Techno-determinism alert! Shhh - nobody mention 4GLs...
I was on a conference call with some representatives from IBM last night, where they were talking about
professional services offerings aiming to help customers "implement SOAs" (I'm not even going to get started on how that phrase sounds all wrong to me - that's something for later). One of the supposed benefits of the result of these services, we were told, will be that "business people are able to compose services into new applications, extend existing services, and so on".
Woah! Hold on. Is IBM sure about this? I've heard this before - and it wasn't true then. Has anything changed in the meantime? I don't think so.
Back in the early 1990s when I was working for what was once known as a "big 6" consulting firm, my best friend was working for Oracle as a consultant. On a number of occasions he tried to convince me that Oracle's CASE tools and "4GL" (or "fourth generation language") would mean the end of the corporate programmer: business software development would be done either by ISVs like Oracle - or by business analysts within enterprises. I didn't believe him - or Larry Ellison's hype - then, and I still don't believe it now. Oh, and excuse me, but wasn't this also supposed to be an outcome of COBOL - in the 1970s? Business analysts doing all the work?
So I have a nervousness when people say such things - because they have a habit of transforming from nuggets casually tossed-about marketing-speak into that most dangerous of things: "received wisdom". And by the way, the resurgence of this thinking isn't just coming from web services and SOA - but also from other material I've read concerning, for example, Microsoft's very interesting work on
software factories and domain-specific languages.
Saying that technology riffs like SOA or innovations like software factories will lead to businesspeople getting intimately involved in development of IT systems, looks to me like a potentially dangerous example of what Ovum analyst
Gary Barnett calls "techno-determinism" - the (sometimes arrogant) technology-centred assumption that just because a technology exists to make something technically feasible, that's what will happen.
All these things are "good medicine" (just as 4GLs were and still are, by the way) - they lead to software tooling environments in which the flexibility to create anything and everything, is consciously traded-off against more productivity within a particular context - which is just what you sometimes need. But software technology availability alone does not, and never will, drive significant change in the way that people within companies work with software - and that's what is required if we are to really ensure that businesspeople play active roles in specifying and developing systems to support business goals. Businesspeople can and should be involved in software development and integration work, and there are some areas which are more amenable for them to be involved in than others. But business involvement requires something far harder than technology invention - the will from all sides (within both business and IT organisations) to drive and support organisational and cultural change.
