Back in the days when it was easy to write interactive applications, I 
made a demonstration (as a Hypercard stack, and in Macintosh Pascal for 
the 512K Mac) that allowed the user to click on points to make a source 
triangle, and any number of destination triangles, to create an Iterated 
Function System.  It also output the coefficients for the affine 
transformations.  This was a great way to get to understand the "collage 
theorem" (see Barnsley's "Fractals Everywhere" book, among many others.) 
  Now I suppose that if I want to do the same thing again, I'll have to 
learn JAVA.

In the meantime, using POVRay, one can manually enter the values for the 
vertices of the triangles.  If anyone wants the code, I'll post it.

Dave Matthews

(With some work, perhaps I can use the IFS below for my "sig")

Post a reply to this message

Attachments:
Download 'collage01.jpg' (33 KB) Download 'collage02.jpg' (50 KB) Download 'collage03.jpg' (35 KB)

Preview of image 'collage01.jpg'

Preview of image 'collage02.jpg'

Preview of image 'collage03.jpg'

>  Now I suppose that if I want to do the same thing again, I'll have to 
> learn JAVA.

No need - I've got the code already.

Erm - somewhere...

*searches around*

Oh well! I _did_ have it... I guess you could always do it with 
PostScript instead! :-D

> In the meantime, using POVRay, one can manually enter the values for the 
> vertices of the triangles.  If anyone wants the code, I'll post it.

Neat.

> (With some work, perhaps I can use the IFS below for my "sig")

Yes, that would seem an optimal way to do it... ;-)

But hey, see if you can make the shape a lil more interesting first tho 
- maybe add a bit of twist or something... I don't know, long time since 
I did this stuff. ;-)

Andrew @ home.

PS. Nasty checker texture :-S

Post a reply to this message

Andrew C on Mozilla wrote:


> No need - I've got the code already.
> 
> Erm - somewhere...
> 
> *searches around*
> 
> Oh well! I _did_ have it... I guess you could always do it with 
> PostScript instead! :-D
> 

Seriously, if you do happen to find the JAVA applet, could you post a 
link in off-topic, or e-mail it to me?  I'd really appreciate it. 
Although I'm probably going to have to break down and learn a usable 
programming language again.  (PostScript?)

> But hey, see if you can make the shape a lil more interesting first tho 
> - maybe add a bit of twist or something... I don't know, long time since 
> I did this stuff. ;-)
> 

I could just flip some of the transforms around, or angle them so that 
they rotate inward, or being more radical and following the lead of this 
very interesting application,

http://flam3.com/

I could compose it with a non-linear transformation (see his 
well-written paper under "math" at the above link).

Mostly, though, I made this POVRay version so that I could make 3D IFS. 
  I can arrange ball-and-stick tetrahedra or parallelepipeds and this 
should be easier than trying to puzzle out the transformations directly.

> PS. Nasty checker texture :-S

Man, no-one likes my baby blue and pink checkers.  Here I am, just 
trying to be gender-inclusive.  Oh, well. ;-)

Thanks for the comments.

Dave Matthews

Post a reply to this message

>> No need - I've got the code already.
>>
>> Erm - somewhere...
>>
>> *searches around*
>>
>> Oh well! I _did_ have it... I guess you could always do it with 
>> PostScript instead! :-D
> 
> Seriously, if you do happen to find the JAVA applet, could you post a 
> link in off-topic, or e-mail it to me?  I'd really appreciate it. 
> Although I'm probably going to have to break down and learn a usable 
> programming language again.  (PostScript?)

I wrote a Java applet to do (hard-coded) 2D linear IFS as part of my 
dynamic HTML assignment. As with everything in Java, doing the program 
is easy, making the output *visible* is almost impossible! Displaying 
text, drawing dots... Java makes these things ridiculusly difficult.

>> But hey, see if you can make the shape a lil more interesting first 
>> tho - maybe add a bit of twist or something... I don't know, long time 
>> since I did this stuff. ;-)
> 
> I could just flip some of the transforms around, or angle them so that 
> they rotate inward, or being more radical and following the lead of this 
> very interesting application,
> 
> http://flam3.com/
> 
> I could compose it with a non-linear transformation (see his 
> well-written paper under "math" at the above link).
> 
> Mostly, though, I made this POVRay version so that I could make 3D IFS. 
>  I can arrange ball-and-stick tetrahedra or parallelepipeds and this 
> should be easier than trying to puzzle out the transformations directly.

Personally, when I do 3D IFS it tends to be 2D ones, tweaked slightly to 
use the extra dimension. Otherwise you seem to just get clouds of stuff 
which isn't all that interesting. (Having said that, it's not something 
I've had the pleasure of playing with all that much as yet.)

>> PS. Nasty checker texture :-S
> 
> 
> Man, no-one likes my baby blue and pink checkers.  Here I am, just 
> trying to be gender-inclusive.  Oh, well. ;-)

*grins*

Andrew @ home.

Post a reply to this message

Andrew C on Mozilla wrote:

> As with everything in Java, doing the program 
> is easy, making the output *visible* is almost impossible! Displaying 
> text, drawing dots... Java makes these things ridiculusly difficult.
> 

That's a shame.  Do you know of any framework (platform specific is 
fine) that makes it easy to write clickable graphic applications?  I 
used to love the old Macintosh Pascal, because you basically said 
something like "open window" "get mouse location" "draw line" -- things 
I can understand.


> Personally, when I do 3D IFS it tends to be 2D ones, tweaked slightly to 
> use the extra dimension. Otherwise you seem to just get clouds of stuff 
> which isn't all that interesting. (Having said that, it's not something 
> I've had the pleasure of playing with all that much as yet.)
> 

That's been my experience, also.  I'm attaching an IFS that I didn't use 
in Paul Bourke's Fractal contest, since it was just a spiral IFS "pushed 
up" a bit.  I'm hoping that if I can visualize them better, I'll be able 
to make them better.  We'll see.

Dave Matthews

Post a reply to this message

Attachments:
Download 'tdifs005.jpg' (38 KB)

Preview of image 'tdifs005.jpg'

"Dave Matthews" <dma### [at] nospamnet> wrote in message
news:409e7878@news.povray.org...
>(as a Hypercard stack, and in Macintosh Pascal for

>   Now I suppose that if I want to do the same thing again, I'll have to
> learn JAVA.
>

No reason to resort to Java.

Google for pythoncard,  its a mix of ideas from hyperstack and VB,
implemented in python.

Very easy to get your head around. Between this, wxWindows, tkinter, etc.
you can build
stuff like you describe very easily.

There's also pyGame which may be even easier for what your thinking of.

Post a reply to this message

new.povray.org wrote:

> Google for pythoncard,  its a mix of ideas from hyperstack and VB,
> implemented in python.
> 
> Very easy to get your head around. Between this, wxWindows, tkinter, etc.
> you can build
> stuff like you describe very easily.
> 
> There's also pyGame which may be even easier for what your thinking of.
> 

Thanks, I'll check into them.  Sounds like what I need.

Dave Matthews

Post a reply to this message