PROGRAM

Workshop (DEVSOFT)

Developing Software Standards for the Artificial Life community

July 25th, half day

Organizers: Clifford Bohm, Emily Dolson, Alexander Lalejini, Nitash CG, Charles Ofria

[website]

This workshop will start a community-level discussion about developing digital data standards to facilitate building experimental and analysis software and tools and to allow for comparisons of results generated by different systems. Artificial life is becoming more complex as the field advances and as computational power improves. There are already a number of issues that are difficult to address (such as tracking a single gene through a genetic lineage, evaluating lines of descent in sexually reproducing populations and other data visualizations and analyses). Many projects have solved some of these problems in individual systems, but these solutions are rarely portable and often must be re-engineered for each system.
Standards allow tools to be developed that can immediately be applied to multiple systems, reducing the inefficiencies of parallel development, encouraging the development of common benchmarks, and facilitating communication and cooperation among research groups. For example, the Robot Operating System (ROS) defines communication and data standards targeted toward the robotics community, which has facilitated massive community package development, sharing, and reuse. In addition to facilitating tool reuse, standards also make it possible to compare results across different systems (even to the point of using the same analysis tools). Further, standards could increase the incentive to develop solutions to some of the more elusive community-wide challenges. This could have the effect of helping to bridge disparate communities.

The Artificial Life community spans many scientific disciplines; thus, standards will need to be robust and will need to be able to respond to regular input from a broad audience. This workshop aims to start a conversation about identifying best practices and developing data and tool standards amongst individuals who design, develop, and use a
wide range of artificial life systems, including artificial chemistries (e.g. swarm or Jordan algebra chemistries), abstract ecologies (e.g. Daisyworld), evolutionary systems (e.g. Avida, Aevol, and MABE), and a range of other communities who express interest. To achieve our aim, the workshop will consist of a short introduction, followed by group discussion and break-out sessions.