CarbonMeta Technologies Inc (COWI)

[aries4747]

Robotics opportunities abound in software development

From Martin Calsyn's "Adventures in Robots" blog
http://robotsoftware.blogspot.com/
(p.s. Note today's entry)

It is not hard to make a comparison to the early 1980's and the onset of the PC revolution. There were lots of homebrew computers and a few commercial ones and lots of failed and failing efforts to market them. If you asked a hobbyist or the proud new owner of a Sol computer why they had one, you might hear 'to store recipes' or some similar answer. It was an excuse, not a reason. A rapidly growing number of people knew how to build and operate personal computers but nobody knew what they were for yet. You could do a few neat tricks with them, but why did you need one?

And then came VisiCalc - the application that first showed us what microcomputers were for and why we needed one. And the revolution began in earnest.

Now, I'm not convinced that anybody yet knows what the killer app for mass-market robots is yet. I know I don't have the answer. But I do know that unlike the early PC days, the step from being able to build the hardware to being able to do something useful in software is much bigger with robots than it was with early microcomputer hardware.

The real need (and thus real opportunity) appears to be in generating the software tools and software foundations upon which that 'robot Visicalc' killer app will ultimately be built. And the message to robot hardware manufacturers, I believe, is to put aside the drawings and plans until you have a deep and credible software plan that explains how people will go about doing something useful with your hardware. If you have a PC based robot that you are demonstrating by driving with a joystick - then stop! Step away from the soldering iron. We can do that with a $50 RC car and $50 wireless camera.

Even if you believe that enough of the basic software problems of navigation and behavior and human interaction are solved, you don't have a credible hardware product until you can demonstrate at least some of those techniques actually operating on your hardware and can clearly articulate how I build on the software foundation you are pairing with your hardware.

Building the hardware is the easy part. Sure, it's complicated and expensive to figure out all the intricacies of design, production and manufacturing, but who cares if what you have in the end is equivalent (and just as software-impoverished) as one I can build with pieces and parts? Yours may be prettier, but it's not any more useful.

And this is where the real opportunity for MSRS and third-party vendors like CoroWare comes in. MSRS can give robot OEM's a quick way to incorporate advanced capabilities and a build a credible story for how their customers can build on those capabilities. It gives them access to the type of software commoditization and standardization that fueled the PC revolution. It gives ISVs (like CoroWare) a place to do software innovation and a marketplace that is no longer tied to the fortunes of one robot manufacturer. And, MSRS and ISV software together with MSRS-enabled hardware make it possible for that killer app to be born and to flourish.

It's the software. The revolution in robotics is a software revolution. We can build more capable hardware than we can make good use of in software. I don't know what the killer app will be, but I do know I want to be a part of the tools and software that enables and runs that killer app.