The confluence of IT and manufacturing

One of my most exciting projects was working for an electric utility in Substation Design in 1995. I was in charge of maintaining an active list of all the materials that were needed for building electrical substations.

The chief engineer’s requirements were basic and simplistic: create something that tracked substation construction materials, like clevis joints and copper clad steel earthing rods. That was it. So that’s where I started, and I built a fairly simple Microsoft Excel 5.0 workbook to track everything. It was called Requisition Purchase Order Information, or REQPUROI for short. (Yeah, my creative naming skill was 1/300 back then; we were limited to 8.3 filenames, after all.)

Not an actual screenshot of REQPUROI, but the same version of Excel.

As I learned more about the business and worked with the engineers, I added more features to REQPUROI. I listened to their complaints about their workflow, understood how they did their jobs, and figured out how I could make their jobs easier. I learned Visual Basic very quickly. Eventually, that Excel workbook could effectively model the cost of creating a substation in terms of the materials. Further, it let the engineers identify opportunities to improve the design with a net reduction in construction costs. My eventual goal for it was to integrate with AutoCAD and the supply chain to track in real-time what materials were at the service center and determine what purchasing power the utility had in negotiating orders for more clevis joints and copper clad steel earthing rods. But instead, I was promoted and my career took off.

I was nineteen years old.

Since then I’ve worked on some really cool stuff, the most recent of which was Future Health eConnect. The company had a legacy product written in Visual FoxPro, and they wanted to do something disruptive. The president told me in my interview, “I have a vision of a new product, and I want you to build it.” So I did, with a lot of help from some great people. I hired a UI designer who had worked exclusively on the Mac and iOS platforms, and we worked side by side for two solid months to come up with the design aesthetic and user experience for eConnect.

And now I can look at it with quite a bit of pride:


There is so much of me in that product, in its visual design, its layout, its features, the way it’s delivered, the user experience. There’s a lot I wish I could improve, but I know that it’s in great hands. I don’t want to go on and on about how great eConnect is here, so head on over to the YouTube channel to see some of its features. Go ahead, I’ll wait here.

Today I was thinking about REQPUROI and I wondered aloud what it would be now, all these years later, if I were still working on it. Some ideas I would have loved to be able to do back then:

  • 3D visualization of the materials with callouts for individual components
  • Direct integration with AutoCAD so the engineer could see the components that make up a deliverable piece of equipment
  • Integration into an ERP system for supply chain management
  • Export of materials lists for the as-builts, so the engineers could verify that the materials were in place post-construction
  • A mobile version of the application with multitouch exploration of the materials and inventory
  • Product lifecycle management capabilities so the engineers could see opportunities for reuse of materials post-extraction and manufacturing
  • Utilization of ISO 10303 for exchange of information between REQPUROI and CAD/CAE as well as Product Data Management systems, internal and external

The possibilities are endless. The budget would have to be!

As an aside, I’ve always had a certain fascination with PLM. When I was at EDS, they owned Unigraphics, aka UGS, now part of Siemens, and I often daydreamed about what it would be like to work on PLM software. A few years later, I authored a proposal for Microsoft to use PLM in the development of the Xbox, but it went nowhere.

All of those bullet points up there are ambitious, to say the least. But that first one, hmm, that one’s interesting. In fact, Future Health eConnect has that feature. In the Subjective Note, there’s a 3D body diagram that allows the user to specify points of interest on a 3D model and annotate them. Here’s an example:


What’s really cool about this is you can rotate the 3D image, and the annotations move along with it, into and out of view:


I can’t take credit for this. That goes to the brilliant graphics engineer on the eConnect development staff. I just think it’s amazing and the functionality has a lot of other practical uses. It’s quite a bit more exciting than a clevis joint:


But the applicability is still there.

Can you imagine building something like a rocket and having a model of the rocket with all of its components, and the components’ components, all the way down to the nuts and bolts, and if you are missing any one of the components from a bolt to the ignition button, it’s indicated on that model? You would have a visual representation of the materials needed to build that rocket from start to finish, along with a printed report of the bill of materials, with integration into the supply chain, integration into the PLM tools, integration into the flight engineer’s laptop and the safety engineer’s iPad, integration with the factory floor systems for construction, with all of that available in a dashboard in the back office. The vice president of production development could say, “We’ve improved the time from design to test by a factor of three.” The vice president of manufacturing could say, “We reduced the cost to manufacture a rocket prototype by 11% over last year.”

As you can see, I’m very passionate about this.

At the utility, once a substation’s construction was completed, a team of engineers would inspect the substation. This process was known as an “as-built.” On the as-built, an engineer would visually inspect various items at the substation, enumerated on a checklist. He would use a spotter scope to look at the parts that were out of view or high up. The engineer tech would have a printout from REQPUROI and check off each item that the engineer confirmed was built as expected.

I imagine in a modern version of that materials tracking system, the engineer could use a handheld device, like a smartphone or tablet, to zoom in on the items and capture images, which could be analyzed on the device against a preset list of expected data points. Each substation would have its GPS coordinates stored in the tracking system, and each constructed component’s image could be geotagged. Back at the office, the data that was downloaded from the device into the main tracking system could be analyzed and spread out into the ERP system, the CAD system, and more.

With that capability now, in 2013, what can we do in 2023?

Belt buckles holding up invention

When I was in San Antonio a couple of weeks ago for the CSCW conference, I went shopping for some souvenirs for my wife. I lived in Texas for 25 years and never gave two thoughts to buying souvenirs, but there I was, buying refrigerator magnets and Alamo stickers. One thing I got for myself was a gigantic belt buckle with “State of Texas” emblazoned across it.

My belt buckle reminded me of one of my first managers from way back when my career started. His name was Bill, and he wore a belt buckle that wasn’t quite as large as my new acquisition but still had some heft. When Bill was a young guy, he worked in the mail room. As he sorted mail, he discovered that the letters and packages were bumping into his belt buckle. The collision damaged the mail and slowed him down. He realized he had a choice: he could either stop wearing his belt buckle (and face the prospect of hitching up his pants all day) or he could change the way he sorted mail.

Bill was an enterprising fellow on his way to becoming a successful engineer, and he was pretty pleased with his mail sorting method. He also felt awful attached to his belt buckle and surely didn’t want to face the enormous void it would leave. So he did what any smart young engineer would do: he turned his belt to the side. For years he wore his belt to the side so that the mail wouldn’t hit his belt buckle, even after he left the mail room and moved into engineering, until one day when his coworkers felt brave enough to tease him about his belt configuration and he returned the enormous buckle to its rightful place, squarely in the middle of his expanding belly.

I was thinking about ol’ Bill and his buckle today in the context of stifling innovation. While Bill’s innovation turned into a fashion statement, what if his cajoling cohorts had put the kibosh on Bill’s idea when he first developed it? He would have at least had to not wear the buckle, and at worst change the way he did things. And that got me thinking further: what if Bill had stopped wearing the buckle and the result was a more efficient sorting method? What if he had created the mail-sorting equivalent of the Fosbury Flop? Did one side-turned buckle prevent mail sorting from entering its own renaissance?

All dramatics aside, there is no obvious right or wrong influence and outcome in innovation. Bill saw a need and created a new way to meet that need. And while his innovation may have prevented an invention, we’ll never know. Meanwhile, somewhere in a Texas retirement home, Bill can lean back in his rocking chair and tell grand stories of all the electrical substations he designed instead of the one story of the time his pants fell down while he was sorting mail.

Remembering (and refactoring) old projects

Some of the things I’ve worked on my career—in fact, some of the most exciting and innovative things I’ve worked on—never really took off. There’s a statistic somewhere about how many software projects fail, another one about how many startups fail, and yet another about how many ideas never go anywhere. All I know is that, like the song goes, ain’t nothin’ gon’ break my stride.

Back in the early Oughts, web services and service-oriented architecture were the rage. Microsoft’s .NET and Sun’s Java platform were the de facto tools and technologies that developers and architects used to build complex SOA capabilities. I was part of this wave of excitement, and I was very lucky to have the opportunity to work on an extremely ambitious project in research & development. This project aimed to unify the enterprise services (web services and otherwise) in a meaningful way.

I found some of the code for this project on an old USB stick and recorded a fifteen minute demo, which you can sit back and enjoy. I recommend clicking through to watch it on YouTube, as it’ll be easier to see.

This is the first time I’ve shown Retrospect since 2002, and since it was a project I was enormously passionate about, all the memories came flooding back and I was immediately familiar with it again.

I also had to fix numerous bugs to get it to work in Visual Studio 2010 and .NET Framework 4. As I navigated through the code, I felt a little sad. This project was abandoned, and the original server holding its source code (mercilessly housed in Visual SourceSafe) has long since been scrapped. The guy I worked on it with left EDS years ago, and the patents we filed lie dormant, never to be granted.

Retrospect does some amazing things that were really ahead of their time, like interface introspection, code generation, interactive code development in a graphical environment, service orchestration, and more.

Looking at the code really showed me how much I’ve grown as a developer, and how much I even knew back then. The architecture uses the MVC pattern, there’s a hearty attempt at separation of concerns, and the object-orientation is ambitious, if not wholly effective.

I sometimes wish I could redo the whole thing in WPF, but I have so many other things I need to do instead.