Common Sense II ... with sincere apologies to Mr. Paine

Perhaps the sentiments contained within the following pages, are not yet sufficiently fashionable to procure them general favor; a long habit of not thinking a thing wrong, gives it a superficial appearance of being right, and raises at first a formable outcry in defense of custom. But tumult soon subsides. Time makes more converts then reason.

--Common Sense; Thomas Paine, February 14, 1776

In the spirit of Thomas Paine, I'd like to offer a few comments concerning personal computers (PC's) and programmable controllers (PLC's). We've managed in the last year or so to give the market something interesting to talk about and let the trade magazines try to digest new buzzwords. Somehow we seem to have fallen into this pitched battle between PLC's and PC's. The debate over PC control versus PLC control, however, provides a misplaced focus for the real issues. This discussion is not one hardware/OS versus another; it's about whole new business strategies trying to emerge against a status quo feudal system. The current state of the controls market, whether it be PLC, CNC, or DCS, can really be summarized by looking at a structure that takes proprietary hardware and software though captive channels to customers held hostage by the sunk costs of their installed base and training.

Now the interesting thing about this is that a lot of people are comfortable in their current position. Change is not comfortable. And, if you're on top, it is down right scary! Yet the mere mortals on the office side of the enterprise have been forced to adapt to ever-changing technology since circa 1982. Somehow we have convinced ourselves that the factory floor is quite different from the office. Really? I think we spend lots of time trying to come up with arguments just to keep the MBAs out of our hair. Oh, we can make arguments that nobody lost their life because the copier went down, etc. But can we really look at the vast majority of control applications and say that the environment is all that technologically different from the rest of the world.

In many plants and factories the physical environment is even better then the office, given strict temperature, humidity, power, particulate, and static control. To be sure there are applications where life is at risk if controls fail. In those cases you need redundancy, back up systems, and physical intervention procedures to ensure that the process will not fail. For most of us, however, systems are designed to work until they fail, and then we want them to fail in a predicable manner. Getting them back up and keeping the failure frequency to a minimum is how we get into these arguments of who has the better technology.

Now in the commercial area we have cost, speed, ease of use, packaging, support, delivery, optimization, and databases. Shouldn't we take a look at the technology that has serviced the necessary quality shakeout of a high-volume, low-margin business, whose costs have been driven down at the same time the familiarity curve as risen? Shouldn't we try o take advantage of the R&D resources of Microsoft, Intel and 3Com, which every day produce high volume, high-quality technology that mere mortals make work every day? Has the commercial world finally produced technology robust enough for the factory or plant floor? Or, is it that those in the factory and plant floor environments have finally become familiar enough with the technology to at least try it? Or, is it that he proprietary price premium has become high enough that we can overcome the pain of change? Maybe the incremental cost of using leveraged technology (e.g. Windows NT, Ethernet, etc.) is now so far below the incremental cost of sustaining the sunk cost curve, that it makes sense to look at what we're really doing. Why should a scanner card with a sub 500 kHz performance cost $2000.00 when a 10 megabit Ethernet card costs $80.00?

The calculator and the spreadsheet

This is not about PCs versus PLCs. The answer to the question, "Which do we need?," is that we need both! Chances are you have a high powered computer right on your desk at work and at home. You likely have a calculator nearby also. Now why would you need a calculator when all those MIPS are eagerly waiting? The answer is that for some problems, it is easier to use. When you multiply 117 times 235, it is easier to do it on your TI than in Excel. On the other hand, when you've got a multivariate problem you will reuse or move to other places, you can't beat the power and flexibility of a computer spreadsheet. The point is, you have to know enough about your problem to know if you have a spreadsheet or calculator problem. If it's a calculator problem, chances are your small PLC will do just fine. If you have a spreadsheet problem, life will be more fun with PLC-based control.

Aha! Now you've got it. Let's just put the PLC on a PC. Problem is relay ladder logic (RLL) on a PC costs you more and does less then a PLC. This doesn't make much sense to us. We often hear the argument that everyone knows RLL. We don't think so. Explaining it to recent engineering school graduates is usually met with the expression, "You've got to be kidding!" The best description we've heard is, "Oh, this is sort of a graphical assembly language." So why not take an in-vogue approach and find a simple language tuned to the problem? Well, if you're talking about solving a process problem, flow charts appear quite logical. It's a technique that everyone does seem to know, or can be picked up intuitively. In Ann Arbor we teach this technique in elementary school.

A new marketing theme

We've become convinced that products and companies never really go away. Remember when PLCs were going to replace relays? Actually, a lot of relay panels disappeared, but the relay itself is alive and well. In fact the relay market is almost twice as large as it was in 1974 when we were just fooling around with this PLC stuff. Similarly, the argument that PLCs are dead is a little presumptuous. Our view of this PC control technology is that it is a new wave--or theme. Relays were a theme and had their own technology adaptation curve (e.g., the product life cycle). Later, this spawned a PLC life cycle curve. PC control is now a third curve and market theme.

Various business reference books (e.g., Inside the Tornado by Geoffrey A. Moore) refer to this as a discontinuous innovation! Or, it could be the dreaded paradigm shock! What we find in our travels around the market is a resignation that PC control is not an if but a when.

This is dependent on a change in user expectation. To explain this let's turn back the clock to when the IBM Selectric was king and the first word processors appeared. Sometimes it's hard to imagine life before laser printers, hundreds of TrueType fonts, thesaurus, spell checkers, and a grammar checker. But, in the early days all we had was soft typewriter. Many secretaries stayed with the tried and true, top of the line IBM Selectric--feeling that, if IBM didn't give them permission to try one of these things then they couldn't even be tempted to try! In fact, the early PC didn't do much more then the Selectric and it wasn't always as convenient.

But then IBM came out with one of these computers the technology was blessed! Later, something even more dramatic happened. New companies, software companies, emerged with stuff that suddenly let these new boxes do all sorts of cool things that the variable Selectric just couldn't match. At this point, the user expectation changed. Today, when we sit down to type something we expect it to check our spelling, have wizards, templates, fonts and so on.

The Selectric disappeared into the room where we go to make carbon copies. When control engineers expect their controls to network seamlessly, share data painlessly, import/export information in a well recognized format, be table driven, and run on ubiquitous computers, the users expectations will change! Then the debate won't be one of PCs versus PLCs, but rather new thinking companies versus the beleaguered defenders of the old. Who becomes whom will be fun to watch!

Ken Spencer
President & CEO
Think & Do Software, Inc.
Ann Arbor, MI 48108