The very first text editor 3689

In an abstract arm-waving way, that is a very plausible argument, but it is not what happened with my physicists. First they are much younger than that. Second, there really was somebody or something at CERN at the time they were there that caused them to become excellent at VMS. Their research was always pushing the edge of what you could get done on a VAX before the heat rest of the universe, so they ended up as excellent designers as well as coders. Everything was efficient by design. Unbelievable as it seems, CERN was an excellent trade school. Even better for this employer, many other employers thought they were over-educated and a bit scruffy, so they were very happy to get an interesting job at all, and I was delighted they were affordable.

suspect that in some cases (here in UK) their refusal to unlearn was they were being taught drivel. At the time (1990's) British CS courses appeared to have been designed at the lecturer's whim. Students would get something like functional programming shoved down their throats with no prac, no textbook, and a set of notes printed on fanfold. It was the fashion for each lecturer to invent his own psuedo language, as if Pascal or MIX (nobody taught in MIX) were not good enough. I could go on, but I'd sound too curmudgeonly.

I think your second point is good, that unsuitable candidates went into CS because it was fashionable. And that is exactly what I was on about in my previous post to Barb. Their uni did them no favours by pbutting them. Trade school or not. You would think "B-tree"and "keys per bucket" would mean something to a CS graduate wouldn't you? Yet we had to give lessons in that stuff before shoving them into working with RMS indexed files. ("they" were customer's people seconded to us.)

Universities could learn a thing or two from the driving test then. They do no favours to their students or to the society in which they operate if the course results do not differentiate by knowledge or ability. In so failing to do so, they are committing fraud.

In particular, they are robbing their good students. It gets so bad that you have to regard a first as a pbutt, and a II-1 as a bare pbutt. What do you give your good students? A gold star stuck to their foreheads?

You are right about PhDs being able to finish stuff. I'd have expected the difference between them and graduates to be less than it was though. High intellectual ability on the other hand, should be apparent in BA results. We are talking university here, not child-minding.

Physicists were taught better programming skills that were CS students. They needed that to get their research done. CS students probably knew more about Turing machines and whether something was NP complete, but they never seemed to have any clue about how to apply it. Indeed, one of my maths PhDs was better than even my physicists in churning out practical solutions to some PDP stuff. (pickup and delivery problem, not PDP-11 ;-). You would have thought CS people would be ace at that. Nope. His technique - straight to a numerical methods text, then to set about improving the simulated annealing recipes and teaching the physicists as he went. I was one very happy pointy-haired boss.

In closing this rambling rant, I'd like to offer what was probably the best predictor of programming skill. High School results. Excellent marks at A-level in numerate subjects was by far the most reliable indicator at pre-job interview time. In a funny way, that helps illustrate how badly some universities let their good students down.

If I were brave enough, I could have hired arrogant teenage whizz-kids before their CS courses screwed around with their heads.

I think CS has improved in most UK places since I had first-hand experience with their output. I sure hope so.

I'm not totally anti CS courses. There was one that turned out some of the best programmers the world has ever seen. I'm thinking of John Lions' course at UNSW. I was lucky enough, at various times, to have worked alongside a few of his students that did the donkey work for his red book. Greg Rose, John O'Brien, Doug Crompton and Tony Martin-Jones. I hated Unix at the time, but those guys were, and are, exceptional. What set them apart was their confidence and bravery. They would take on, and succeed with stuff that would have left me a gibbering wreck. I blame John Lions. I think they might as well.