@cstross @Shunra
Yeah, those days were full of surprises and mis-forecasts like that.
Anent newspapers:
I worked a consulting gig with a guy whose previous client was the NYT, around 1980. They wanted a *custom* word processing application, to be integrated with their custom editing and typesetting pipeline. This was before vendor-provided, reasonably standard word processors were a thing, beyond toys.
He learned several things, but the ones that stick out in my memory are:
(1) User interviews are an unreliable way to get requirements. You can implement *exactly* what they ask, and they will say, "No, not like that." This is a truism today, but in 1980 it was a frustrating thing to have to learn.
(2) Management *liked* that the skills learned using their custom application were not portable, so people couldn't switch jobs. Management equally *disliked* that they couldn't just hire people who already had those skills. This is a periodic reminder that management can be kinda parasitic at times. (Admittedly, not *all* the time, but enough to be annoying.)