Yes. Lapack was the successor to linpack and I seem to recall some of the linpack routines going back much further than the eighties. MATLAB (which existed before the commercial release in 1984) was built on linpack.
Totally feel that — “LLM chat hoarding” is becoming its own category of second brain, and it’s easy for it to turn into a pile of half-finished plans.
One thing I’m exploring is treating chats/projects as an inbox that only becomes useful when you force a tiny extraction step: turn each chat into either (a) one concrete next action, or (b) one reusable artifact (template/checklist/snippet), and ignore the rest until you explicitly pull it.
What’s the biggest blocker for you: picking the single next step, or remembering which chat had the good idea when the moment comes? Details in my HN profile/bio if you want the angle I’m validating.
The fact that the Apple II met the new FCC requirements was a major competitive advantage for Apple, and there have been rumors over the years about how that happened. The higher emissions allowance was why you saw the big shift from monsters like the Atari 800 (heavy cast metal frame, aluminum or pot metal) and Commodore PET to lighter chassis like the Atari XL series and the Commodore VIC-20 and C64.
The old FCC Standards kneecapped Atari. I think Atari would have had a much showing against Apple had they not had to have the heavy and expensive cast box inside every 400/800 and the increased cost for "smart" peripherals versus "dumb" slots. Those Atari machines are arguably more technically advanced and capable than the Apple II. The cost of FCC compliance drove up the price and hurt their market share.
I've always thought the whole Apple / aftermarket RF modulator trick was a bit underhanded.
When I was a kid, I had a green screen Apple, and I wish I knew about 3rd party RF modulators. (It didn't work with an Atari-style modulator.) I never saw a setup like that, so I wonder how common they really were.
Maybe 'typical' only in the very early days? I just never saw one plugged into a television rather than a computer monitor, nor did I ever hear about such a thing. I would have loved to play those games in color.
After the initial 2001 model, Commodore used a mix of materials, with some models made of all metal and some of a metal/plastic hybrid (metal base, plastic top), according to this website: https://www.zimmers.net/cbmpics/cbm/PETx/petfaq.html (look for "WHAT MODELS OF THE PET ARE THERE?")
The chassis (the bottom black part) was always made out of metal. But all the white part above, on the very first PETs (the ones with the rectangular keyboard) it was made out of metal, and on all subsequent ones (with normal keyboard and green-on-black screens) it was made of this plastic material. Source: Personal experience (I'm old).
One of the books I first tried to learn from was _Miniature Projects for Electronic Hobbyists_ by one Ken W. Sessions. It was really a unijunction transistor project book. You know how some people will say you'll learn maybe not the best mental models for electronics from Forrest Mims' books? (I don't get it, they seem OK to me) Well, Forrest has nothing on Ken W. Sessions. The circuits worked, though.
The TL081 family of op amps is definitely in the "old parts you should (sometimes) use" category (for price/performance) but there have been much better op amps for not much more money for a long time now. Both the now-ancient 2nd edition and the aging 3rd edition of Horowitz and Hill recommend the LF411 for a JFET-input general purpose jellybean, for example.
For learning you some analog, I would recommend the LM358 op amp and the LM393 comparator and all the old National material on them. The LM358 has its quirks so you won't think op amps are ideal but you still see them in a ton of stuff, because they're dirt cheap and adequate for a lot of things, and the LM393 comparator is still so good you really should know why you're choosing anything else - micropower or exotically high speeds or whatever.
Sure, it's kind of cringe. I've shipped stuff with 555 timers in it. I'm not proud, but I'm not telling what either.
You can do almost all of the 555 tricks with comparators and then some, and you'll learn more doing them. Check out the old National Semiconductor application notes for the LM393. You're more likely to see comparators used for little bits of analog/analogish-digital glue in professional designs.
And there's another reason not to recommend them - no one sane has used the bipolar version since the Carter administration but they're still out there and it's another pitfall for beginners.
That's the only use for one that isn't (always) a design smell - it makes a really nice missing pulse detector, better than you can easily do with comparators. But if you have the budget, a purpose-made watchdog chip or a tiny microcontroller really can make a better watchdog.
Supervisor chips are not complicated. In some ways simpler than a homebrew analog watchdog, and the good ones will handle failure modes a simple watchdog won't, like those that result in an oscillating output.
Yes, a simple purpose-made chip designed to be used in safety-critical situations, with high tolerances for voltage etc, would probably be better. Although one thing the 555 design has going for it is that a seasoned EE could take one look at the physical circuit and know exactly what it does.
But I would never trust anything that ran software for something like this.
The chips themselves add a bunch of new failure states to consider beyond software bugs, too. Maybe a bad wire or component puts too much load on the microcontroller's wee internal pin drivers and they melt into a permanent "on" state. Or a voltage fluctuation browns out the chip on boot, partially randomizing its RAM or registers. Or the chip manufacturer fixes some errata or discontinues a particular part number and now a pin you've left floating has become a hardware heisenbug. Or the wrong bit flips in your EEPROM after being in a hot machine for a few years. Suddenly a boring 555 looks pretty good. (Keep in mind, we're talking about "turn off heater after pulses stop", not "abort launch sequence if tank 3 pressure low". The latter is way above my pay grade.)
For every task you could also use a 555 timer for (with dedicated analog support complexity,) you are talking about tens of lines of user code at most.
Even if you had to do everything directly with registers, the amount of C or Rust here is minuscule.
Not if you go to the cheap "Asian brands" like you're thinking with micros, plus with your cheap micro you'll need a reset controller. And budget isn't all BOM cost.
reply