🚀 go-pugleaf

Roy Smith <roy@popmail.med.nyu.edu> wrote:
> So, what I'm looking for are some topics in computer graphics which
> will grab this student's interest, but which are free of higher math,
> and of a scope which could reasonably be expected to be handled in the
> first few weeks of an intro programming course by a student with no
> prior programming experience.

The LOGO language is designed for precisely this kind of thing.  LOGO is
an interpreted Lisp-like language with "turtle graphics" drawing
primitives along the lines of

   FORWARD n    move forward n units in the current direction
   LTURN a      turn left through angle a
   RTURN a      turn right through angle a
   PEN UP       stop drawing
   PEN DOWN     start drawing

These primitives, plus arithmetic and control structures for iteration
and recursion, are plenty to start drawing complicated figures.

-- 
Gareth Rees

roy@popmail.med.nyu.edu (Roy Smith):
< I'm really struggling to find programming assignments I can give
< which will be useful and interesting to him.

    I don't know how much you're planning on teaching, but imaging
operations on graphics files might be one answer.  On the programming side,
he will have to learn pointers (memory alloc & free), multiple indexing
(simple pointer and doubly-indexed arrays), and file I/O.  In addition, he'd
have to learn some file format and be able to read and write a file of that
format (I don't have a particular simple one in mind, though).

    The graphics side is very simple -- just about anything you do to an
image is magic when you're new to it.  For example, read in a simple image
and then write three files:  one red, one green and one blue.  Or write out
the four CMYK files.  Other possibilities include smoothing (just average
neighbors in some way) and sharpening (smoothing with different components
=^).

    Don't know how well this will mesh with his expertise, but I expect that
it might excite him enough to plow into it.

                                         Steve <hollasch@acm.org>

roy@popmail.med.nyu.edu (Roy Smith):
< I'm really struggling to find programming assignments I can give
< which will be useful and interesting to him.

You might try Barnsley's 'Chaos Game'.  It's generally the first thing
I do when starting with a new graphics system.

Here's the description:

1. Open a drawing area on the screen

2. Pick three widely seperated points on the screen (pixel coordinates,
most likely).  It's nice if these points form an equilateral triangle,
but not necessary.  Call these points the 'attractors'.

3. Pick a starting point in the drawing area.  Make this the 'current
point'.

4. Randomly pick one of the three attractors.  

5. Make the new current point the average of the coordinates of the
old current point and the attractor randomly chosen in step 4.

6. Plot the (new) current point to the drawing area.

7. Repeat steps 4-6 for a set number of times (100,000 may be a good
place to start - this number depends on the speed of your computer and
the resolution of the drawing area.


Magically, a figure will appear on the screen.  It is a fractal called the
Sierpenski gasket.  The three attractors form the corners of the base 
triangle.  This figure is the result of taking a triangle, connecting
the midpoints of each side, and removing the triangle in the middle.  
Recursively repeat until you get tired :).  The first stage looks like
this:

           /\
          /__\
         /\  /\
        /  \/  \
        --------

I hope you can get the idea.

A few notes on the procedure.  You will want to pretty much fill the
screen with the triangle, so choose the attractors along the top and
bottom edges.  If the starting point is chosen as one of the attractors,
you get instant convergence; otherwise, you will want to not plot the
first few (say 10) points.  This works well using all integer math in
pixel coordinates.  The randomly choosing 1 of 3 is probably the only
somewhat tricky point - but it will work quit well with just modulo'ing
the integer returned from standard random number routines to base 3.

One nice (minor) modification of this procedure is to change the 
color of the point plotted, using a different color depending on which
attractor was used to make the current point.  Nice bright red, green,
and blue work pretty well.

This basic algorithm is called an 'Iterated Function System' and leads
into Fractals and Fractal Image Compression.  With a teensy bit more
math, you can do other figures.  One very interesting one is a fern
leaf.

-- 
--> Mike Brindley
    brindley@ece.orst.edu

roy@popmail.med.nyu.edu (Roy Smith) wrote:

>I'm in a sticky situation and need some advice.  I'm teaching a class
>in introductory C programming to a mixed group of people, ranging from a
>faculty member here in the Microbiology department, to graduate and
>medical students, to one person (a young adult) who is a talented graphic
>artist but has a weak background in math and science, and no previous
>programming experience.  On the other hand, he's bright, and a wiz at such
>Macintosh packages as Photoshop, Swivel-3D, and the like, as well as such
>conventional media as pencil and paper rendering.  My artistic
>development, on the other hand, reached its zenith when I discovered in
>the 3rd grade that if you mix enough colors of paint, you get brown.

>     I'm really struggling to find programming assignments I can give
>which will be useful and interesting to him.  When I wrote the syllabus
>for the course, all my programming assignments were aimed at people who
>had strong math and science backgrounds (find prime numbers, solve the
>ideal gas law equation, stuff like that), but these are totally lost on
>him.  He is eager to learn, but is falling behind in the class because of
>this.   I've already had one other person of similar background drop out
>of the course because "I'm a visual person, not a programmer", and I'm
>determined not to see that happen with a 2nd student.


How about some programming assignments for basic biological principles?
Plotting population growth, or allele selection.  This may be interesting to the
biologists, and only need a high school level of education?


>     I don't have much experience in graphics programming, but all the
>examples I can think of in that field seem like they quickly end up
>involving math which is beyond this student's capabilities.  So, what I'm
>looking for are some topics in computer graphics which will grab this
>student's interest, but which are free of higher math, and of a scope
>which could reasonably be expected to be handled in the first few weeks of
>an intro programming course by a student with no prior programming
>experience.

For an actual graphics or imaging project, i'd try to frame it
around the concepts you are trying to teach.

For your file IO sections, a graphics file manipulation may be a
good project.  I'd use XV to display the images,and create/read a PPM
(or other simple format) images. The for and while loops lessons 
can modify the images, convert to gray scale or blur the images.

Or you could do a simple OpenGL program, the samples in the OpenGL
programming guides are really simple and appropriate for a first
C class.


Alex

In article <roy-2211951523470001@mchip8.med.nyu.edu>, roy@popmail.med.nyu.edu
 
says...
>
>     I'm in a sticky situation and need some advice.  I'm teaching a class
>in introductory C programming to a mixed group of people,
[snip]
>to one person (a young adult) who is a talented graphic
>artist but has a weak background in math and science, and no previous
>programming experience.

Fractals, recursive subdivision (turtle graphics!) or Mandelbrot/Julia sets 
(Yeah, the math is complex, but if you break it down to arithmetic it's 
trivial).

My first ray tracer was done with an average high school algebra education, and 
I'm no math wiz (But I'm persistent, which counts for something, so that may be 
a little more complex than you want to try).

Some image processing can be done brute force (Blurs, rescaling, have him play 
with "brush stroking" algorithms).

Some simple non-linear simulations (bouncing ball) done in 2d, perhaps even in 
3d if you've got a graphics package he can tie it into.

Dan