Don't Reinvent the GUI Wheel
When I first saw a demo of our new Data Visualization & Control (DVC) Toolkit I thought: “Boy, do I wish I had this back when I was a coding hack.” More years ago than I care to think about, and certainly more than I’m willing to admit, I frequently came across the problem of validating my latest algorithm on real data. Like many others, my usual answer was to add software test harnesses to read in data from a disk file to simulate the input, and to use fprintf to output the results to a file to then view with Excel or gnuplot. Sure it worked, but how much time did it take to implement, not to mention the static nature of the test scenarios? Want a new input data set? Better be ready to sit and type in a whole new bunch of values, or write yet more code to generate the data.
I once spent many weeks writing a graphical script language to render animated drawings of a rolling mill to show areas of concern such as possible bearing wear as predicted by anomalies in measure vibration (a very crude precursor to our advanced analytics?).
Now developers of embedded systems have the option to use the DVC Toolkit which allows for simple connection to data input that can be varied with graphical knobs and sliders, and to view the output in real time just like hooking up an oscilloscope. Want to view the images you are manipulating from cameras? It can do that too.
All it takes is to learn five function calls and insert them into the application to implement points of data injection and extraction, then design your control and display widgets using a high level tool, and off you go. The application runs on your embedded system under Linux, Windows or VxWorks, and the display runs on either the embedded system or on a remote PC connected via Ethernet.
Simple, effective and a great timesaver. Maybe I should start coding again if it’s this easy…