Most publications and even websites maintain a pretty tidy appearance with all the words fitting in nice, even columns, and articles coming to a conclusion just as they run out of space. This is a gross deception because every story has loads of worthy but unused material, and dozens if not hundred of loose ends that will likely never get followed up. Just as many stories never get covered or told, even the ones that do are incomplete snapshots.

Readers can always pursue with their own research, of course, but I can't remember hearing about any that did. Still, it might be a good idea because, even though there's supposedly nothing new under the sun, what I find is really new is a deeper understanding of what was previously thought to be completely familiar. For instance, after 56 years of showing no interest in professional sports, I just learned my parents were big New York Rangers hockey fans before I was born. D'oh! It would have been cool to go to Madison Square Garden for more than the circus.

Anyway, one of the biggest loose ends running through my coverage around several recent Control cover and features articles is a spooky sensation that Internet-fueled, server-based digitalization of process controls along with everything else may be a lot more unnerving when it actually happens than it is to just talk and write about it. It's kind of like jumping in a pool you think is 3-4 feet deep, only to discover it's 12 feet deep. Remember that one?

I mean, I've been covering the emergence of smaller, faster, cheaper microprocessors, software and networking for a long time as they've spread ever more quickly onto plant floors and into process applications. I thought I knew what was going on as programmable I/O and Ethernet networking turned former control cabinets into virtual devices running on servers, and many servers subsequently gave way to virtual versions on fewer and fewer servers.

Likewise, long-established, proprietary control systems that charge by the installed tag are beginning to give way to programmable controllers that have far less costly licensing plans. At the same time, these controllers are facing the advent of developers building far less costly controllers on generic silicon platforms like Raspberry Pi and Arduino. Ironically, the trail doesn't end there, because many of these boards may only cost $40-50 each, but I'm told some suppliers are making similar devices for $3-5 each.  

Economic and financial upheaval aside, what makes me queasy lately is that, not only are the panels and computers disappearing, but the usual interfaces and even the network itself seem to be vanishing, too. In the "Interfaces lighten the load" feature in January, Beijer's Jeff Hayes described "screen-less HMIs" like Opto 22's groov solution that handles data, but farms out the actual display task to whatever screen the user wants to use.

Likewise, I haven't had the chance to cover his organization yet, but Tom VanNorman, systems engineering director at Veracity Industrial Networks, was recently telling me about software-defined networking (SDN), which to my fevered mind sounded like a virtual or network-less network. He reports that an SDN switch can access anything on its system without needing to be tied to any kind of traditional network or virtual local area network (VLAN), and that each of these switches has its own user-defined rules and policies for better security. Ulp, here's where that falling off the railing or into the deep end feeling kicks in.

I promise that I'll try to get a handle on these newly unsettled topics and report on them shortly. I may have mentioned it before, but I still think the only way to cover, understand and work with this increasingly squishy and intangible world is to focus on where the calculations are occurring. This is because, no matter how virtual and intangible software-based monitoring and control gets, they still have to do their jobs somewhere. At least, that's my hope.

Jim Montague is executive editor of Control.