A look back at Microsoft in manufacturing

Historically speaking, the record might show modern-day machine control, interfacing, and data collection didn’t start until Microsoft operating systems and applications came to the factory floor. And when exactly was that? The jury still is out on precisely when, but it’s been rolling along over these past 10 years.

The timeline on the relentless roll out of Windows operating systems can be found at TheElderGeek.com.

You can’t go to a manufacturing facility of any size and not find a PC on the factory floor that isn’t running a Microsoft OS. From Windows NT to Windows CE to Windows XP Embedded, the “preferred” solution for on-floor processing is Microsoft.

It goes beyond HMI and PLC programming software. Arguably, Microsoft touches every aspect of manufacturing software that most plants and machines use.

Where were we without Microsoft? Why have we had this love affair with its software for so many years, especially during the 1997-to-present era we’re examining for Control Design’s 10th year celebration?

First Responder
I once wrote an article on Windows-based ICOM programming software for Allen-Bradley PLCs. Scott Zifferer, the visionary president of Icom ripped into me about it. “How can you tell me I’m doing something wrong when I’m the only guy out there doing Windows-based programming for PLCs?” he raged.

I’d written, “ICOM was first off the block with Windows-based programming. Is it a new way to program? That depends on whom you ask. Is it a better way to program? Same answer. Is it fast enough to be productive? Probably not, unless you have a 486/66 with 16 MB of RAM. But it’s the current future of programming environments.”

That article was published 12 years ago, not during our 10-year look-back period, but you get the idea. Look that those horsepower requirements!

Windows 2.1 actually was around in 1987, and you needed a 286 processor with 2 megabytes of RAM on an extended memory card to run it. Very bulky. 1990 was the year that Windows 3.0 was introduced, and soon after 3.1, and 3.11, which gave the masses a networking platform without having to use Novell Netware. A home-based network? That was unheard of.

Windows 98 was released in 1998-1999, hardware to run it was getting faster and cheaper, and the early adopter curve was well passed. We were off to the races.

Ten years later, the love-in continues.

Dennis Morin started Wonderware with a few of his friends back in the 1980s. Intouch was the first HMI product to run under Windows, and the first industrial application that ran on top of windows.

This is was this app that brought Windows to the factory floor, and started the whole damn thing. People wanted an alternative operator interface other than a Xycom computer and firmware-based graphics. The development of the final systems was painstaking and time-consuming. The newbies have no idea what we went through.

Wonderware made it easy, and they took over the market.

Steve Rubin, founder of Intellution agrees. “We were slow off the mark,” he says. “We didn’t think Windows would move that fast, if at all.”

The move—or non-move—to Windows made companies and broke others. Ask anyone who shot the wad on OS/2 development. IBM and Microsoft said they would work in partnership, but soon it was clear that Gates had other ides, and the bi-partisanship disappeared. USData was an HMI/SCADA supplier that was squashed when users decided that OS/2 and UNIX were no longer for them. That was their market at the time. IBM tossed in the towel on OS/2 in 2000.

Windows did not provide any top-floor connectivity until Windows NT—the first true Windows OS—arrived around 1998. It allowed for server-based applications and networking that challenged UNIX based systems.

Enter PC-based control. Windows NT became the standard for real-time control. Many companies developed networking solutions based on NT, and migrated them with clients based on Windows 95/98. Solutions allowed for systems to run as servers, and clients were on the floor. Wow.

The pervasiveness of Microsoft is as simple as realizing “they were first.” Apple had a GUI before Microsoft, but didn’t work with people who might provide guidance for product development in the industrial community. They really didn’t know—or maybe care—anything about machine builders.

The Evangelists
Microsoft presented papers at our conferences, and helped the industry understand the new strategy. Don Richardson was the application developers customer unit manager for industry. At the OMAC users group meeting in 1999 he made it clear that Microsoft was more than just HMI applications.

He suggested that from embedded devices to back-office systems (NT), Microsoft was there. Windows DNA technology, coupled with COM/DCOM began a path towards high productivity and ease-of-use gains that allowed us to accelerate our solutions, and become more productive.

The tie-in of ERP/MRP/MES systems really was where Microsoft wanted to go. This can be seen with the accelerated use of SAP and BAAN systems. Regardless of the back end, the front end is Windows. Even most Oracle database applications have Windows front ends.

DOS-based development was very different. Greg Wood, principal with Sandstone Software, suggests the move to a protected environment (windows) made everyone nervous in the development community. The methodology of using the development products (compilers) changed weekly.

But once the learning curve was climbed, he suggests that the speed at which code and products could be developed increased 10-fold. I think we’ve seen plenty of that over these 10 years.

Microsoft allowed basement companies to form, and create reasonable software because it was “easy.” The migration from DOS to Windows brought many developers with it, but many new disciples of Microsoft were spawned as time moved forward.

QNX was a superior operating system for many reasons. But you needed to write C code for the applications. While the Canadian-based company still is successful, it was considered niche. It got caught by the Microsoft wave, and reluctantly (at the time) put a Windows front-end onto its client-side applications.

Easy networking, client/server applications, MDI (multiple document interface), OLE (object linking and embedding), and DDE (dynamic data exchange) took care of the factory floor. No more proprietary hardware/software as such.

Replacement technologies already have uprooted many initial technologies as we move forward into open software, SOA (service oriented architecture), and web-based deliveries.

Flow-chart programming in DOS was very hard to follow, and kudos to Ron Lavalle for creating the first flow charting application in DOS. Moving to Windows was a dream. Mike Klein, who founded Steeplechase Software, once told me that, if it wasn’t for Windows, his MBA-based business plan would not have gotten off the ground.

And now AJAX, OPC, XML are at the forefront, but also are indicating that the dominance that Microsoft once enjoyed might be flattening out.

Lee Lane, Logix marketing manager at Rockwell Automation, thought the biggest benefit from the past 10 years in the plant itself is data integration, and the coupling of the IT departments and the controls group. This is largely due to the wide-spread use of Ethernet.

You can’t get data unless the architecture is present. I’m not saying that anyone can design a network, but Microsoft has made it easy, allowing many points of entry into the processes and machines we control and monitor to allow this integration.

And with this integration comes the availability of information everywhere, and more web-based applications. We only need an open-source (Linux) client running open-source browser (Firefox) for the client side. Where to get the information from is still Microsoft’s territory on the server side.

Bill Lydon, of Applied Marketing Concepts, suggests that Microsoft has been astute at pricing its software to allow their customers to be ‘comfortable’ with Microsoft solutions. It is only on the server side that the costs tend to be high.

He submits that “Microsoft Windows, with all its warts, delivered the most value of any other choice.”

The low-hanging fruit has been long picked. The technologies we enjoyed 10 years ago are getting old or have disappeared. Real capability is masked by high-performance hardware, and costs are under the microscope. The advent of security issues and hacks (they have always been there) has forced us into a different world.

While the general consensus is that 95% of all users accessing web-based systems and the Internet are using Internet Explorer these days, any vulnerability can turn into a big deal, and will stir the pot for users to switch to a different browser (again, see www.TheElderGeek.com).

While the tide will turn slowly and our investment in Microsoft still is high, the balance point is unknown. Most of the gurus in the biz figured that PC-based control would revolutionize machine control. It seems we’re still waiting.

Microsoft, we’re still in love, but I suspect that a rear-view mirror look over the next 10 years will yield a different landscape. I’ve been right before, you know.