Re: evaluation of OMG DDS as a controls protocol for EPICS

[Kay-Uwe Kasemir <kasemirk@ornl.gov>]

On Feb 15, 2007, at 10:46 , Claude Saunders wrote:
> Just out of curiosity, what other RTI S/W have you used? Other than  
> NDDS, I'm only aware of their SkyBoard product (which is very odd  
> and interesting at the same time).
At LANL, I once used a demo version of RTI's "StethoScope",
a GUI for Strip-charting vxWorks variables.

We also had one license for "ControlShell":
A Java-based IDE for drawing State Diagrams as well as data flow,
creating C++ skeleton code for vxWorks.
When clicking on the state diagram transitions or blocks in the
data flow diagram, you can enter C++ code snippets for entering the
state, processing a block etc.
The result becomes a compiled vxWorks binary.

The data flow blocks are a bit like EPICS records. The control shell
library provides the framework for initializing and processing the  
records.
Each block that you define compares to a custom record.
You can create blocks for e.g. PID and reuse those,
but I don't think you get a lot of predefined blocks similar to
records in EPICS base.

The state diagram tool was actually real nice, because you could
also use it at runtime to monitor the running state machine:
The 'active' state was highlighted etc.

So it's similar to the IOC side of EPICS, but Control shell didn't  
include
a network protocol nor client side tools beyond the state machine  
monitor.
When I now look at the RTI web page, "ControlShell" might have been
renamed to "Constellation":
http://www.rti.com/products/robotics_controls/documents/ 
ConstellationOverview.pdf

-Kay