Just get rid of the penny!

I wouldn't call that either compiling or interpreting. That's front-end lexical analysis and parsing, which for BASIC was not exactly a conventional parse tree, but close. It remained in memory, and that's what was read when you typed RUN. I'd have to fire up my old Apple ][ to be able to tell you for certain what the on-disk format of a saved BASIC program looked like, but it was probably stored in some hybrid of the tokenized format plus the REM comments. (I probably knew the answer to that 45 years ago.)

It wouldn't surprise me if they had considered compiling that tokenized data down to the same UCSD P-code that the Pascal system used, but AFAIK, neither of Apple's BASICs were ever ported to the p-System, although there was a different BASIC from UCSD that was available.

When I worked in the HP-UX file systems group, we had Hans Reiser come in for a day and talk about possibly including ReiserFS with HP/UX, either instead of, or in addition to, VxFS from Veritas. IIRC, this was about a year before he got in trouble and we realized just who we had been dealing with.

Our impression at the time was that he was no more eccentric than any of the other super-geniuses who worked with us at HP or who had built better mousetraps on the outside. We were hoping to swing a deal like we had with Veritas, where we got the VxFS code in source form, built it into HP-UX, and distributed it for free with the OS instead of as an extra cost add-on like it was with Solaris. In return, we received late-beta quality code, tested the bejesus out of it, fixed most of the bugs we found ourselves, then sent the fixes back to Veritas, where they could do whatever they wanted with them, which was to integrate them back into their code base nearly 100% of the time. Part of why this worked was because Veritas and our organization were physically located relatively close to each other, so face-to-face communication was easy and frequent.

We decided not to proceed with Reiser, not because he was a nutjob (which we had no way of knowing about), but because he wanted too much money, did not want to share the source, and because his core developers were in Russia and other countries and were contractors instead of employees, so they could disappear at any time. For someone who was going to be a critical third-party contributor, it looked too brittle for comfort.

Not impossible, especially in an OS environment (either monolithic kernel or microkernel and separate FS code) with well-thought-out VFS and I/O layers, but yes, very difficult. Even writing one from scratch and putting huge resources towards every known type of testing may still leave in some bug that doesn't manifest in the field for years or decades. It's especially frustrating when you're trying to hit the sweet spot of the fastest-but-most-expensive versus slow-but-cheap-and-reliable, like bcachefs does.

It's all about viewpoint and use case. I use ZFS in my NAS that's running FreeBSD-based TrueNAS, and it's a beautiful thing that I turn on and it Just Works. But if I were developing something that needed the capabilities of both Linux and ZFS, it's not clear whether ZFS and the hoops I'd have to jump through would be more work or less than an alternative like btrfs/bcachefs/etc.

Which also raises the question: Why don't we all just call it ECMAscript? Has the trademarked thing whose source files end in .js diverged from ECMAscript enough to consider them two different things?

In some fields a master's is useful, like CS, where it indicates high level engineer versus a PhD which indicates a focus on pure research. In others, like chemistry, a master's is an afterthought, and a PhD is the only thing that matters.

Definitely the right thing to do, but they may not be permitted to if the product uses proprietary third-party code, e.g., using a Wind River-derived OS that Just Works instead of porting something to whatever SOC they use.

You're so right about executives. Not only are they worried about this quarter's results above all else, they tend to be nontechnical, sometimes even the CTO. Unless they're smart (and humble) enough to listen to the people they hired to actually know and do this stuff, they're always going to be vulnerable to one-size-fits-all-and-cures-all "solutions" sold using slide decks full of buzzwords like "zero trust". (Yet the company is supposed to put 100% trust in the vendor!)

So what are people whose data centers are on air-gapped networks supposed to do? (Serious question.)

That's for a very loose definition of "educated". My daughter was in high school when Covid hit, and her junior year may as well have been cancelled. Everyone slept through their online classes or spent the time scrolling on their phones with their computers' cameras turned off. Science labs were nonexistent, or watered down so that they could be done with what could be found around the house. Everything from routine homework to large term-paper type projects were relaxed in size, scope, and firmness of due dates. Almost all extracurriculars were cancelled, as were art shows, music performances, and anything resembling field trips. The usual college-application process (Common App, FAFSA, campus visits, SATs/ACTs, etc.) was severely disrupted, and many more kids defaulted to community college. And this is in a majority-minority district that punches well above its weight in terms of graduation rate, college attendance, etc.

The younger kids had it way worse. Try getting a first-grader to sit on Zoom all day!

There is no way anyone could credibly claim that the Covid lockdown "proved" that in-person schooling in actual school buildings with human teachers is an unnecessary waste of money. I absolutely would not want what my daughter went through to become the norm!

Use ksh93 if you need floating point, e.g.:

$ echo $(( sqrt(2) / 2 ))
0.707106781186547524

[Honestly asking here.]

What's stopping a government from just reducing the amount of gold each unit of currency represents?

Also, ISTR that the US in particular couldn't go back on the gold standard because at current dollar-denominated prices, there isn't enough gold on earth to cover the amount of dollars in circulation.

Very true. I love this one-sentence definition: "Economics is the study of desires."

It's like those people who want to go back on the gold standard. They complain that there's nothing supporting fiat currency; it's worth precisely what someone is willing to pay for it (by whatever mechanism). OTOH, what is gold worth? Precisely what someone is willing to pay for it!

The funny thing is that I have some IDE drives from the early 1990s that still work fine. Their cost per megabyte was horrendous, and they were fast only in comparison to floppies, but there were fewer technological shortcuts they could get away with, so some drive models were famous for being rock-stable for decades. (Mine are now backed up to a RAID-10 NAS whose SATA drives I'm pretty sure have a lower MTBF.)