I just worked on the code and more recently the source release. MitchW is in the process of performing a source release. His timeline is currently unknown to me, I know I've asked him more than a few times.
You can keep waiting, I know I am. In fact I have an entire site that is waiting on the source release in order to be launched. It focuses on future development of the Terrarium. Fortunately I'm able to work on the versions of the source I have available to me, so I'll have some interesting things to release that work with the source release as soon as it becomes available. Unfortunately, since my code is based on the source release you won't be seeing anything from until then.
If you are truly focusing on AI, and not necessarily on code hosting as the Terrarium is, then you may think of replacing the exceptional portions of the existing Terrarium codebase. If you don't host code, and instead interpret an AI language you get over all of our hosting problems.
The Terrarium is an application, it is not an engine. Don't expect to get source code in your hands that is easily extensible to the point of creating new games out of it in the course of a few days. It is highly tuned to what it does, and isn't focused on being a general hosting framework. Looking at the tabs at the top of the page here, you may see TaskVision/FotoVision/IssueVision. Take any one of those applications and try to quickly and easily create a new application out of it. It won't be easy, because each the applications are tuned for a specific goal. That is the difference between an engine and an application. Terrarium is the later, an application, and you won't find it easy to implement things like new commands/abilities or auto-discovery. It will give you some great code modules that you can use in creating your own separate application (or so I hope). |