The application that calculates what we owe and what we are owed went live in 1998. It is reliable, it is cheap to run, and its replacement has been proposed and deferred in four separate planning rounds because it would consume most of a decade and most of a budget. I expect it to be running when I retire.
Four people understand it properly. Their average age is fifty-six, and two of them have told me informally that they are thinking about three or four more years. That is the real timetable my department is working to, and it is shorter than the system's.
This is not a story about mainframes or about a language nobody teaches. Two of the four are perfectly comfortable with modern tooling. What they hold is not a skill anyone can hire, it is accumulated knowledge of why this organisation's rules are shaped the way they are: which of some nine hundred customisations exist because of a regulation, which exist because of a merger in 2006, and which exist because somebody insisted and has since left. None of that is written down anywhere that would survive their departure, and the reason is not carelessness. Nobody funds documentation of something that already works.
The supplier side is thinning at the same rate. Our support partner has two engineers who can answer a genuinely difficult question about this product and both are near the end of their careers, which is not a fact you find in a service level agreement.
What we are actually doing is unexciting. Pairing two younger engineers onto it permanently, not as a rotation, and accepting that this costs delivery capacity now. Paying above the market band for the skill, openly, and arguing about it annually. Working through the customisations in order of risk and recording the reason for each, which after eight months is about a fifth complete. Removing the ones no longer needed, since every deletion is knowledge we no longer have to preserve.
The thing our industry talks about least is that the constraint on modernising an estate is rarely technology and rarely money. It is a small number of people, and their availability has a date on it that nobody puts in a plan.
– Serguey Shinder
Top comments (0)