By March of 1993 I had completed Helix4, another such system of self-replicating programs. I haven't created one quite like that since, but am tempted to do so. It may seem like beating an idea to death, but there is something that is particularly alluring about that particular form of digital life, and that is the fact that all of the evolution is competely open-ended. It's like creating the laws of physics and watching life evolve within it. A system more like BugFest may be more interesting to watch, but all of the evolution and progress within the system is limited to the initial design. All that sort of program can hope to do is model a certain aspect (or several aspects) of processes in the natural world, such as natural selection, population growth cycles, regional diversity, predator/prey balances, and so on. A more open-ended system doesn't specifically create these models and impose them on the simulated creatures, but rather provides an environment in which creatures can develop on their own to express behaviours that were *not* specifically designed into the system. Tierra, Helix4, and systems of their sort are much more open-ended as such, and also offer a lot of options from the designers standpoint, such that each one of these implementations may have strengths and weaknesses not expressed in other similar systems. Additionally, I'm not completely satisfied with my previous attempts at Tierra-like artificial life. When I go back and look at Helix4 (and even more so when looking at Helix1), I see a multitude of things that I would do differently now.

However, I shall not limit myself to any one form of ALife. I'm also trying to put together ideas for a system in which creatures behaviour and physiology are completely generated by a genetic code. This would place it somewhere between Helix/Tierra and BugFest in terms of its design philosphy in the sense that it would attempt to achieve a certain level of open-endedness that is absent from an ecological simulation such as BugFest, but it would also attempt to "look" more like an approximation of life rather than just watching a bunch of colored bars representing blocks of memory that are in use. Please keep in mind that my attempts at creating digital life are not entirely geared towards proving any scientific theories or studying any particular phenomena, but rather are an artistic expression and a learning experience for me. And what do I learn? Each successive ALife project tends to do several things for me. I get more practice in abstract visualization of complex structures and interactions, and I tend to develop new methods for storing data and processing data more efficiently, and I also create code that is often of use in my other programs. But I also learn a few things about life too. It's amazing how often you can see something unusual happen in one of these little ALife models, and then see the same process described on a nature show on public television. :-)

And thus the work will continue, and in many cases it will continue to be inspired by the projects that I read about or see myself. Sometimes an interesting and completely new idea can be generated by building on the ideas that have come before you. Sometimes all it takes is simply getting into the right frame of mind.