Byte Magazine: Declarative Languages (1985)
archive.org
archive.org
This year, PROLOG turned 50 years!
I remember buying this one, and seeing an Amiga a week later in the mall. The Prolog article blew my young mind. How in the heck is it figuring this out, BASIC doesn't do this?!?
* not least because the industry was infinitely more (historical invocation of the word) artisanal and the work personalized, I very much want to read. If I may be so forward, it sounds like you have personal history close to this. I'm seriously interested in doing anything possible to help get history like this published with effective general reader marketing and can tap a variety of serious resources and experience.
Even Motorola gave up on the 68k line and every computer manufacturer that depended on it such as Apple, Commodore, Atari, Sun Microsystems and many others either scrambled to switch to a new CPU or went out of business.
The computer press of the 1980s tells a compelling story about the rise of the 68k but I've never seen a good account of the fall other than the account of why the BBC Micro didn't use it.
1986: Motorola mc68000 - 32 bit is beautiful
1991: Intel 80486DX2 - ugh, how ugly, those segment registers!
1993: HP PA9000 - the arrival of the RISC panther brings enlightenment at the speed of light
But seriously, that issue of Byte magazine reminded me that such mags no longer exist. Now it's just ads and tests.
https://en.wikipedia.org/wiki/Intel_i860
but they didn't throw the x86 under the bus for it's name. If they had, Intel would be a has-been chipmaker the same way Motorola is.
In 1964 IBM realized it was a revolutionary idea to keep the same architecture from one generation of computers to the next. Intel was the second company to take this vision seriously and realize it and that's why Intel not only made the first microprocessor but it is still a dominant producer today. (Alternately the Apple II has no heirs because there wasn't a progression to a compatible Apple 3, Apple 4, etc.)
Apple had a dual strategy for far longer than is generally remembered, the Macintosh was a hit but it wasn't a home run for a long time.
No intel is abandon x86 all the times. That is why until the competitor call them out they just will work like Motorola did. And that is why the cpu they have never got real update for decade. Once again amd save the x86 days.
And for intel move to risc … balk the wrong tree and give up even ARM as it is not invented here.
Just a badly managed engineering firm. Need a woman I think. Or do not ignore woman when come to innovation. And keep the family.
?
The 68K was significantly faster. Faster clock, wider bus, many more and much wider registers and wider/faster ALUs.
Dhrystone is around 20x faster on a Mac than on an Apple //e:
https://netlib.org/performance/html/dhrystone.data.col0.html
In terms of real experienced performance in the applications people ran at the time the 68k was a disappointment.
And we didn't actually do much 32 bit math on the 68K except for address calculations, we mostly did 16 bit math, and in fact only the address ALU was 32 bit (2 16 bit ALUs to be precise), 32 bit ops on data registers had to go through the 16 bit data ALU twice...which means that if you had to do 32 bit arithmetic, you could win some perf if you could express it as address calculations (LEA, I am looking at you!)
And of course 16 bit math was common enough that the Apple II included a virtual 16 bit machine in its ROMs, for code-density purposes, but with an obvious further speed hit [1]
But even comparing 16 bit instructions 1:1 on a 7 MHz 68K (Amiga, Mac) with 8 bit instructions on an Apple II the 68K is faster, and the 1:1 comparison doesn't make sense because the 68K has so many more registers and more powerful addressing modes and of course does more work per instruction, and and and.
So not really sure where this idea of slow 68K came from.
Maybe it is due to Wirth's law: "Software gets slower more quickly than hardware gets faster". The 68K was so much faster that people were much more ambitious in what they tried.
3 registers. In two modes. None can hold more than 16 bits. Sometimes only 8 bits. Direct page (nicely movable) so you can .. sort of... have more registers. But none of those have an ALU and they take more cycles than a register.
You can address 24-bits, but never hold a pointer to any of that in a single register. So real fun doing framebuffer operations, etc.
Writing a compiler for it totally sucks.
The 816 looks great on a spec sheet until you actually try to write programs for it.
People did neat things with the IIgs. But I'm pretty sure they did it with their teeth clenched.
Also, the big reason the 68k eventually fell out of favor was because it (1) wasn't ready on time for the IBM PC, and (2) couldn't keep pace with Intel on the low-end.
Intel made similar blunders (i960, i860 among others). But while they were making those blunders they never killed off their x86 line, and continued to work on it. And they threw enough engineering $$ at x86 that they pulled the Pentium out of their hat and made x86 really fly, despite not being RISC.
Motorola could of, but never made a 68k ISA equivalent to the Pentium. Freescale kind of did with the ColdFire, but many years too late.
Anyways, it's really apples and oranges when talking about 68k applicability in those old home computers.
Yes, the 68k was not nearly as cycle efficient as a 6502. Memory access and interrupt responsiveness was way lower. And for the Amiga (and Atari ST), yes, the 68k may not have been the ideal CPU in that sense.
But if you look at the 6502 instruction set... there's just no reasonable way to make it 16-bit -- let alone 32-bit -- without major hacks. The 65816 does it with terrible hacks.
There really was no other option at the time other than the 68000 or x86. The 65816 (in the IIgs) came later, and was as I said, a hack job. x86 was gross. NS32xx was full of bugs and probably had the same laggard cycle efficiency. Z8000 was also weird and segmented and probably slow.
68000 got you 32-bits, big linear address space, and seemed to have a future. And the Mac was using it, so.
I loved writing assembly for the 68k. The big endianness is weird in retrospect, but it was fine at the time. Once they brought out the 020, 030, etc and got a cache in front of the memory bus. it really started to fly, performance wise. It's a shame that they EOL'd it.
PowerPC was nice, but ended up being effectively abandoned as well. And the move from 68k to PowerPC was a real struggle for Apple. MacOS "Classic" was really hardcoded for the 68k and I can't imagine the millions of man hours that happened at Apple to make that transition. System 7.whatever on PowerPC was crashy as hell. It took them years to make a stable system out of it, and by then OSX was almost done. Then they only got another 5 or 6 years before Freescale just dropped the ball on them and they had to jump to Intel.
Now you know why they'll just make their own processors now.
I remember being partial to the 6809 - It seemed like Motorola was able to do the 8bit (6800) to 8/16bit (6809) transition with that processor pretty well. But IIRC the only "mainstream" system using it was the Radio Shack color computer. I had the z80 based Timex-Sinclair color computer (I think this was known as the 2068) but I wanted to play with 6809 assembly programming so I built an expansion card with a 6809 on it and created my own 6809 assembler in Z80 assembly - I could assemble 6809 code on the TS and then send the resulting program over to the RAM in the expansion board and signal it to run the program and then have it send results back to the Z80 side - I remember being so excited when I got that all to work (Oh, to be young again and have time and energy for such projects :)
> I loved writing assembly for the 68k.
Later on I got an Atari ST and bought the full development system (C compiler, and a bunch Xeroxed docs). 68K assembly was indeed so nice - it's just sad that Intel won and we were stuck with x86 ISA for so long. I really never got into assembly programming for x86 because it was so ugly by comparison I just couldn't bear to look at it.
Just shows you Motorola's bizarre self-destructiveness. No continuity from 6809 to 68k. None from 68k to 88k. None from 68k to PowerPC. Then dead.
The 68k ran rings around the 6502 in terms of speed, addressing (well, duh), bus, memory access, ... basically, everything.
The 16 bit Intel competitors were technically inferior to it but... they won out, for a variety of reasons I won't get into here.
The Amiga was a superb machine that I loved with all my heart, but then, Windows 95 shipped and... basically nuked the entire field.
And I don't see how one could expand the 68k into a 64-bit design, but that was still a few years down the road by the time the 68k was in decline (sad to see, I was a fan of the 68k architecture).
The hallmark of the Apple II was its simplicity. A person could understand all of it, from the hardware up, quite easily.
The Amiga, OTOH, is a very complicated machine. There is no denying it was an incredibly capable machine back then, but, ultimately, its complexity became an issue when VGA and PCM sound became a thing on PCs and the ISA bus made it trivial to add new capabilities to a PC while Commodore had to push out a new computer (or, at least, a new chipset) for the same capabilities.
Also seeing "Computing at Chaos Manor", Jerry Pournelle's column, in the ToC gave me the warm fuzzies. He always came across as, uh I don't know, likable? Like somebody's magically technical super-nice granddad/uncle or whatever. When I read Byte I had no idea at first about his books, all I knew about him was what I gleaned from the columns. So weird. I guess in a way he was an influencer waaay before the term even existed? :)
And Niven isn't?
Niven said a good way to help hospitals stem financial losses is to spread rumors in Spanish within the Latino community that emergency rooms are killing patients in order to harvest their organs for transplants. “The problem [of hospitals going broke] is hugely exaggerated by illegal aliens who aren’t going to pay for anything anyway,” Niven said.
I think he knew more than me about computers (not hard, because I was living in a hut in Africa, without running water or electricity, most of the time I was reading Byte), but much less than a lot of other people who could have written a better column.
https://dl.acm.org/doi/pdf/10.1145/800087.802799
I worked on the HOPE interpreter (using POPL-2) when I was a post-doc in Edinburgh.
Like, they have schematics! And talk about expanding the 640k barrier!