There is neat sketch on openprocessing by Chinchbug that demonstrates the use of threads in processing. Here I've adapted it to use the convenience method thread to run a function in processing (uses reflection under the hood). Since I was using processing-2.0 no point in me trying to publish it on open processing.

Announcing a near release candidate of ruby-processing-2.0, based on processing-2.0 and JRuby-1.7.4, available at github. Included with this version are many worked examples (including many vanilla processing examples translated to ruby). This release has been extensively tested on linux, and quite near releases are reported as working on macOSX. I am particularly keen that Windows users of previous versions to give it a go. The monicker 2.0 is doubly appropriate for this version of ruby-processing as JRuby-1.7.4 and hence this version of ruby-processing defaults to ruby-1.9 (which can be seen as near proxy for ruby-2.0). 

I have created an example that demonstrates the use of the new PShape api (since processing-2.0b8) to easily create complex 3D shapes in ruby-processing see my blog.  PS: I'm still looking for volunteers to my fork of ruby-processing see link here.

I'm looking for testers of my ruby-processing fork available at github see also my blog here. Includes many worked examples from vanilla-processing release (shaders, new pshape api etc).
Features:-

1. jruby-1.7.3 (defaults to ruby 1.9)
2. all sketches are now "bare" sketches matching vanilla processing

Recently I've been messing with ruby-processing again, the adventurously inclined might like to try out ruby-processing updated for processing-2.0, that I have been working on.
However the purpose of this post is to announce (now see cf3ruby below) a library for ruby-processing that facillitates the writing of context free sketches in ruby. This library is very much based on Jeremy Ashkenas context-free.rb except it has been update to ruby-1.9 syntax, and the DSL is more in line with current c++ context-freeapplication (shape replaces rule). Tested as working with current version of ruby-processing-1.0.11. For an example see my blog.

Anyone contemplating using github to host their library be warned that they are deprecating the downloads section in 90 days (starting 11 December 2012). I've got a couple of such libraries and would like to continue to use github. These libraries have a relatively low volume of take-up and are probably for the more technically inclined anyway.  Any bang goes the idea of switching processing distro from googlecode to github, unless there is an autobuild!

Has your library started to fail with processing-2.0b7, especially when you expected a mousewheel action?
If so please vote for my requested enhancement here. I have detailed the problem on my blog here.

Since processing-2.0b5 the use of an external editor is being replaced with the ability to run pde scripts from the command line (actually as you probably already know, this a two stage process, the pde script gets converted to a java file using a pre-processor, the java class is then compiled and run).  

I have forked the joons library (from googlecode) to github  joons-fork.  I did this because I wanted to learn how Joon Hyub Lee had managed to implement sphere primitives (something I would like to do with my povwriter library). Anyway I could not resist hacking the library, so I've updated it to work with processing-2.0, and re-based it to use sunflow-0.73 (cf sunflow-0.72). But I didn't stop there so I have also changed quite a bit of code (so if I've introduced errors blame me). You should know that not all PGraphics has been implemented in the joons library and shape(PShape) in particular has not been implemented. Read about the original library here, see my experiments with new library  here and here.

Fairly recently I created and ArcBall module in pyprocessing, and I wondered if could I translate it to work with processing.py ( the reason this could be considered heresy is that the excellent peasycam library, and processing.py are both written by Jonathan Feinberg). I've published the ArcBall with my other processing.py experiments at github for anyone interested. The interesting thing I learned was that it is possible to import elements from processing into a module (for ruby processing this is possible but best achieved via a proxy mixin).

This library like my original povwriter library exports the processing sketch as a PovRAY SDL, then povray does the raytracing, if you use povray-3.7beta (currently RC-6) then it is multithreaded so you can make use of your multicore processor.  Although this version of the library is not currently available for auto-download, I have change the distribution to be a single download, with one the sample files Install.pde to install required template files and to set the path to povray.

Previously I have always recommended ( and used) the latest NVIDIA drivers for NVIDIA graphics cards on linux. However it may not always be necessary to use manufacturers drivers in the near future. I have just done a fresh install of Aptosid (that defaults to use the Nouveau drivers), Aptosid is supposed to be leading edge Debian but the default kernel is only 2.6.37, I tend to roll my own kernels so I'm using a 3.5.2 version (built with Nouveau driver options). I installed libgl1-mesa-dri-experimental for gallium 3D acceleration, and here's what I get when I run:-

1. glxinfo|egrep '(direct|glx|OpenGL)'

direct rendering: Yes
server glx vendor string: SGI
server glx version string: 1.4
server glx extensions:
client glx vendor string: Mesa Project and SGI
client glx version string: 1.4
client glx extensions:
OpenGL vendor string: nouveau
OpenGL renderer string: Gallium 0.4 on NV96   <I've got a budget priced 9500gt graphics card>
OpenGL version string: 2.1 Mesa 8.0.4
OpenGL shading language version string: 1.20
OpenGL extensions:

Anyway the 3D performance was pretty damn good with both processing-1.5.1 and processing-2.0a8 and even more heresy seems to run pretty well with openjdk.

One error that was reported in the console on a few sketches, but only with processing-2.0 was this one (also with oracles jdk):-

1. join_values_nofail:398 - failed to coalesce values

The TexturedSphere example has a comment that it could do with ArcBall rotation (else there is reversal of controls) I have today posted a modified example including ArcBall control at open processing. I have also updated the sphere to be a retained shape in anticipation of processing-2.0. Unfortunately this means to view the sketch you will need to download it as openprocessing does not support opengl sketches of any kind!

I have created a simple image processing library that can produce pixellated images (from either jpg or png files) via processing and the CF3 application.
There are some very inventive people over at contextfreeart and my idea was based on an original idea by guigui. But you should take a look at what kipling came up with using my data.cfdg file produced from a black and white image of Che  Guevara. I've published the library at github. 
Suggested method of working is to use the example sketch (which acts like an application), this creates pixellation using CIRCLES (cfdg is run from processing), you can adjust the size of the pixels whilst the app is running (key in 1 .. 9 key followed 'r') and the cfdg is re-run. When you've got something you like then you can go in for bit of offline editing of the design.cfdg (the small file).

I have made available a simple library for use with the Hemesh library and processing-2.0a6 (or more recent svn versions for the adventurous). You can download the library as a zip or if you want the source, git clone the repository.  Either way I have included an example of usage that I have also posted to openprocessing, which sadly does not currently support opengl sketches of any kind ( I thought this was/is going to be fixed for processing-2.0) and this will be a real issue once processing-2.0 goes mainline.  In my sketch I have made use of the enableShape() functionality to add style to the retained shape, which I assume to be the most efficient method and modifying ambient, fill etc. I am contemplating creating a Material class to store this style information in a convenient way before applying it to the retained shape (I think it could be convenient to create say steel, chrome, glass marble as pre-made materials as in povray includes for example), or is there a better way?

There have been earlier discussions here about using using enums in processing, now I've found a half decent reason to use them in processing-2.0 to control state. In this example at openprocessing I avoid unnecessarily updating the retained shape (Vertex Buffer Object see Andres blog) using enums as a crude state control. I hope that this sketch is useful for anyone who wants to take advantage of the new retained shape functionality. It was expected with the introduction of processing-2.0 that opengl sketches would display in the browser, sadly this is still not the case at openprocessing (although for a brief period my sketches did work there as Andres mentions).

Gerd Platl did a great job of improving perfomance of toxiclibs mesh sketches, by using the GLGraphics library here, this is great for processing-1.5.1, but sadly unavailable to those without a decent graphics card, and does not work online. For processing-2.0 there is the ability to use the retained mode with PShape. I have just posted a sketch to OpenProcessing here, that illustrates how you can use the retained mode with toxiclibs (which sadly also doesn't work on line).  However anyone who can view these sketches will appreciate the fantastic performance improvement available using newer OPENGL features compared to the original processing so I think Andres has done a great job, and it is unrealisic to expect any OLD_P3D renderer to be always be compatible.

I have returned to developing my povwriter for processing-1.5.1 (which I am developing in tandem with a version for processing-2.0). The most recent added features for version 0.83 is the addition of ArcBall manipulation of the sketch, and a MouseWheel "zoom" feature that works by adjusting the scale. For the ArcBall there is an ability to constrain the rotation to one of the 3 primary axes (x, y, z). Tested as working on linux ( calling for feedback from windows/mac users), requires an installed version of povray, preferably 3.7 beta rc5.

This is mainly for the benefit of fellow library developers, whose library depends beginRaw() / endRaw(). I have had some success in updating my povwriter library (export to povray), to work with the current svn version of processing-2.0. For the new version of processing your raw export class should inherit from PGraphics (rather than PGraphics3D). For some reason or other I found keyboard entry to be pretty dodgy (perhaps it doesn't get polled too often), so I was prompted to create my own simple button gui, this worked rather well see my blog.

It wasn't until I came to post here that I realised there had been some recent discussion about ArcBalls here:-
https://forum.processing.org/topic/help-on-learning-how-to-implement-an-arcball-style-rotation, but I have created a new thread because I'm sharing a sketch with you.

There exists a tutorial library here but that is pretty old, and further the ArcBall logic is hidden in a library. However it was sketch by Ariel Malka that I found initially (which looks suspiciously similar). I've updated Ariels sketch to use PVector, in place of a custom Vec3 class, I have also made use of the new PShape3D class (via PShape interface in case you get confused). You can see my sketch here at openprocessing, I hope you find it interesting/useful.

The beta 1.70 version of the Hemesh library includes support for export to PovRAY mesh2 objects. I have posted on my blog an examples of using the latest (svn version 108) version, which demonstrates the use of this tool. This example also demonstrates a reason why it is not always an advantage to always include "calculated normals" with your mesh2 object.