[NeXus-tech] quick thought on application definitions

[Pete R. Jemian]

Recent workshop at DESY has me thinking and that is dangerous ...

This might be content for the "manual" so excuse me while I think out loud.

Conceptually, application definitions are similar to the Java notion of 
the interface.  In Java, "an interface is a group of related methods 
with empty bodies."
(http://download.oracle.com/javase/tutorial/java/concepts/interface.html)

As Mark K says, an application definition forms a contract between the 
provider and the consumer that certain named data items are guaranteed 
to be available.

To me then, a scientific technique might store all data items within an 
application definition.  However, we expect that a beamline will store 
its information in a component tree based on the base classes and then 
provide one or more subentry application definitions with links from the 
expected names of that application definition to the actual data in the 
component tree.  Different instruments will likely use different links 
to meet the same application definition.  In some cases, it may still be 
necessary to define reductions of data from the component tree in the 
application definition but that seems more of a detail than a rule.

Pete



-- 
----------------------------------------------------------
  Pete R. Jemian, Ph.D.                <jemian at anl.gov>
  Beam line Controls and Data Acquisition, Group Leader
  Advanced Photon Source,   Argonne National Laboratory
  Argonne, IL  60439                   630 - 252 - 3189
-----------------------------------------------------------
     Education is the one thing for which people
        are willing to pay yet not receive.
-----------------------------------------------------------