Wednesday, October 10, 2007

Classifying Innovation

Masood Mortazavi (Sun Microsystems) quotes a distinction from Thomas Hughes between radical and conservative inventions.
"The system-originating inventions can be labeled radical, the system-improving ones conservative." Thomas P. Hughes (2004), American Genesis: A Century of Inventions and Technological Enthusiasm, 1870-1970

Masood claims Java (originally developed by Sun Microsystems) as a radical invention. But what exactly is the system that Java has created? (Obviously we're not just talking here about software systems written in Java - otherwise every minor programming language would count as a radical invention.) Does it mean something like "ecosystem"? And does one's view of this system depend on one's position - inside the Java world or outside?

Perhaps it's noteworthy that I can talk about a Java "world" at all. Perhaps this means that the Java world itself is "The System" for the purposes of Hughes' definition.

But I think the existence of "Systems" or "Worlds" in this sense is pretty subjective. I'd prefer to interpret Hughes' distinction as a spectrum (some inventions are more/less radical than others) rather than a rigid classification.

But there is another more fundamental problem with Hughes' classification, which is that it appears to confuse invention with innovation. There is often a huge gap (mental as well as temporal) between the invention of a device and the emergence of an innovation. The world of recorded music may be traced back to the invention of the gramophone; the world of telecommunications may be traced back to the invention of the telegraph or telephone; but the emergence of these worlds cannot be attributed solely to the invention of the device. Some of the inventions that powered the industrial revolution (including the steam engine, invented by Hero of Alexandria) were known to the ancients, who regarded them as toys and failed to appreciate their radical potential. The iPod doesn't get a Nobel Prize, despite the protestations of Fake Steve Jobs.

As for Java, I'd prefer to regard it as an innovative synthesis of earlier inventions. Object-oriented languages were invented in the 1960s (I learned Simula at college in the 1970s), but the OO world really only emerged in the late 1980s.

So I think the radical/conservative distinction applies better to innovations than to inventions.

There is of course a further problem - the later innovation distorts our perceptions of the earlier invention. It is now practically impossible to view the ancient steam engine without associating it with what it became nearly two thousand years later. Key inventions are disputed (telephone, television, calculus), and the invention itself becomes a social construction.

Which means the classification of inventions and/or innovations becomes an act of interpretation (hermeneutics). Does labelling something as "radical" tell us anything useful, or is it like ranking Reubens above Ingres?

No comments: