Documentation and Implementation Status

Swarm is an open ended system which is meant to grow in response to the requirements of the user base, either by inhouse development or through user re-contributions. We are therefore maintaining a list of the most popular requests (both in terms of tools and libraries) so that groups of users can recognize common requirements, make more informed suggestions and so forth:

A 'Double' Space Which could deal with notions of "distance" and answer questions of the form: "which other objects are within X radius of me"? An initial implementation of such a space has been re-contributed by Ginger Booth and may serve as a good foundation for such a space.

Complete Batch-Mode Support Swarm can now run in batch mode, which should allow the user to organize large parameter sweeps over the models s/he is implementing. However, we still need to provide adequate and standardized support for file operations (it should be easy, for example, to load the parameters of an experiment from a file). In order to do this we will provide File objects which will allow users to avoid ad-hoc coding of their file-I/O. This support will be in place well before V1.0.

More Analysis Tools The averager object can generate the Mean, Max, Min, and Count of a given input stream. We would like to add similar tools to calculate entropies, mutual information and other such measures.