Video download link | md5sum 7ed7e54f0466c32647944ab8cbe47e68



Summary: Lesson of the week (or the month) is, stop letting corporations break what already works just to introduce something newer (at risk/cost to the clients, not those corporations that use early adopters as unpaid testers)

OVER the years we wrote many articles (e.g. [1, 2, 3]) that explain how simplicity... well, basically keeps things nice and simple, whereas unnecessary complexity or early adoption of "new things" means risk. This is common knowledge that many people, especially mission-critical operations, have come to expect.

More than a fortnight ago BT said it would move me to fibre-optics (no more copper) and it put in a lot of effort to convince me to go along with that plan. I accordingly scheduled downtime on numerous occasions; I foresaw completion more than a week ago. But it all went very wrong (see my series; it has 6 parts: Part I, Part II, Part III, Part IV, Part V, and Part VI). This morning I got fed up by it (among other issues) and the cancellation was confirmed a couple of hours ago. I wished I had never agreed to that foolish plan, which to BT was about cost savings (to BT, not to me). The lessons learned may actually extend somewhat to debates we nowadays have in the software -- not the networking/transit -- world/domain/field. They try to make the Web all about bloated frameworks and the same goes for programming and operating systems. Simplicity, or sufficient abstraction to help us know what's going on under the hood, is being abandoned in the name of "power" or "security" or supposed "ease of use". But we end up with more problems than before, the entry barrier is raised even further (discouraging so-called 'inclusion' and 'diversity' because training and experience levels go up a notch) and many practical things get worse, they're not improved. These days many security issues in systemd are falsely attributed to "Linux" and in the past week even weak passwords were being blamed on "Linux". It's astounding, isn't it?

GNU is still minimalist if not 'brutalist'. Except GCC, which is absolutely massive (because there's lots of hardware support, hence the same problem as Linux). This atomic UNIX mentality or modularity helps keep freedom in fact; more people can participate. ⬆