1. Decision Analysis needs a GUI, any GUI. This will be relatively
straight-forward, with one big caveat: one must be able to program GUI's
in *some* (read *any*) Python GUI toolkit. How about a PythonCard
version, a Tkinter version, a Java Swing vesion, a wxPython
version...let a thousand flowers bloom...
2. The simple underlyng logic behind decide.py needs to be extended. Any
one with a even a modicum of statistical knwoledge (as in more than me
;-))) should be able to come up with more interesting algorithms to use
to supplement (or replace!) decide.py's current logic.
3. PDA's. ...I think that decide.py's kind of fun, easy, and
lightweight programs are prefect for use on PDA's. I already have
decide.py working perfectly well on my Sharp Zaurus, which is cool. But
a version taylored just for a given PDA, with a lightweight GUI
taylor-made for the PDA, with sliders say for data entry, would be
really fun.
4. I have non-working code that uses simple semantic analysis of random
questions to identify type of question and keywords involved, , and then
uses the WWW and Google to do simple statistical anaysis to "answer"
random questions. Could be a part of a new decide.py logic module
someday, but the algorithm needs a lot of work. I will be uploading this
code to the Sourceforge site, as soon as I get it into reasonable
shape.(The existing code can be found on my web site, just look for
askMerlin.py)
Anyway, all are welcome. Hey, I won even turn away any non-newbies who
might show up! ;-)))
But for newbies, this is the one place on the net where you can code but
not be ashamed of your newbieness. Just have fun.
By the time I get back from my two week trip starting Monday July 8, I
hope to see some contributions and some questions. Address email
questions to rdsteph at earthlink.net or just post them on the Sourceforge
site.(That's if I survive *coach* round the world lfight with stops in
several countries ;-((((
Decision Analysis, think of it as a poor man's extensible, expert
system, semi-intelligent AI ssytem.
More information about the Python-list
mailing list