So, it all boils down to these checkpoints:
- Figure out what paramaters will be needed.
- How do the parameters relate to each other.
- Draw faces using these parameters.
- Create the GUI.
- Make seed values work.

It’s working like a charm!
Unfortunately, Illustrator seem to have some problem in displaying a HUGE number of particles moving about (go figure!). On these screenshots there are somewhere around 500 particles. Haven’t tried out just how many there can be at the same time but the frame rate took a severe nose dive at 2000. Hell, even 500 of ‘em little bastards was enough to drop it to barely acceptable.
I will try to find me a decent screencaster so I can post some moving picture magic. Also there are a lot of stuff I plan to add to the system, such as pulsars and vibrating particles.

NOTE: This is a continuation of my earlier post; PARTICLE, SCHMARTICLE.

UPDATE (Nov 22 – 18:50)

Oh, and a neat trick to keep the anchor point count to a minimum – and thereby, hopefully, improving the frame rate – is to draw the particles as a line between two points with same coordinates and setting the stroke cap as ’round’. This way you get a circle with just two anchors instead of four!

This will, no doubt, be the biggest undertaking – with regards to the complexity of the code required – I’ve done to date so there will probably be a lot more posts around this topic in the future as this was more of a mission statement than anything else.

Although having nothing to do with me, I found this nice treelike structure that exhibits the Fibonacci number properties (the increasing branches) which is not very hard to achieve with an L-System. At its most basic form it would look like this.

Axiom: A

A → AB

B → A

I’m fairly happy with the script as it gets the job done and it is available for download here. There are a lot of tweaks to be made, for sure, but these curves are all produced in Illustrator. This is one of my favourite grammars, the Cesaro’s Sweep. It is actually a variant of the Koch Snowflake but I think it looks like trees. Another one of many variants of the Koch curve. This is a Quadratic Koch Island. The last one is called a dragon curve. The keen observer will also recognize this as the curve from the Jurassic Park book. There is also a pretty handy Java applet by Christopher G. Jennings to play with here.

Someone scribbled some incomprehensible symbols on a wall at my work some time ago. I have no clue what it reads or if they are even real letters. Probably just the work of some bored-out-of-his-wits zombie drone. Nevertheless, they looked somewhat real to me and seemed to follow some kind of inherent system. I let this percolate for some time until finally, the other day, I started to sketch out some very basic glyphs, as I had my morning coffee at Wayne’s sometime around two pm.

I quickly came up with 15 different types of symbols to be used as pairs, one upper and one lower. I subsequently organized them in a table and had the color coded like this.

The blue are glyphs that never will be used for reasons I will go into further down. Red and green are glyphs that are duplicates and the gray are totally unique. As my script works now I use all but the blue glyphs.

Now, the script just randomly picks two glyphs and puts one on top of the other. Nothing more than that. Well, other than I don’t need to have them pre-made as illustrator objects as each one of the 15 glyphs just exists as sequence of four ones and zeros in an array. The numbers tells the script what side, or sides, of a square will be drawn. Get it?

Well, like this.

First image is just to show in what order the sides will be read. The second image represents a glyph that in my array consequently would read 1100.

In order to not have the script draw the same line twice – as is the case with the blue ones – I put this rule in that don’t allow the upper glyphs third side and the lower glyphs first side to pair up. Got that?

First pair is not allowed, instead there are two other combinations to choose from.

Aside from that, I put in a bunch of other features such as, glyph size, stroke width, stroke caps, glyph spacing, number of glyphs, a unique glyphs only-button et cetera…

Thick stroke, square caps and tight spacing.

Slim Stroke, square caps and tight spacing.

Thick stroke, round caps and tight spacing.

Well, if anything it does kinda resemble some sort of alien sci-fi font… for whatever time I would need one…

Jared Tarbell’s homepage Complexification is another page I had found earlier. He has made some amazing scripts using Processing and his random line, water color script very much resembles a huge city viewed from above.Anyway, I threw myself head-over-heels into this undertaking and spent two nights trying to make my own city generator and came up with this.Now, I must say that it looks fairly satisfying but it only works a fraction of the times I run it and I have no control over any of the parameters I should be able to control such as population density, map type (raster, radial or a mix of the two), road curvature, road classes and their potential to spawn new branches.

As in Introversion’s Subversion I guess that I would have to make use of some complex L-system but that is a little too much above my competence level, for the moment.

Every time the tool chooses a tile it will first pick a random number between 0-100. If the random number is less than or equal to the percentage input it will verify that no neighboring tiles require a connection and then pick an empty tile. If, however, there are tiles next to it that do require a connection it will simply disregard the sparseness. This way the percentage could be somewhat misleading but over a long enough period of time I guess it will even out.

I have recently added some functions to two of my scripts; aMaze and the Tile Toy. The maze script can now save it’s wall data to a file which allows me to use it inside my Tile Toy script creating mazes with a completely different look. The appearance of the maze is now totally based on what kind of tiles I create within Tile Toy.

At the moment it only supports ‘single file’ tiles therefore only five tiles are needed (since it is a perfect maze it has no need for a solitary tile). Actually that number is closer to four, as a four-way intersection seems to be very rarely used.

I really can’t say how this project came about. Probably happened during one of many late nights rummaging Wikipedia…

Anyhow, I read about random maze generation and was fascinated how easy it seemed. Well, it wasn’t. Not for me anyway. Thankfully I found this blog which helped me out alot.

As with all my scripts so far, the moment I get the idea, all sorts of useful possibilities pop into my head. This was no exception.

But, as with all my scripts so far, when I’m done it doesn’t seem that interesting any more. Somehow the fun part is getting the idea and realizing it!

Finally my new script works!

The idea was to make a script that would, out of limited number of tiles, create seemingly infinite variations of map- or roadlike patterns.

The tiles needs some rules in order to pair them up so I drew some temporary lines representing roads just to make it easier to understand how to match them up with each other. I soon noticed that I only needed six different variations in order to make a continous map/pattern without any of the lines ending abruptly.

No matter what tile the script chooses, it will always be possible to put another tile on either side of it and make it pair up with the rest of the pattern since all the roads lie on the symmetry axis. This is, on the other side, only true if each tile is ‘single filed’. The number of tiles needed will increase dramaticly with every additional ‘file’.

In the single file tile set above there are 16 possible ways of placing the tiles:

• Tile #1: 4 rotations
• Tile #2: 4 rotations
• Tile #3: 1 static
• Tile #4: 2 rotations
• Tile #5: 1 static
• Tile #6: 4 rotations

But, in a tile set with a higher number of files simply rotating the tile in 90º increments won’t do since it’s symmetry axis will be offset.

Evidently, reflecting a tile both vertically and horizontaly yields the same result as rotating 180º.

The script has to consider the fact that each tile should be able to both rotate and reflect.