The end-goal of representing knowledge in this system is to be able to analyze and understand it. A large number of different analysis programs exist within the open-source movements and it makes sense to find a way to access their functionality without having to program it within our application. We have developed a system called Neuroscience Analysis Web Services (NAWS, http://www.sourceforge.net/projects/naws) that allows developers to publish analysis programs as web services so that they may be used from within any neuroinformatics system (including NeuroScholar). Users may upload data to a NAWS implementation, execute analysis programs on their data (asynchronously if required for a long analysis run) and then return the results. The analysis programs themselves are independent of the web service framework. This permits existing tools to be incorporated easily into the system as long as licensing conditions are adhered to. We are implementing and testing several analysis methods using methods from the ‘R’ statistical analysis package (http://www.r-project.org/), extended kohonen’s maps and others. The use of web-services avoids the need for the developers of different systems to repeat the implementation of these methods in separate systems. It also encourages the use of a single representation of data for exchange and communication.
In addition to the obvious utility of having access to methods of data analysis from within NeuroScholar client systems, the development of this system can be used for to perform computations that could not be performed at the client side. The server side NAWS functionality can be extended to load balance analysis method runs that require a lot of computation power and memory across a cluster of machines. This may include computationally intensive programs or programs that are not written in Java (Optical Character Recognition, PDF file parsing or multivariate statistical analysis). We plan to extend the NAWS engine to support an even wider set of different open-source computations to a NeuroScholar client.
The NAWS client can be accessed here.
The client and server side NAWS software can be downloaded from here.