I kicked the door closed, sat down, and sipped slowly at my Evian.

I’d been here three years. It wasn’t the first job of its kind I’d taken. I’d drifted into positions like this much as I’d drifted into my career in software development in the first place.

Leaving school, fairly bright but hopelessly unfit, I’d had vague TV21 dreams of becoming a scientist — an astrophysicist maybe, probing the far reaches of the universe, or a space engineer, building and controlling rockets and spacecraft. I was bright enough for college, but a few “down-to-Earth” harangues from my dad the accountant had made me see the wisdom of keeping my options open.

I got a place at Warwick University, where I read math. It was a bright, friendly place, the math faculty at the time was sparky and innovative — the home of then-trendy catastrophe theory — and I soon found myself forgetting the ostensible reasons I’d gone there. Working my way tidily through the groves of axioms, postulates, and corollaries, I quickly hit my intellectual limits, but I discovered in myself a deep appreciation of logic and order.

In my last year I traipsed around the milk round of potential employers, trying to find something that would plug into that interest in mathematical logic. I found it in software development — which brings a wry smile to the lips of all my acquaintances who have ever found themselves staring at a blue screen with a baffling error message.

But software ought to be logical. The math underlying relational databases, for instance, accessed by virtually every Internet user every day, is pure and beautiful. There is a whole discipline called “formal methods” in which you set out what you want to achieve and write a program that is a self-proof that it will do exactly what it’s supposed to do.

That was the dream, as I began my career — first in Manchester, and then, inevitably, in London, the center of everything in Britain. When I could afford it I took a small flat in Hackney, and started a gruesome daily traveling routine by bus and Tube. But as I started work, first in the software development departments of large corporations and then in independent development houses, I soon found that rigor was expensive — less so than the cost of fixing all the bugs later, but an upfront cost virtually nobody was willing to pay.

Eventually I drifted into testing, the one place where you are supposed to be rigorous. For a while I prospered. The fashion was for development methods that were, if not formal, at least structured and so open to inspection. I would draw up my test plans covering every conceivable condition the software could take, with predictions of how it should react. I turned up errors at every level from typos in the code to compilation into machine code to fundamental design flaws — but that was okay; that was the job, and it was satisfying to make things better.

But there was a constant pressure to cut costs on testing, which higher-level managers could never quite figure the benefit of, and endless turf wars between competing teams of developers and the testers come to rip to pieces “their” code. I started to be bypassed by development managers who could boast they were delivering something of direct benefit to the end user — and who, unlike me, had significant budgets and teams to run.

Not only that, they were all tall men. It’s always tall men who get on in management hierarchies, no doubt some deep primate thing. I’m a man but was never that tall, so was stuffed from the start. My trace of a Manchester accent didn’t help, either.

And then, in the nineties, a new wave of software development techniques came along. The new languages were much lower level than some of those in the past: that is, closer to the machine. As a developer you could deliver all sorts of fancy miracles. But your code would be dense and highly interconnected: difficult for an outsider to read, hard to test, all but impossible to maintain. In the wine bars and pubs of post-yuppie London I would rage against this retreat from the mathematical high ground to a kind of medieval craftsmanship, and the lower standards it would bring. But the tide was against me, even as giant applications in the stock exchange and the health service crashed and burned, even as every user of PC software howled with rage at errors so fundamental they should never have gotten past the most elementary level of inspection.

Long before I was out of my thirties my career was stalling. I still had choices, even stable employment of a sort. Testing was never going to be fashionable, but you could hardly run a respectable software development shop with no testing effort at all.

And so here I was at Hyf. I was aware that I was really a kind of totem, a personalized embodiment of the company’s illusory commitment to “high-quality deliverables.” But I’d stayed there, for three years already. Whatever I thought of the job I had bills to pay and a pension to build up. And, just sometimes, I managed to get some work done that satisfied my need to carve order from chaos — a need, as I was going to find out, that went deep in me and my family, indeed.

If I sat up in my chair I could see the end wall of the office, a slab of Victorian brickwork capped by the curving roof of the old station structure. I was struck now how good the brickwork was compared to my father’s house. A station clock maybe six feet across was set into the wall, a translucent disc marked with big Roman numerals and two spearlike hands. The back was faced over with glass that revealed the works, which still operated. Sales types would use it to impress clients. I stared at the big minute hand long enough to see it wobble its way through two, three, four minutes. It was a relic of vanished days, I thought, days of heroic engineering. There have always been engineers in my family.

It struck me suddenly how young everybody was here — everybody but me, that is. None of them was interested in the brickwork.

The big station clock reached eleven-thirty, and I hadn’t done a damn thing all morning. In the afternoon, I told myself, I would resume the good fight. For now, I shut down the computer, gathered my jacket,

and walked out for an early and long lunch.

It was a gray day, unseasonably cold for mid-September. I bought a small orange juice and an avocado and bacon sandwich from a Pret A Manger. I walked as far as Saint Katherine’s Dock before settling to a bench and eating.

Then, restless, cold, reluctant to go back to work, I walked toward Liverpool Street.

On impulse I stopped into a cybercafй. It was half empty, despite the time of day, and the customers were either eating or chatting rather than logging on. I bought my time credits and a tall latte, and sat at an empty terminal, placing myself as far from anybody else as possible.

I logged onto my home email account, got into a search engine, and typed “Mary Queen Virgins” into the query line.

Of course I could have done this at work; most people would, I suppose. But my strict but useless sense of what was right tripped me up. I had always felt uncomfortable purloining the firm’s resources, from computer time to paper clips, always aware that in the end somebody somewhere would have to work a little harder to make up for my petty theft. Or perhaps it was just that I wanted to keep my private affairs out of the office.

Most of the results were dross: straightforward crank sites put up by religious nuts of one kind or another, a remarkably large number of churches with similar names, and the usual irritating clutter from the high schools and colleges that have developed the antisocial habit of placing the entire contents of their course materials on the public Internet, thus baffling every search engine yet devised. I skimmed past most of this stuff. I felt confident I could discard anything from outside Europe — in fact from outside the single-currency Euro zone, since I knew my father had once used euros.