[ANN]: Gradient Kaleidoscope Galleries (and Rev 3.5 still crashing)

[Wilhelm Sanke]

As the "Gradient Kaleidoscope Tool" stack is nearing completion, but 
still crashes Rev 3.5 more or less frequently, I have uploaded a number 
of examples produced with the Kaleidoscope tool to demonstrate what can 
be achieved with it.

Either go to <http://www.sanke.org/MetaMedia> and scroll down on the 
left side until you see "Kaleidoscope Galleries",
or go directly to

<http://www.sanke.org/MetaMedia/GradientKaleidoscopeGallery1.htm>  or
<http://www.sanke.org/MetaMedia/GradientKaleidoscopeGallery2.htm>  and
<http://www.sanke.org/MetaMedia/GradientKaleidoscopeGallery3.htm>  

The examples have been produced by mainly applying two basic techniques 
- described here in a much simplified way:

1. Putting two copies of an image on top of each other, setting their 
blendlevels to 50, setting a different angle for the upper image, and 
then taking a snapshot.

2. Selecting a triangle with an appropriate base width from the original 
gradient image (or: any image of course can be used) and then arrange 
these triangles to form three-sided to multi-sided polygons.

The original image need not be an exact square, but can be of any 
different width and height. The triangle elements will be adapted 
automatically to their required height and width to fit into the created 
polygon if you have for example a "wider" or "heigher" original image.

Presently I have implemented scripted procedures in the tool stack to 
produce

- a triangle-polygon, consisting of three smaller traingles
- square, consisting of four triangles
- pentagon, 5 triangles
- hexagon
- octagon
- rotated octagon
- hexadecagon, 16 triangles
- 32-sided polygon, and
- multisided polygons.

Up from a hexadecagon it might be better to apply an oval mask, but I 
have also included hexadecagon examples with discernable sides.

The trickiest part for me was to find the algorithms for the 
triangle-polygon and the pentagon, e.g. for the triangle-polygon you 
have to compute the relative height of the smaller 3 triangles to fit 
exactly into the bigger triangle (which is roughly "width of the 
triangle divided by 3.47") and then you have to find a way to position 
the triangle parts in the script to form the respective polygon, which 
is somewhat difficult because rotated images with an angle other than a 
multiple of 90 have rects of a different size. Finally, you have the to 
find out the rect from which you need to take the final snapshot from 
the grouped smaller triangles.

A number of examples especially in Galleries 2 and 3 have a somewhat 
"naturalistic" look because they display certain irregularities that 
were brought about by different - and often repeatedly applied - 
algorithms, like skewing, mirroring, changing color values, 
enlarging/stretching, despeckling, additional gradient overlays, 
lithography, sharpen, and smoothing filters etc..

Concerning the crashes:

Quite regularly, about every three days, but sometimes even more 
frequently, I get crashes with Rev.3.5.. Stack and IDE vanish at once 
and completely, in rare cases I get a warning from the Windows system 
that the Rev.exe file has problem (when working on my Mac, I get the 
same kind of crashes). I have had the case of one more corrupted stack, 
which however was intact  in so far as I could retrieve the scripts from 
the stack with a word processor (In another such case I had reported 
earlier a substantial part of the corrupted stack had even be deleted - 
and another - not corrupted - stack, from which I had try to copy 
objects had been partially damaged, i.e. a number fo objects were 
missing afterwards including such element I had *not* tried to transfer).
I have not yet been able to find out the exact cause and its location in 
the scripts for these crashes. I have the feeling that somehow "toxic 
waste" is accumulating in stack and/or engine which then causes crashes 
when a certain level has been reached. It may well be related to the 
frequently applied procedures of various kinds of snapshot-taking, 
masking, setting angles, and grouping or regrouping of images.--

I hope for the next version of Rev to offer a better platform for such 
tasks, and I need such a safe basis before I go public with my Gradient 
Kaleidoscope stack.
Another reason for the delay is that due to other predelictions and 
obligations programming with Revolution is now less at the center of my 
interests than it has been before.

Best regards,

Wilhelm Sanke