Many years ago, when I was working at the BBC and writing end-user applications in a 4GL called SeaChange (which had very crappy debugging tools) I was reduced to kneeling on the floor and listening to the hard drives to check what was going on. I seem to remember reading about Ken Thompson doing something similar (not comparing myself here).
And another physical hard drive story - the first one I ever got bought, for a DEC Rainbow (PC class computer), came in a crate the size of a small fridge (like the ones you get in hotel rooms) and took about an hour to unpack.
Ooh and mentioning DEC and disks I remember when I first used a DECSystem10 repeatedly submitting "mount prog" requests (to actually physically mount the "prog" disk pack so that I could use LISP) until the very cross operators (remember them?) rang me and said "we are doing scheduled backups, you idiot". I was always nice to operators after that.
Ha, yeah I remember doing many repeated mount/unmount requests on the school’s DECSystem20 before I realized it was a request for the human operator to find a magtape and thread it through the reel-to-reel.
Mercifully, I had nothing to do with tape drives then, much horror later. And shortly after we converted from the DEC10 to a couple of IBM 4381s, who's backup regimes I simply did not want to know about - fingers in ears and screaming. But I did like them and wrote a distributed printing system for them in Rexx.
I was assisting a team from Digital (DEC) recommissioning a VAX 11/750 system in our computer room. The DEC guys forgot to lower the feet on the disk unit and when the disk pack confidence test started, the whole unit lurched forward and two of us had to hold it down so it didn't damage any of its ribbon cables.
Later, when the DEC guys were pushing the unit back in to place, one of the ribbon cables got caught between two metal edges and was cut in half. Fortunately, we had a spare.
Unless coupled with extensive verification, these stories should be taken as urban legends: told, re-told, and expounded upon by people who like the attention.
Of course, if you simply like to read unsubstantiated legends, there are plenty available:
I have no extensive documentation. Heck I've never mentioned this to anyone that was not in the room at the time, but in 1996 - I did this. With a smaller disk - an 8" disk mounted high in old tower PC (newer than the disk), and some x86 assembler.
This feels like physics to me - I'm not sure why people need extensive documentation to believe that throwing a drive head back and forth would cause a machine to vibrate (in fact even at just 8" those large drive heads make a machine jump and stutter) and if one was lucky do so consistently enough to walk the machine forward.
Considering that only recently the first-wave of RTX 3xxx cards had a flaw where uncapped FPS on videogame menu screens caused them to spike current outside of the intended design envelope and completely brick the devices with unfixable damage, I think it's pretty safe to assume there's a long history of "user doing something programmer didn't expect ends very badly".
> they told me that one time they let it go so far that it pulled its own cable out of the plug
Not sure I believe this story as IBM channel cables have a big ass screw in the center that is used to make sure they don't come loose. I can't remember if it's even possible to fully seat the connector without tightening the screw but certainly it was not a practice to leave it undone.
There is another story there about someone writing assembler program calculating 50000! for a DEC PDP-10. The PDP-10's registers are memory mapped to locations 0..15, so you could write a program that ran entirely from the registers. According to the story the registers (which were TTL) overheated and died.
I'm old enough to have programmed a PDP-10 when I was at Uni, and yeah once you grok'ed the architecture writing a program that ran in the registers is a challenge you set yourself. Mine computed primes using a sieve. It didn't kill anything.
A person of my acquaintance owned a PDP-10 which he had in his apartment. Originally he had no memory for it and could only run programs from the registers and I remember thinking how strange that was.
I heard of a story that a Drum-drive started walking until it touted a table (over night).
There is also another story i think from DEC, that there was a Company who had like 10x more head-crashes (always on the same day), bit later they found out is was the Lorry who delivered goods one floor below.
Many years ago, when I was working at the BBC and writing end-user applications in a 4GL called SeaChange (which had very crappy debugging tools) I was reduced to kneeling on the floor and listening to the hard drives to check what was going on. I seem to remember reading about Ken Thompson doing something similar (not comparing myself here).
And another physical hard drive story - the first one I ever got bought, for a DEC Rainbow (PC class computer), came in a crate the size of a small fridge (like the ones you get in hotel rooms) and took about an hour to unpack.
Ooh and mentioning DEC and disks I remember when I first used a DECSystem10 repeatedly submitting "mount prog" requests (to actually physically mount the "prog" disk pack so that I could use LISP) until the very cross operators (remember them?) rang me and said "we are doing scheduled backups, you idiot". I was always nice to operators after that.