Hi Richard (regarding L-systems & IFS types in Fractint, & their future development), Sorry it's taken me so long to respond. I've been away from the internet a good deal this summer, in the high Caucasus mountains of the Republic of Georgia... I've discovered a programme which seems to be the LS equivalent of Apophysis in 2d or XenoDream in 3d: Context Free Art http://www.contextfreeart.org/ I describe it thus because it allows LS programming plus a lot more non-LS work, and because its language is at least related to LS. Are you aware of it? As far as IFS goes... one difference between Apohysis IFS and that of Fractint is a great feature of the latter about which I have not found any literature, so I might as well claim credit for discovering it unless someone else can correct me. This is that if you start off with a shape like a square filled *solid* with holons, and then simply alter the numbers in the 7th column - those which deal with probability - you end up with fantastic fractal textures instead of the solid fill. (These probability numbers should still sum to 1.) I have done a huge number of explorations of squares in this manner, some with triangles, and also discovered new textures for the Dragon Curve and TwinDragon. I was waiting for the new Fractint version to reveal this, but now seems a better time to point it out, in the context of a plea not to do away with such a possibility. (It doesn't seem to work in Apophysis, or at least works differently; I haven't tried it in XenoDream yet.) I don't understand all of the details of your letter below about IFS, but am eagerly awaiting a demonstration of what it all means. Tony Hanmer On 8/1/07, Richard <legalize@xmission.com> wrote:
I'd appreciate input on you regarding the L-systems stuff I posted to the list recently. For L-systems I want the code to ultimately approach the realistic plant rendering capability of Prusinkiewicz's code. We're a long way off from that, but that's the target.
For IFS, right now the code only supports probabilistic 2D/3D IFS. I want to add more IFS types like language restricted IFS. For rendering, I want to use OpenGL to give interactive manipulation of the IFS rendering as well as a design interface that's better than editing transformations coefficients in a text file. I also want to use "point sprite" rendering for IFS which will give them more "body" with fewer iterations and still be interactive. I've also thought about doing a potential rendering technique where rendered iterations define a charge in space and the rendering draws the potential resulting from such a charge. Finally, I want to extend the text file specifications so that you can specify the transformations in a transform language rather than as just raw coefficients. That way you could naturally right s(sqrt(2)/2)*r(180)*t(0.5,0.5) to mean "scale uniformly by sqrt(2)/2, then rotate by 180 degrees about the Z axis, then translate by (0.5,0.5)". I've come across a very nice recursive descent parser framework for C++ -- the spirit module in the boost library -- that will let me make a number of improvements across the board in fractint, both for formula types and for IFS and L-system types. -- "The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download <http://www.xmission.com/~legalize/book/download/index.html>
Legalize Adulthood! <http://blogs.xmission.com/legalize/>