>Did the "killer poke" reduce Commodore PET sales?
Speaking as someone who owned a Commodore PET, knew others who owned them, occasionally worked at a sort of "pre-Internet Internet cafe" with dozens of PETs available for public use, and who sold a Commodore PET into the used market during their heyday, no.
PET's were pre-consumer-Internet and pre-BBS, so if someone was in a position to damage your machine, they were there in the room with you. There was no risk of random network viruses.
Sneaker-net sharing of programs on cassette tapes made tape-borne viruses impractical, to say nothing of their likely not even having been invented yet.
People did not set their friends' machines on fire, either digitally or with matches.
> PET's were pre-consumer-Internet and pre-BBS, so if someone was in a position to damage your machine, they were there in the room with you. There was no risk of random network viruses.
This makes me wonder. The beginning of the Windows XP era saw an abundance of worms, and the eras immediately following it saw plenty of RCEs on systems where admin privilege was typical.
With only a handful of BIOS vendors and most of them having ROM update utilities that worked under Windows, I'm a little amazed that we never saw widespread malicious code that just purposely bricked machines upon the next reboot.
Beats me. I read about the PET a lot in computer magazines but they seemed rare on the ground compared to TRS-80s, Apple ][ and Atari. The only time I saw PETs in person was at a summer program at a high school in Milford, NH. That school also had a PDP-11 and maybe 20 or so VT-100 terminals plus 1 really awesome GiGi graphics terminal, with that who would mess around with the PET?
The one thing I liked about the PET was that it had a really cool set of graphic characters in the high bit character range.
The Amstrad CPC used the 6845 CRTC and there were lots of other similar warnings about changing the vertical height too much or reducing the sync widths. At the time, people gave all sorts of dire warnings about the monitors being fried. Years later, I have loads of these machines and several monitors, and the only effect I've observed with out-of-spec sync signals is that the monitor can't synchronise properly and you get the rolling screen effect. So, I think all these warnings were theoretical worries rather than based in fact.
Interestingly, it seems some people have also tried to inhibit the CRTC vsync signal and create their own with a similar trick to the PET POKE - see 7.3 on page 40: https://shaker.logonsystem.eu/ACCC1.7-EN.pdf (and a comment on page 41 that they have seen damaged chips resulting from this - the PPI is the 8255 bi-directional parallel port driver chip, used on the Amstrad for scanning keyboard, driving the sound chip and miscellaneous such as the VSYNC line).
I don't know if it affected sales or not, but I'm going to guess not. Ordinary people would never have heard of it, and Commodore nerds kindof enjoyed that it existed.
I don't remember people talking a lot about it, but it was widely known. The only time I ever heard it being brought up was as a counterexample to the assertion that "you can't actually break a computer by typing the wrong thing at the keyboard". That was said a lot to reassure normal people who were afraid to touch computer keyboards out of fear of breaking expensive equipment.
In practice, the "killer poke" didn't really live up to its name. What it did was set the monitor sync rate to something the monitor couldn't handle. That would eventually damage the monitor, but you had to let it sit in that state for a long time (putting up with the terrible squeal it caused) for that to happen.
Same thing with a TRS-80 Model 1. You could make the monitor go crazy by poking to 16407. (It would recover with a power cycle, but I don't know if any of the values would damage the monitor over time.)
I would say no because it seemed to me at the time that people understood the damage the PET’s killer poke could cause was more theoretical than actual.
I took it more as a kind of lesson that software could sometimes break out of the normal execution model it runs in, sometimes in very unexpected ways and with possibly significant consequences. (I was just a kid, so probably my mother lead me there. She was an experienced systems architect, or whatever they called it at the time, and had plenty of practical experience with machines that were a lot more raw than the PET.)
I liked the PET, and while it had a better basic, monitor, and character set, it was lacking a graphics capability like the Apple ][. That’s what reduced PET sales. E.g., a C64 was basically a PET plus graphics & sound (w/enough memory to use them) and sold like gangbusters.
The VIC-20 was basically a PET with color and no built in monitor, and it sold very well too. The key was that it was much cheaper - the PET line machines were thousands of dollars whereas the later VIC-20 and C64 were hundreds of dollars and provided great value for money.
The PET (with monitor) started a $795. The C64 was initially released at $595, which is pretty much the same considering that didn't include a monitor. But I would assume most of C64s were sold at a much lower price. I don't know its pricing history but I seem to recall the price dropped pretty quickly. My (very imperfect) memory says I got one in the early '80s about $399 and I think it was down to $299 not too long after that.
I remember sometimes getting a bright flash on the screen, after which it took about a second to go back to its normal geometry, if an ancient CURSOR magazine program briefly set the "killer bit" to be able to run some code quickly. On those ancient first generation PETs, the killer poke just caused video noise if left on, by removing the "avoid writing screen memory while the display hardware is reading it" lockout.
On middle generation PETs - the dynamic RAM, pre-6845, small screen ones, the POKE did nothing.
It was the later, big screen PETs - both 40 and 80 column - that were vulnerable. It seems the POKE - which set an input pin to output - actually prevented the vertical deflection from operating. I can see that frying something if left on for a while.
This is all from memory so some of the details may be wrong. It was known, but not considered a big deal. But were there a lot of consumer sales of PETs anyway? They seemed to be more of an institutional (school) machine. Consumers really got into it with the VIC20 and of course the C64 later.
We had these in our high school computer classes my junior or senior year and we joked about doing it (learned from either Compute! or Byte magazine or some bbs thing) and our teacher also knew about it and immediately sternly warned us not to do it. Computers were so expensive and difficult to access at that time (depending upon your age/income/location) one would have to be pretty irresponsible to do it. For reference one of my classmate's father worked at IBM and they had a rare IBM PC at home with a diskette drive, and our only other computer usage reference was the mini computers we got access to at our community college that via something like a VT(some number) with green or orange text terminals for DEC and access to some other mainframe that printed out reams of paper when we played Zork on it.
Yes, people talked and joked about it. I had a couple of friends with PET's and they would not let you near them if you joked about it. I doubt it impacted sales, but it was definitely part of the culture. Most 8-bit machines had some well-known locations that you could poke values to and make them crash in spectacular ways. My ZX-81 sure did ;-)
By the time the Apple IIe, Commodore-64 and IBM PC came out, they were just a lot better and/or cheaper and that's what really killed the PET..
which would let you drop to the command line in a heavily tricked-out version of CP/M. He told me it was unbreakable and go ahead and try, so I started up BASIC and wrote something like
10 FOR I=0 TO 65535
20 PRINT I
30 POKE I,0
40 NEXT
which got to 100 and something before the machine crashed and disconnected. He didn’t think that counted as “breaking” it but he did say he was amused that it reset the real time clock to zero.
