Good ideas tend to split & bifurcate, until the original term is considered simply common sense and the parts people care about are individual subfields that were trivial when they started. Recall that "industry" used to refer to anything that was mass-produced in a factory, which, back in the 1820s, basically meant textiles. As the general techniques proved their worth, "industry" split into subfields, such that we got the "oil industry" and "steel industry" and "auto industry" and "aviation industry" and so on.

I think a similar process is happening with software, where the "software industry" is bifurcating into embedded software, social software, payment software, etc, much of it in support of an existing industry; "software is eating the world". I suspect that we're in the very early phases of this. In the early history of the industrial revolution, most industries were ways to produce existing goods (clothing, iron, tobacco, energy) dramatically more efficiently, much like how much software today is embedded in old-line industries' line-of-business apps. But starting around 1900, you started getting industries (electricity, aviation, pharmaceuticals) that were largely unrecognizable compared to their substitutes. You're seeing the very beginning of that now with specifically Internet- and cloud-based industries, but it likely has another century to go.