“…just making the point that software needs ongoing support not because it changes but things around it do (including but not limited to users, other systems, hardware)”

“Propeller head” info follows:

Largely, such problems are due to two trends:
1) ignoring the hard-won knowledge of the older programmers
2) using “modern”, “late-binding” programming languages

For instance, when Linux used SVR4 boot process and init, the init program itself (the “mother of all” programs that run on the system) was short (less than 100k source code) and had not needed any bug fixes for years – it did one thing, and did it well. Since Red Hat insisted on using systemd to replace init, this init process is now multi-megabytes of code, some of which runs in a version-specific run-time environment. This cancer has insinuated itself into almost every distribution of Linux – Red Hat (and it downstream “free” version CentOS), Debian, Ubuntu etc etc. You CAN get a system without it, but it is very difficult and requires considerable knowledge to make it work.
With SVR4 init, you could start and stop any service (networking, web, mail etc), add USB devices etc etc and the system would not need to be rebooted – it might require SOME aspects to be restarted, but anything not affected just kept on going. With systemd, if something breaks it (and as well as taking over the init process, it has hooks into USB, desktop bus, name resolution (DNS), system logs and this list keeps getting longer), the system needs a reboot. And the “breakage” may be because you updated an OS run-time environment, or because you got a “desktop bus” message that is mal-formed (ie “buggy” third party sofware).
All this to get a system that boots to a desktop quicker – for an OS that is largely used on servers, not desktops. I have seen quite a few “embedded” Linux systems (modems etc) and most simply use older versions because those older systems are fully deterministic and reliable, unlike the current “desktop” versions of that OS. Sure, they take a bit longer to boot up, but they don’t fall down seemingly on a whim, or because you added some hardware or service – that new hardware/service might not work, but everything else does, and you can debug, add and remove drivers etc etc without needing to reboot the system. Oh, and SVR4 log files are in “human readable” form, so you can simply send them to a printer if you want, or use any of the many text processing tools (grep, awk, perl, lex & yacc etc etc) to extract highly useful data summaries etc from them.

In short, my beef with the younger generation programmers is that they haven’t taken the time to see what is already available and what the capabilities are, and they certainly haven’t embraced the Unix/Linux “method” for writing programs – “do one thing, do it well, and do it as fast as possible”.

This from someone who has written Unix (SunOS) low-level device drivers in C, used recursive descent and table-based parsers to create interpreters of “unique” languages, and written web-sites including HTML, Javascript, AJAX calls to server-side perl scripts using database backends to automate several tasks including customer invoicing, stock control and production information systems, RADIUS server user configuration including access control lists and IP pools, an XML-RPC based customer API into a domain name registration system, a customer configurable SD-WAN service with several user privilege levels, and even an automated fault-finding system for custom SCADA system cards. For pretty much all of these, once in working order and with no known bugs, they continued indefinitely as designed and didn’t “fall over” just because of some minor OS update or hardware change – they were robust because they were designed so, because they needed to be, and because I tested them with the assumption that the end-user is both a complete moron that would do things any thinking person would realise were inappropriate, and also an evil genius who would try any- and everything to break it, just for the “fun” of breaking it.

Here endeth the rant… 🙂

” How much wiser it would be to keep our coal-fired power stations going and start to revisit some of the beserko ‘science’ on which this whole fiasco is based.”

Or even, for heavens sake, keep the coal-plants going until it is obvious (from actual usage data!) that they are no longer required. At least that would have been a “Plan B” – if renew-a-bubbles actual work for a full year and the coal plant is idle all that time, then hurrah, turn it off. If it’s still needed, better keep it going, eh?
Ah – flawed. Sensible, see? Therefore, against GovCo policy (all sides). So not going to happen. Bugger bum piss poo fart. Oh well, better get used to blackouts I guess…

“Bear – What would the compound word for “involuntary transgenderization”?”

something with “permastrapadicktome” in it for the F to M, I s’pose… Other than that, I dunno.