Sat, 07 Apr 2007

The Proper Subset

Of late, various ATE industry pundits and marketeers have been yaaaakking about how the "software is the instrument" and pointing out how Virtual Instruments are a superset of Synthetic Instruments and how they knew this paradigm shift was coming all all along.

I do agree with them -- to a point.

My reservation is this. By calling SI a subset of VI they seem to imply that a VI solution (e.g. LabView) has SI totally covered. This seems to be what some people call "greedy reductionism". The disparaging words "just" or "nothing more" are implied. Humans are just animals; Synthetic is just Virtual. They make it sound as if there was nothing more to humans than just a bunch of animal instincts, or nothing more to synthetic instruments than just a kind of VI. (Some people may think consciousness itself is just a computer program, and thereby also nothing more than a subset of VI, but my examples digress.)

My reply to this greedy reduction would be that steak is just a subset of food, but that doesn't mean you can buy a good fillet mignon at a greasy spoon diner. SI may be a subset of VI, but that doesn't instantly allow you to conclude that LabView (or any other generic VI tool) does a good job of assisting the implementation of the specialized subset that is SI. Quite the contrary. One would expect that a more specialized tool, focused on the subset in question, to do a better job than a general purpose tool. You can usually get a better steak at a steakhouse (not that there's anything wrong with the T-bones they sell at Waffle House).

If the generality of interfaces to measurement specific hardware was removed from the VI developer's toolkit, restricting their options, then the developer would have no choice but to work within the narrow SI concept. Add a rich enough pallete of measurement abstractions and the resulting real-world measurements synthesized would have a much better chance of being hardware abstract, and not bound to particular natural hardware instruments. This would lead us, in turn, to synthetic measurements that are not shackled to legacy hardware, realizing the end goal of the synthetic instrument approach.

There will be kicking and screaming, of course. Developers who are accustomed to all the freedom that a LabView or a Simulink bring them might balk at accepting a more restricted development environment. Does anybody remember when MS-DOS and Win3.1 was the standard and developers kicked and screamed about the restrictions that Windows NT would impose on them? They wanted the freedom to write software that interacted directly with hardware. They didn't want no stinkin' OS standing in their way. But customers wanted a computer that didn't crash. The only way to get that was hardware abstraction. In the end, after the debacle that was WinME, even the developers realized the error of their old way.

Measurement software technology is at the same sort of juncture. We would love to keep the generality and freedom of the general purpose Virtual Instrument tools, but ATE customers are fed up with losing their software IP investment when the hardware goes obsolete. They want a way to keep their measurement IP forever, even though the hardware gets updated periodically. Synthetic Instruments are the proper subset that reaches this goal.

Posted Apr 07, 2007 at 16:00 UTC, 560 words,  [Permalink

Buy My Book!

Synthetic Instruments: Concepts and Applications

Contact Me

©2006 C.T. Nadovich, All Rights Reserved