Interoperability Is a Lie

Posted On Sep 5 2007 by

Since I’ve said it a few times, I thought I would get it down in print before it got taken out of context or worse. “Interoperability is the biggest lie in automation today.” The word is thrown around as easily and meaninglessly as “friend.” Interoperable is, at best, an adjective for standards-based systems, and at worst, a hack to cover up the fact that different systems are not at all meant to speak to one another. The former case is so rare as to make it the exception; the latter case is perpetual job security for systems people.

Function: noun
: ability of a system (as a weapons system) to work with or use the parts or equipment of another system
Webster’s Online

I’m not sure why Webster’s chose weapons when online systems would have made a good example; perhaps they wanted to emphasize the interoperability of pieces and parts, though computer hardware would have done okay for that. In fact, hardware interoperability is much less of a lie than software interoperability. Systems folk are much more used to after-market memory, cables, etc. We’re spoiled now by the (nearly) Universal Serial Bus. But the time is not so distant that we should have forgotten that hardware used to be as proprietary as most software.

Why am I so worked up about this? Well, first of all, I care about standards, and interoperability without standards is more often than not an end-run around them. Second, unlike loose use of the word “friend,” interoperability can cost time and money. It’s the same reason I object to the oft-overused “automagically.” In my experience, there’s no magic in IT—just lots of hard work. The more we describe it as magic, the less respect there is for hard work.

So what? Well, I’ve been advocating breaking systems apart for some time now. The dis-integrated library system is reality—the classic ILS, link resolving, metasearch, ERM, patron self-service, digital asset management, CRM (e.g. PeopleSoft, Banner), “next-gen” catalog interfaces. Would you believe someone who told you they had a Shangri La where all these pieces interoperate? I’m generally skeptical in thinking that any two of them can. Sometimes a little skepticism can be healthy.


[This post originally appeared as part of American Libraries’ Hectic Pace Blog and is archived here.]

Last Updated on: January 19th, 2024 at 12:22 am, by Andrew K. Pace

Written by Andrew K. Pace

5 responses to “Interoperability Is a Lie

  1. Interoperable is, at best, an adjective for standards-based systems, and at worst, a hack to cover up the fact that different systems are not at all meant to speak to one another.

    Of course we use the former (standards) to beat up on those who insist on giving us the latter (closed systems). Since we (the consumers of library automation systems) can’t rely on regulation (as in the government dictating how tax records are sent to the IRS) to make it happen and there isn’t one 800-pound gorilla (as in Walmart dictating the use of RFID tags on shipping containers) to insist that it happens, interoperability for us means we need to cajole, embarrass, or simply “route around” providers that don’t give us open systems.

    It is work, but it is necessary and generally fun work.

  2. I think I have only heard the term “automagically” used with a hint of sarcasm by the designers and administrators themselves. I think they are making light of the hard work and expertise that is involved. A lay person does not want to hear about how hard my job is or how much of an expert I am, especially when that lay person welds airplane parts together all day, and removes metal splinters from their skin all evening. I think automagically is a great descriptor. Skepticism can be healthy as you say, but so is humor, and automagic is funny.

    I think your objection with “interoperability” is a semantic one. I object to the use of the title “Systems Analyst” in place of “Chief Programmer” because systems analysis is used in much broader contexts as well. I object to the overuse of the term “long tail phenomenon.” But I am not going to change the world by objecting to how people use language.

    I think interoperability is the flexibility of a system component to communicate with another system component. Components don’t have to get an apartment together and share clothing in order to provide a useful function. If two components can send data back and forth, and even translate it in a consistent way, I would say there is a degree of interoperability going on. Perhaps, to compromise, it is more correct to say that interoperability is a standards based process–a goal that one works towards–rather than a state that one claims to have achieved.

  3. Andrew, I too find the use of “automagic” to be funny. It’s like something I heard first maybe fifteen years ago from someone at a support desk, when a piece of software was clearly going south: “Computers: labor-saving devices!” It inevitably cracks people up.

    Interoperability… yeah. But we also have a responsibility to describe “operability” in a way that is realistic. We owe that to *ourselves,* so we don’t give ammo to those who are tempted to route around standards — uh, I guess I’m saying vendors.

  4. “A little skepticism can be healthy?” You bet, but let’s direct it properly. Vendors are the symptom, not the entire problem here. As a librarian, a vendor for many decades and a long time dedicated supporter of our standards organization (NISO), I really think this profession also needs to take a look in the mirror when complaining about interoperability.

    I totally agree with Peter that there needs to be an 800-pound gorilla here. It does exist and it’s called the total library profession! Unfortunately, libraries are so schizophrenic, so unfocused, heading in so many directions simultaneously, chasing so many different goals that the 800-pound gorilla looks more like a few thousand marmosets all chasing after their own needs and desires. If librarians want interoperability, then they need to do the following:

    a) Support your standards organizations that develop these standards. (How many libraries are actual full voting members of NISO? When asked, they say they leave it to their “vendors” to represent them. Anyone besides me see a problem here?!?!?)

    b) Help your standards organization and/or professional technical organizations (LITA?) develop test-beds and published compliance ratings about interoperability. You’ll have to charge someone money for this, because it isn’t cheap or easy. But if libraries truly want this, then they should be willing to pay for it. Then vendors won’t be able to “route around” complying.

    c) Let’s get some clear, national definition of what interoperability for libraries IS. Even if every major library became a member of NISO, we’d still have problems coming up with a standard that was tight enough for seamless integration and that worked with minimal configuration, because all those thousands of approaches used by each library would have to be accommodated for the standard to get passed. Some larger definition of the goals to be achieved would minimize this greatly and help actual interoperability result. And yes, we all know technology changes quickly, so even a national definition would be under constant revision. But we must, MUST have an overall plan, to understand where we’re going with this concept and how we’re going to get there and how to focus our limited resources. Again, some division of ALA (LITA – Andrew?), maybe OCLC, but someone please organize and develop a strategic plan so we can all work towards it! Otherwise, we’re all left to define and implement interoperability as we see it – and you, as the buyers, are left hoping it all comes together cleanly and functionally, which of course, it does not in the current environment.

    Vendors typically deliver what they’re customers ask for – it is how they stay in business. Yes, there are those vendors that do things in less than admirable ways – and comprehensive compliance testing would weed them out. But if as a profession, you’re unhappy with what ALL your vendors are delivering then it might be time to look at what you’re ACTUALLY asking them to deliver (and at what cost).

    I assure you, having worked in, and leading automation firms, I’d rather write to a standard any day of the week than spend my time trying to do the same thing a hundred different ways. It’s boring, it’s tedious and it’s frustrating. Interoperability is not a lie, it’s a requirement. When it is in place we can all focus on how to do the more exciting and differentiating things that libraries excel at in the information landscape. So, let’s work together to define interoperability and how to implement it. The pieces are there, they’re just waiting for leadership and it has to come from the profession that vendors serve and participate in. We, as vendors, can play a role, but ultimately you dictate what gets delivered to you.

  5. Well, not to focus on the tree and ignore the forest, but in my experience the term “automagically” has been used for a good reason, i.e. to make what is being done seem a lot more complicated than it really is. It’s a snarky way of saying, “sorry my poor, uneducated user, but this is simply far too complicated to waste my time explaining its operation to you”.

    And, from my perspective, it tends used by people with little to no patience in working with end users.