bioMoby - Automated registry and discovery of Web Services
There are still
some problems with web services. One problem is simply finding them.
They may be out there, but how do we know that they exist, and how to
use them? The second problem is knowing what type of input the service
expects, and what type of output the service generates. BioMOBY [http://www.biomoby.org/] creates a
framework for the automated registration and discovery of web services.
One of the most innovative ideas behind BioMOBY is that services
are discovered based on the type of data that you are using.
A client program
begins by creating an XML description of the data. The description is
sent to MOBY Central, which maintains an up to date registry of all
MOBY services. MOBY Central returns a list of services that can work on
the type of data being sent. The program (or the user) decides which
service to call, and invokes that service. The service processes the
data and returns the output, again in MOBY-compatible XML. Since both
input and output is XML, you can go through many cycles of
analysis using BioMOBY.
Note: This browser client is NOT MOBY. It is only a simple interface to
MOBY. Anyone can write a client program that uses MOBY.
Using almost any
piece of data as a starting point, bioMoby makes it possible for the
user to pipeline data from one object to the next. At each step,
bioMoby discovers services appropriate to a given type of data, which
return a choice of objects for further discovery.