Craig's Reviews > Interop: The Promise and Perils of Highly Interconnected Systems

Interop by John Palfrey
Rate this book
Clear rating

by
3958765
's review
Jul 30, 12

Read from July 09 to 28, 2012

John Palfrey and Urs Gasser outline interoperability basically as interconnection between systems. A third collaborator—say a technologist or a system’s theorist—might have rounded out the definition just slightly but crucially to an increasing interconnection between systems. Implicit in the book's (mostly impressive) analysis is the myth of equilibrium: that some point down the line we get to a desirable (or acceptable) state, and nothing progresses from then on. Maybe the authors would object to that characterization since it’s not consciously implicit, like an imbedded ideology, but it's absent nonetheless. To be fair, if interoperability was better understood there’d be more books on it, and John and Urs do sample from a good deal of expertise, but the resulting advice—to aim for optimal interoperability rather than complete interop—comes across usually as a decision can lead one outcome almost as easily as it can lead to another or if you're not careful things can go down a more harmful route. Interop’s benefits are increased efficiency, effectiveness and flexibility, but too much interop can lead to privacy and security concerns, homogeneity and lock-in effects. Careful: “The optimal level of interop is highly context specific.”

Interoperability should be investigated in more of an evolutionary, lifecycle framework where, in pretty imprecise terms, we first see two distinct systems with no mutual or highly asymmetric information. With information exchange over time (and possibly a little goading) they reach a point where I think John and Urs would characterize things as optimal. But information exchange doesn’t stop there; it continues until there is something close to perfectly mutual information. At perfect information—i.e. complete interop—there are no longer two interoperating systems but one fused system now primed to begin the process anew with another distinct system.

And this does manifest in a different policy outlook. The way diesel pump nozzles won’t fit a gasoline car’s tank and vice versa functions as an analogy in the book for why some systems should be purposely non-interoperable. But the question for innovators (analogous to policy makers in examples of regulation) is not how to make gas pumps less interoperable, but how to make different engine types more interoperable. At the current state of technology, fine, it’s not possible, but long term considerations are critical for good policy.

John and Urs do an expert job at maintaining objectivity with one exception that might be a stumbling block depending on where you fall on the issue of internet privacy. I mention it because it’s in many ways a continuation of the last point. One of their down-the-road interop challenges we’ll face is the Internet of Things. “It is not clear to many people whether in fact [the Internet of Things] should be built at all.”; “Although the grand vision of the IoT, in which billions of things are connected with each other is still a dream (or for many people, a nightmare)…”; and “And yet despite these early successes, the vision of the ubiquitous Internet of Things—a system that connects large parts of the physical world with the digital—remains primarily on drawing boards and in computer science laboratories. Some people may prefer it to remain there, merely a concept for experimentation in universities and corporate R&D facilities.” Here's another issue where the authors think the judicious use of purposeful non-interop might be the way to go. With so much interconnection the IoT poses a huge threat to personal privacy. But privacy itself seems to be another area where there’s too little not too much interop—i.e. too little interop between private, public, professional and all the other ways we compartmentalize our lives. Part of the solution is allowing time to do what it does to social norms. It’s a question really of how much we want to delay technical innovation to allow ourselves time to adjust—and how possible it even is to delay.

The authors do a good job defending the market while allowing that even a laissez-faire government constructs a medium in which the market functions. They elaborate on this in a number of helpful ways but are almost silent on patent and copyright reform. Although it’s a subject that warrants its own book, if government is never not in the background, poorly conceived patent and copyright laws serve as barriers to interoperability in all the wrong places. (Larry Downes’ The Laws of Disruption handles this subject nicely).

Despite a few shortcomings this is not at all a worthless book. The authors are both out of Harvard’s Berkman Center for Internet and Society. Berkman is multidisciplinary enough to generate a lot of really fascinating material, but it is a part of the law school, and in the end Interop does get something of the law academic’s treatment. If you're still undecided, here's a nice little video of a presentation the authors did at Berkman for the book's release. If you're satisfied after watching it you probably won't get a lot more out of the book. If it excites you, read away.
1 like · likeflag

Sign into Goodreads to see if any of your friends have read Interop.
sign in »

No comments have been added yet.