charting a course through the IT industry
the journal of Michael Werneburg
twenty-seven years and one million words
For many years, the information technology field had major problems to solve. Nowadays, we're very keen to unlearn from our experience.
The earliest drivers in IT development were real problems like "how to model and instantiate business workflows" in an effort to improve productivity, reliability, and accountability while at the same time reducing duplication of work. Slowly but surely, some techniques and technologies became obviously critical.
Among the chief technologies was a very simply software development life-cycle (SDLC) that started with a defined problem, broke it down into a stated architecture, and then built to that architecture. Much the same way you'd build a house, design a city, or tackle any project.
Among the technologies that won out were the relational database. This complex software allows users to safely store and retrieve data reliably.
Real SDLC's died an early and painful death in the 90's when a tidal wave of "web developers" got into the act and simply short-circuited the entire thing. Now they talk about themselves being "agile" and able to quickly produce results. It no longer matters that the results map to the requirements or that anybody understand what's been built. This is okay, it seems, because we no longer seem to have the slightest twinkling of care about our own future needs. Because all that matters now is "cheap solutions".
A lot of software development is now done on an anonymous basis without any connection between the business and the software developers. The latter are likely housed in a remote location in an area where the same language is not typically spoken. And the software creation is divorced from any expensive business analysts because those types were shown the door years ago. This allows solutions to come quickly and cheaply (can't beat those rates from country x!).
This availability of cheap results let the business believe they have a functioning software team. So they carry on, happy to have today's solutions in place. Kinda.
But without the input of software specialists who understand the business, there's no one who's really looking out for the long term interests of the business. The business users don't understand this, and they don't see the danger. But they've made their numbers for this quarter, and these days, there's every chance that next quarter's problems really are someone else's problem.
So things continue for a while, with the result that the business systems are simply a massive collection of hacks that once solved a host of problems but now for some reason it can't be bent into shape to fit today's problems.This is no longer just the domain of the web developers; it's permeated dangerously throughout core systems such as those running our financial systems. I know because I've seen it first hand. Time and again.
It was this trend that convinced me that there was no future for me in software development. I went first into database systems, then into technical infrastructure, and finally into process design and implementation. barely IT at all anymore, my last job resembled a switchboard operator and author of TPS reports.
The RDBMS mentioned above is now dying from the same root cause. It's all expressed quite well in this video. Build a quick fix and .. fuck it.