Have you let a user wreak havoc on your machine?
Heck no.
I have wreaked havoc on a user's machine, but only after severe provocation.
WHO, ME? Mistakes happen, and when readers of The Register make them, we like to share those tales of woe in a Monday column we call "Who, Me?" It's our guide on how not to get ahead in the modern workplace. This week, meet a reader we'll Regomize as Clay, who in the 1980s was the most junior member of the management team on a …
Many years ago I was at a conference where someone was demonstrating some newfangled speech recognition board (this was way before a standard PC had the power to do it in software). After demonstrating commands like "TYPE file.txt" etc. there was a short pause, during which a loud voice from the back of the room called "FORMAT C colon". Fortunately the technology wasn't quite up to modern standards, the only response was much laughter (from the presenter too, after an initial look of panic!).
I may have mentioned this before, but in the late 80s I shared a PC with a young engineer who had a bad case of Dunning Kreuger syndrome when he had his hands on the keyboard.
At one point I physically removed the keyboard from his possession just before he hit ENTER having typed FORMAT C:
Whilst I had recent backups, it was the days of 360K floppies and I didn't fancy reloading the entire 15MB of software and data from floppies.
I fully admit to threatening him with physical violence if he tried anything so stupid in the future, with the advice to RTFM before hitting ENTER.
That was OS/2 when they introduced the "killer feature" (IBM's term, not mine) of VTD (voice type detection) that would finally allow OS/2 to overtake Windows.
There were only three problems with that.
First, the VTD in OS/2 sucked, being discrete. That mean you had to speak. every. single. word. as. it's. own. sentence., with a discrete (hence the term) pause after each world. William Shatner not withstanding, that's difficult for most people.
Second, third party VTD packages for Windows appeared, and they weren't discrete. Sure, they cost $50-$80, but having your "killer feature" be overtaken by your competitor is not a good look.
And finally, after the initial curiosity, no one really used it. OS/2 was mostly using in corporate environments, where there's already enough background noise. Now imagine having your 50 software developers all talking to their PCs, and the resultant chaos from one developer's overhearing his cube neighbour's verbal commands.
The "FORMAT C colon yes yes" joke is, sadly, an urban legend. It was routinely shouted out by audience members at IBM demos of OS/2, but it didn't do anything.
Speaking of early speech recognition, as it were, is anyone here old enough to have attended the Seattle World's Fair (which brought us that now-ubiqutious modern wonder of transportation, the monorail) in 1962?
I remember being fascinated by seeing and (I think) getting to use the IBM Shoebox, a gizmo, according to Wikipedia, "able to perform mathematical functions and provide speech recognition. It was capable of recognizing 16 spoken words, including the digits from 0 through 9. "
For small values of capable, obviously.
That and seeing an IBM 709 computer (I think -- might've been a 7094), which produced reading lists based upon input from a questionaire, probably put me on the trajectory to doing whatever the heck I've been doing for the past few decades.
> that now-ubiqutious modern wonder of transportation, the monorail) in 1962?
Ubiqutious? A couple amusement parks and a few airports? I am fairly sure there is no monorail in my state or the next few states. I have seen Newark Airport and Disney Florida but they didn't take over.
Also: The 'Hotchkiss Bicycle Railroad' was a purpose-built monorail on which a matching bicycle could be ridden. Four examples in the 1890s.There is another vintage monorail with steam engine still run as a tourist fun-ride, name forgotten.
In 1989, I was working as a temp doing various office stuff. I'd barely had any experience with computers, except a couple of in-house systems, and had never used floppies. One employer gave me a floppy and asked me to format it. I didn't know how, so she impatiently explained what prompt to go to and type format. I got a prompt asking if I wanted to format [something I didn't understand], and since I'd been told to format, I clicked yes. It ground away for a really long time, and eventually the person who'd given it to me showed up to see what was taking so long. Turned out that the part of the command I hadn't understood was C: (I later learned about A:, but that was later!). Yes, I'd tried to reformat the whole drive, causing some panic when they realized what had happened. No panic for me, I was still blissfully ignorant. I learned later that the office had tried to get my temp agency to pay for it, but never heard or cared to find out beyond that.
FORMAT without arguments defaulted to C
That must be what happened! I recounted my adventures trying to format a floppy in 1989 above, and that must have been what got me. I don't know if the person giving me the assignment didn't say "type in format a:" and I only remembered the "format" part or what, but it was C: that got (partially) formatted. This makes me feel slightly less guilty, although I didn't feel very guilty in the first place.
I've recounted before, the ACT Apricot always used A: as the primary disk, so if an HD was installed that would be A: and the first floppy would be B:
Swapping between IBM compatibles and Apricot's was hazardous, particularly as I was often retrieving data from Apricot machines:
Format A:
Oops!
Ctrl-C
Get out Norton Utilities diskette
Undelete
Fortunately the few times it happened I knew what was needed for Undelete to save the day, i.e. the first letter of the filenames.
> what was the objective of the DEL *.* command?
TFA> used the only command he knew
Objective: showing off literally the only thing he knew.
So, from your list, "ineptitude" is the best fit. Although, the desire of a manager to meddle does appear to be part of the pact they sign with an "evil 3rd party"*
* snappy dresser, often in red to match the colour of the "ink" used on the contract.
I had a user with a DOS machine that wouldn't boot. They admitted to trying to delete a subdirectory. After a few questions about how they'd gone about this, I realised they'd done the following
cd \dir
del *.*
dir
dir showed them what they thought were two files, '.' and '..' so they typed
del .
del ..
Norton Utilities undelete fixed the problem. I still smile whenever I think about it.
del *.*
"In next week's thrilling installment of Who Me?, John reboots the wrong server and then puts an asset tag on the left side rather than the right side."
(And, yes, I have contributed to On Call and other articles in the past, back when El Reg was somewhat half decent.)
> C:\>dir a:
> C:\>del *.*
Just shows the un-unix evil of having disjoint file systems ;)
Each file system being being assigned a device letter and working directory.
C:\ A:
A:\ cd xyz
A:\XYZ> C:
C:\> DEL A:*.* REM will delete files in the working directory on drive A viz A:\XYZ
Back in the days of msdos 3.3, I just joined the drives under C:\ - not so important for the floppies but for another 3 or 4 hard disk partitions the only sane way. Most (~all) of the code I was mucking about with, was for Unix systems which didn't grok drive letters. An undocumented dos call allowed setting the switch char to '—' and path separator to '/' which largely completed the illusion… until I could get a real *ix system. :)
Strikes me that the only command an apprentice manglement apparatchik knew was del *.* speaks volumes of the destructive mindset of the whole accursed boiling.
The original example doesn't demonstrate any problem with disjoint file systems, where I can easily do a ls /mnt/sda and then rm * without recognizing that those weren't going to operate on the same files.
Your new version doesn't demonstrate Unix's superiority either, since if you wanted to delete all the files in the A: drive, you would obviously type del A:\*.* which would not care about the working directory. In Unix, leaving out the path separator would be no better rm -r /mnt/sda* could do even more damage and rm /mnt/sda* might do nothing or might do something but almost certainly not what you were going for. Neither operating system tries to or does prevent you from running completely valid commands that are not what you want to run.
Not correctly deduce, as the delete example works as long as I own the directory I'm in and, as long as I also have the ability to delete things in the mounted disk, I would have every reason to expect it to work when making the mistake. But you may deduce as you please, even if you get it wrong.
DOS 3.3 came out in 1987. Mark Williams Company's Coherent was released in 1980, but consumers had to wait until 1983 to purchase it directly. It was a pretty good learning tool, and could be used for light UNIX development work ... and was over-all much better than the MS/PC-DOS of the day.
For the adventurous, at around that time, shell accounts on BSD systems were usually available for the asking at your local Uni, even if you weren't currently going to school there.
For folks with money burning a hole in their pocket (or a Boss who knew a good use of a budget!), the so- called "AT&T UNIX PC" was a real UNIX machine that one could use at home in the early '80s. I lucked into a later model in the mid 80s, the 3B/1. Came with all of 2Megs of RAM and a 67Meg HDD and was very usable ... ran SysVR2 with a lot of 4.xBSD improvements.
Reminds me of the story I read ages ago about the user that was told that you must format new disks before using them (back in an era before floppy disks arrived pre-formatted).
A box arrived some weeks later containing some software on a set of floppy disks. The user wanted to use them, but the disks were new and they remembered their training...
A long LONG time ago, when I was in university, I had a girlfriend that worked for the Home Economics college admins or knew them well. Being an engineering student I knew my way around the computers of the day. Her department head wanted to pay me for computer work. I think I earned $12/hour formatting a few boxes of floppy disks.
Might have been one of the easiest jobs of my working life. Unfortunately or fortunately no "who, me" came out of it.
Many moons ago, I had a class at the Uni where (numerical analysis?) "homework" was handed out on floppy disks. These contained a small program which presented tasks and then gave us 3 shots at solving these. The results were graded by amount of tasks solved with 1 to 3 tries, and tasks not solved.
Being wary of problems with strange floppy disks, the first thing I did was to run chkdisk.
This found and fixed disk errors, so I congratulated myself for having dodged that bullet., until I tried to run the program, which promptly failed.
I found out later that, in order to prevent disk multiplication, the disk had been equipped with disk errors which the program needed to see before it was run.
Having optained fresh working disks, a group of us sat down and started solving the identical tasks as a group, making sure that not a single stundent needed more than 2 tries for any task...
So you cannot FROMAT a floppy. .:)
DOS always muddied the water by conflating the physical formatting of the media and disk partitioning with the construction of a file system.
I recall tormatting floppies even pre-formatted ones before use was advisable as it would pick up defective diskettes and "adjust" for any differences between the manufacturers formatting process and your disk drive's "peculiarities."
Floppy disks (and drives) would have taken a close second place to printers for a the privilege of being the greatest bane of the BoFH's life. Even outside the dismal MS ecosystem floppies had quite diverse uses from boot disks for various computers, network hardware boot and configs, lab instruments etc.
It was confusing reading the article and seeing this:
The junior had little choice but to comply, so Clay sat down at the machine and used the only command he knew.MORE CONTEXT
At first I thought they'd found some undocumented feature in MORE that managed to wreck machines. El Reg's article interruptor can turn up in strange places.
That would be Bill Atkinson, at Apple.
Early nineties and I was the main autocad operator for a company straight out of school.
I ran a 486sx25 with maths coprocessor and a 1024x768 display. This cost about £5000 and had a 100Mb hard drive.
I had a few days off and came back to find I couldn’t do any work as someone had wiped the drive containing all the data files and they were in the middle of restoring the data. I think was either an ID10t or VDU fault.
I used this back in the day to quickly find and remove the temporary files Windows 3 and WordPerfect would leave scattered around the computers at the small translation firm I worked for.
At the time of this story (mid-1990s?), all work product was handed around the office on floppies, which typically meant the final work product usually ended up on the boss' PC (a blazing-fast 386DX40 IIRC).
One Friday afternoon while doing a bit of hard drive cleanup, I somehow instructed it to delete all files on his hard disk (no idea what the keystrokes were, but...)
Queue the weekend spent playing with undelete and renaming thousands of customer files, one at a time.
A few months later, I installed a coax network and a file server (well, headless PC) as well as starting a regimen of taking tape backups of all work files and I've been suitably paranoid about backups ever since ;)
See Midnight Commander, a clone of the old MS-DOS utility Norton Commander, that has been tweaked for the *nix environment. It's a useful tool, and a lot more powerful than it looks at first glance. I use it near daily, and have done for around 30 years. 40 years if you include the DOS version from Norton. Try typing mc at a command prompt. Recommended.
Free tip: Be extremely careful if you choose to run mc as root ... it will do exactly what you tell it to do. Don't say I didn't warn you.
Around forty years ago, a customer admin was working his way through the AOS/VS Commands and Utilities guide when he reached FORMAT. The good news is that he then got to test his backups, and the better news is that they must have been OK, since there wasn't extended downtime drama.
(Unlike the customer who called me up to say that she had put up a tape, and entered PCOPY at the console. The dialogue had reached "Is this a tape-to-disk PCOPY?". She had entered "Yes", and only then thought to call us. I think that it was the best part of a week before they were running again.)
50 years ago I joined IBM and one of my first tasks was to stress test CICS (like an operating system for doing business transactions).
The stress test had every one in the test groups running tests. They were told to logon and issue any command on the "command reference". If CICS stayed up for 3 hours the test was successful.
About 2 hours in, I tried the CSMT,P,SHUT,IMMED command and suddenly there was panic. CICS had shutdown!
The person running the test asked "did anyone issue the CSMT,P,SHUT,IMMED" command. I put my hand up.
They restarted the test and pretended the shutdown didn't happen.
Next year they had automation which could simulate end users, and they didn't suffer from a dumb end user again.
Dumb users are to be coddled, at least occasionally ... We had a gal who could break any software in less than five minutes at the keys. She was an 'orrible computer user because she broke them accidentally, not a malicious or mean bone in her body.
HOWEVER, I discovered she had the raw talent to be able to recreate the steps she took to crash the thing. We pulled her into software QA pronto ...
I used to work for a high street retailer back on the early 90's.
Our EPOS system was prehistoric even for the 90's, with no mouse, just a menu you worked through with the cursor keys and short text commands.
One day the store manager decided to help out in the stockroom. His first task, to move a product from one shelf to another, this was done using the relocate command, and required three parameters; product code, current bin location, new bin location.
He forgot the first two parameters, so what he did was move all 20,000 product codes to the new bin location, so the next pick list that came through had all 10 items on the list showing in the same bin location.
It took many late nights to move all products back to their correct locations, one item at a time.
He was never allowed to step foot in the stockroom ever again...
"Our EPOS system was prehistoric even for the 90's, with no mouse, just a menu you worked through with the cursor keys and short text commands."
I think every EPOS system in the 90s (especially the early ones you reference) were text based. I didn't see a single mouse through the newsagent and off-licence I worked in up until 2000.
Several years ago I was at a local tractor parts store. The owner was having trouble with his ancient POS system. After a couple minutes of swearing at technical help, and at least one "turn it off and then on again", he hung up the phone and dug under the counter and came up with an old pad of blank invoices. He allowed as to how he would have to hand write it because the computer was broken. Being me, I asked if I could have a look at it. He looked at me funny, because he's only known me as a rancher/farmer, but said OK.
I came around the counter & eyeballed the situation. Turned out the mouse didn't work. At all. Just an immobile mouse pointer in the center of the screen. I showed the guy the tab key and a couple other keyboard shortcuts for future reference, and took a closer look. The drivers were loaded, and running. Probably a dead mouse (tractor parts counters are rough on everything). I just happened to have a spare serial mouse out in the truck (don't you?). Pulled the mini-maglite out of my pocket & dove under the counter to swap 'em out ... only to discover that the original mouse had managed to work itself loose. So I plugged it back in, and all was well with the world.
Well over ten years of daily use, and the computer had to wait until I was in the shop to spit out it's mouse. Now one of the biggest gossips in the county knows I'm a computer guy. Hopefully he'll keep it to himself ... (In the eight or so years since the above happened, he has called me three times for computer work ...and never told anybody I'm a computer guy. Starting a couple years ago, I've been giving him a good bottle of whiskey (or whisky) at Xmas as a thank you for keeping quiet.)
As of DOS 2.0, "del *.*" added an "Are you sure (Y/N)?" confirmation prompt specifically because of user errors like this.
I remember this because a co-worker accidentally did delete his day's work once by mistake, because he mistakenly did confirm it, and he promptly went on a tirade about how it really should have asked him twice.
He wasn't big on personal responsibility, as you can imagine.
I do remember this happening on a Control Data 6400, however. Their were two operating systems for it at the time, SCOPE and NOS, with NOS being newer and shinier. Both had a purge command to delete files, and a purgall command to purge all. The purgall command was intended for scripts, so it didn't prompt for confirmation.The designers explicitly named it purgall rather than purgeall so that typing errors due to muscle memory wouldn't destroy the user's work by mistake.
The administrator was logged into a user's account to help him with something, and joked that he'd made a mistake, so he'd restart. He type "purgal, type=a" (type of files=all), as the horrified account owner thought all his files would be deleted. The admin was joking, because he left out the final letter of the purgall command, which meant the command would be rejected.
As it would have. On SCOPE. However, the system had been upgraded to NOS the previous week, and NOS's upgraded shell had a neat feature of auto completing incomplete user commands if they were unique enough to be recognized. Which purgal was.
So instead of seeing the expected "Format error on control card" error message that he expected to see, the admin was horrified when his joking command echoed back "Purge complete".
Needless to say, the admin had to explain to the user's boss that the user's "the dog ate my homework" excuse was actually true, with the admin being the dog in question. Which was appropriate, because he was in the doghouse for quite a while after that.
Another tale from the days of CLI:
The CLI for RDOS (a 16-bit, two-ground OS) had a MOVE command with the destination directory as the last argument. The CLI for AOS/VS, the successor to RDOS, had a MOVE command with the destination directory as the first argument. A co-worker, who hardly every worked with RDOS, forgot the distinction while trying to help one of our customers. He told our boss, "Don't ever let me work with RDOS again.") In fact, we fairly shortly upgraded that customer.
I never fell into that particular trap, which owed something to caution and plenty to luck.