Since I quoted the Bréton proverb ‘Why do simple when you can do complicated’ last week, it’s been turning round and round in my head as it seems to pretty much set the scene for all that prevents technology from working. I remember a conversation with an analyst a few years ago, where he likened trying to change the enterprise environment as the equivalent of mining. I paraphrase, but his analogy referenced the largest mineshaft in the world as being roughly a kilometre deep, compared to the 13,200km-diameter planet we occupy. So, he said, while any effort to change might feel like it’s doing a great deal, it’s equally only scratching the surface.
This is one of the reasons, if I may, that big consulting firms are a bit of a con. Not that they mean to be, but, honestly, who can name a single large company that has re-invented itself as a slick, genuinely agile business? All these terms like digital transformation are very nice, and they certainly look good when wrapped in some futuristic stock image but, frankly, it’s never going to happen. Which is one of the reasons my money is on everything-as-a-platform.
Back in the day, when I was an OO consultant, I learned that the thing is less important than the interface: once you have posted an API call, you don’t care what happens behind the scenes as long as you get the response you need, in the time you need it. We’ve been through various waves of this, from right back at Yourdon and Constantine’s 1975 work on structured programming (maximise cohesion/minimise coupling), through (indeed) OO, to component-based development, web services, RESTful interfaces and the (ahem) rest.
The point is, “the enterprise” as we know it can also be accessed as a notional platform of services. Back in the day, I had to deliver some UML training to the British Library but on the first day, I was told how the organisation had an embargo on any new IT spend: essentially, all activity had to be on maintaining existing systems. The five-day course became an exercise in working out what BL’s users wanted to do, then mapping that onto what ’services’ BL had available: it was all a facade, behind which was a set of monolithic systems. But, as I was told several years later when I bumped into the head of IT at a Microsoft event, the idea of seeing existing systems as a platform of services… had worked.
That’s not to say it’s a panacea, or magic bullet or whatever, we all know how there is no such thing (and indeed, file “Technology X is not a silver bullet” alongside “Tech Y is dead, long live tech Y” and other rinse-and-repeat articles that do the rounds about, well, every new trend). And indeed, whatever you do, don’t let any such notion fall into the hands of the enterprise architects, who will map the heck out of it and end up with service catalogues so thick, you could prop up a car with them. But seeing existing, joyously legacy technologies through the lens of an interface means you can let them be for a little longer, putting the effort into what you can build on them rather than trying to change them wholesale.
Just a thought. And speaking of simple, here’s an article for this week.
Five Questions for: Melissa Kramer of Live UTI Free
As I note in the pre-amble, not every innovation needs to be a smorgasbord of buzzy technological terms, nor should it be: we can all learn from this healthcare example in which access to good information can go a long way towards better diagnosis and treatment of this common condition.
That’s all for this week, until next time! Jon