The interface is developed in Java. It initially attempts to determine the available databases to which it can attach by asking CORBA for a list of registered the objects of class WaferDatabase (the IDL class).
The list of databases is then used to create
radio buttons at the top of the interface.
When the user selects a database by clicking on the associated
radio button, the application
obtains a reference to the appropriate remote database
(the name associated with the radio button is adequate based on the
Dynamic Determination of the
Available Databases).
Then it invokes the objects()
method for that database and inserts the resulting
lot names into the List widget.
At this point, the user can generate reports for a given lot by selecting a lot name from the list and selecting a report type from the radio buttons below the list. To generate the report, they press the Create Report button. The event handler obtains a reference to the WaferGraphics Server and invokes the appropriate method for that object as described in the IDL definition of the WaferGraphics Server.
The Plot, BoxWafer and Pareto buttons are simple methods that require only the names of the lot and database. The code is as follows (from generateReport) and assumes the the Sserver is a reference to the WaferGraphics Server, lot is the selected lot from the list and db_name is the current database name.
if(type.equals("plot")) {
Sserver.plot(lot, db_name);
} else if(type.equals("boxwafer")) {
Sserver.boxwafers(lot, db_name);
} else if(type.equals("pareto")) {
Sserver.pareto(lot, db_name);
The histogram invocation is different and requires extra steps from
the WaferReport application.
The arguments to the histogram method are the name of the lot
selected by the user (used as a label for the plot)
and an array of doubles representing the bin counts for the
histogram.
This is obtained from the database using the
yield()
method with the wafer name.
This is implemented in the code as:
} else if(type.equals("yield histogram")) {
// this is special since we compute the yield and
// pass it to S.
// Get the current database and narrow it to
// a wafer database. Then, compute the yield.
SServices.WaferDatabase db = list.currentDatabase();
double[] yield = db.yield(lot);
// plot the histogram remotely.
Sserver.histogram(items[0], yield);
}
resolve_initial_references() with
the string "LocationService" and narrow the object
to the ObjLocation.Agent class.
ObjLocation.Agent method
all_instances_descs()
instance_name field
to obtain an array of Strings identifying the databases.
These can be passed to the Naming Service to obtain a
reference to that instance.
How do we do this in other ORBs? More on this in the future.