Lower voltage diode the other side of a step up transformer?

Nope, that's what he has always said.

The post with the link will be on groups.google but not easy to find because he changed his nym a lot more in those days.

Since he uses Opera as a news client and hardly anyone else does, it may be findable that way.

It is hardly surprising that a link to a photo no longer works.

Ok, after I suggested P Hucker had an IQ of 75, you replied with "No one with 75 can manage an honors degree in physics. . "

I therefore assumed his degree was in that subject. Microelectronics is more an engineering degree.

He's had a few nyms, but most of his posts were under Wilkinson, Macaw or similar. I've had a look. There was a suggestion that he posted it using dropbox but can't find any post that includes a dropbox link to a degree certificate.

I still don't see how he got a degree in microelectronics. Perhaps he told them he was on his deathbed and awarded one in sympathy?

Irrelevant to what that uni chose to call his degree.

More than a few in those days.

Most is irrelevant, what matters is which one he used for that post.

I don't recall which he used and I don't think he ever used dropbox then.

Your problem.

That's pathetic.

Maybe not, his use of Opera is relatively recent.

Maybe <honours physics> would find it. That can't be too common in uk.d-i-y

Probably doesn't work in American, where they think they got bombed on the 9th of November.

My Bing AI friend (who is American) says that it does work.

What does it do.?

It's effective as a joke.

Many years ago, i made a bit of Visual Basic to transfer some data between MS Access and Excel. It all worked OK except that *some* of the dates were transferring incorrectly. It took me a while to notice that the faulty dates had day numbers 1-12, and that the months and days were reversed when this happened.

Both Access and Excel were of course set to UK dates and times, but something in the transfer function (the built-in bit) was assuming US dates. The thing is, it was programmed to notice when this would produce an impossible month, and it then reversed the day and the month, thus making it harder to spot the problem than if it had left them alone. Fixing it was a matter of transferring in a different format and rebuilding it on arrival.

The PDP-11 was a pure octal machine, even the memory management registers were shifted by 6 bit (two octal digits).

It was easy to write small programs directly from the front panel switches. The only nasty thing was to calculate the relative branch offsets :-)

In 1975 introduced VAX was a pure hex machine.

In the early days, the Intel documentation for the 8080 was octal, which was very natural due to the instruction formats (8 regs etc.).

For some reason, Intel changed later on to hex for 8080/8085 which made manual programming much harder. Apparently they understood that the hardware architecture was so weird that they started to support PL/M so that users did not have to program that awkward architecture in assembly. Now the same legacy is in 8086/386 and partially even in the 64 bit processor models.

I still remember a bunch of PDP-11 opcodes. The instruction set was wonderful in octal.

Like, nobody needs a computer at home.

x86 was a toad at birth and has only got worse.

Nah, like systemd, X windows or a Porsche, it was a really bad idea at the get-go but has successfully got to the stage where it is tolerable

It was only supposed to be a stopgap until Intel got the iPAX 432 sorted. I wonder how much money they poured down that rat hole before abandoning it?

That's certainly a matter of opinion, not necessarily fact.

It's been successful beyond measure, regardless of the original architectural warts (most of which haven't been even used for 40 years). The next generation will boot directly into 64-bit and all the legacy nonsense will be gone.

HP also tried a super-CISC computer. It failed too; it was pig slow.

And which was that?

The 1000, 2000 and 3000 series were quite successful in the day.

The snakes project started around 1982 (when my college roommate started at HP Fort Collins) and because the rather successful HP-PA.

The "Classic" CISC 3000.

There was a fad for hyper-CISC machines with opcodes like "eval", which would execute an arbitrarily complex stack-based floating point expression. The reaction was the opposite extreme, RISC. Let the compiler do the thinking, not the microcode.

I programmed that machine in '78 and '79. It was a nice robust system handling both batch and interactive users. I don't recall it being "pig slow" (although at the time, my only comparison was the PDP-8). And I was accessing it at 300 baud.

The architecture was somewhat unique, being a stack-based segmented machine (no general purpose registers) - not surprising given the heritage of the machines designers (ex burroughs). SPL/3000 was fun to program in - rather close to the machine.