I like this mini-project-approach: has something from those old
Meccano-manuals: look at the pictures and (re)build some examples with a
step-by-step-help, then start changing the end-result a little bit, and then
fantasy from scratch... 

I could say now, that my 14-year old daughter wants to start with Smalltalk,
but to be honoust I want to learn it at last myself also (again): I want to
learn it for adding educational functionalties to these also wonderful swiki
things and as a powertool for prototyping fantasies (Look at your users,
what works, what fials, what inspires them to new wishes...)

In style with these manuals it would help to show at the beginning of every
mini-project a picture/ working model of 'The' endproduct.

Maybe is it fun to create a collection of mini-projects: for example I
remember that in the first release of Hypercard there was a small and
elegant demo of a histogram-tool. You had to fill in a row of numbers, then
a push on a button and there it was, a nice simple Graph or Pie ... 
(Dan Shafer.. Dan Shafer... know this name, so I walk to my bookshelfs,
There it is, a book I bought ten years ago, introduction in Digitalk/Windows
Smaaltalk... (shit) this graph-demo was already in that book, I forgot so
much.. terrifying.)

Maybe two examples for intermediate level, focusing on  automatic drawing on
the canvas: 
- 1. From the days of AI there was this game animal guessing, often used in
the LOGO-area. I wrote in these days a small minimal-esperanto-basic example
that also draw the resulting knowledge-tree on the screen. 
- 2. Another example I still like is a MariMaxCorrelation example from Byte
1982? (I still have the basic listing: In this demo you had to collect data
of different aspects of different computer-brands like TRS-80, Apple-II,
KayPro CP/M, and then the program draws a relationships-tree.

I like your bye-the-way-style, for example the part where you explain that
it could be handy for the eventually user to draw a box around things that
belong together... You could here also give a complete lecture in
userinterface design, but that wouldn't that be boring(?) and distracting
from the hands-on exercise. I prefer these small thoughts in the sideline.

What makes these mini-projects most valuable, are the remarks that try to
explain the way of thinking as a Smalltalk designer when you try to
dissamble the problem and try to reassemble the solution:
For example, You give a golden tip for a beginner... but that one is at the
sametime difficult for a starter: "use simple functions instead of the more
powerfull ones... =KISS?".
But how do I now what the simple functons are? 

(Study and compare.. there are so many.. where to start.. Less code is
simpler? ... or...
Now I remember my struggle from ten years ago..) 

A little list with start-objects to begin with would help here? (Again: like
the meccano-box has a list of parts you need for assembling that
mini-project, explaining in the sideline little-by-little how the incredible
rich toolbox is organized.. (Reading the emails of experienced programmers I
understand that this toolbox-endeavor never ends..? 

(Or am I as a beginner on the wrong mental-track?)









-----Original Message-----
From: Dan Shafer [mailto:dshafer at yahoo.com]
Sent: Monday, April 09, 2001 4:47 AM
To: Squeak List
Subject: My First Doc Contribution


I have just uploaded to the Swiki my first contribution to Squeak docs, a
probably over-long tutorial on how to build a simple project entirely using
direct-manipulation Morphic techniques.

Please check it out and make appropriate corrections, comments, and
feedback.
Now that I've made this first conceptual break-through, I plan to do several
more such things of increasing complexity.

Thoughts and criticisms and suggestions welcome!

It's at http://minnow.cc.gatech.edu/squeak/1827


__________________________________________________
Do You Yahoo!?
Get email at your own domain with Yahoo! Mail. 
http://personal.mail.yahoo.com/