This thread is to post the most ridiculous, entertaining, or uninformative and all but or even totally useless errors published by software. The more ridiculous the better. I’ll have a go starting with two.
A mainframe OS called ACOS4 (circa 1980s-90s) required files to be pre-defined for capacity ala IBM ‘technology’. After running for sometimes days and if you hit the capacity limit (and had 20-30 files open) the only message was ‘File error’ and your application just blew off without a trace, no recovery possible. Possibly a contender for a lifetime achievement golden gong.
To modern times, this from Tomtom. There was no way to try logging in again without closing and re-opening the MyDrive application; thoughtful of the Tomtom development team, not. However closing, opening and immediately retrying resulted in a successful login.
“Guru meditation number 00000003.00C06560” or any one of a hundred other hexadecimal numbers. These were OS exceptions from AmigaDOS. The authors were not only nerds but revelled in it, too many in-jokes were barely enough. At least you could get a guide that told you the plain language meaning with a list of likely causes.
I have a standard QWERTY keyboard, with US English layout. It has keys for insert, delete, home, end, caps lock and all sorts of other functions - but no “any” key!
I received a call from an old boss quite some time after leaving a company saying "you know that code you wrote for xyz - well it’s telling us ‘the code can never reach this point; bailing …’ " … it took me some time to compose myself, took even longer to work out how it got there, but as always, it ‘made sense’ in hindsight …
Unix/Linux is a bountiful source of funny errors - Aieee and Oops to name just two …
Or “Something really bad has happened”. Or don’t bother to handle or interpret for the user, just let the database engine pass through failures from SQL or other esoteric cant.
An NCR Tower mini computer that I used for a period in the mid to late 1980s ran an early version of Unix. It would often provide this “helpful” error message: “Not a typewriter “.
I’ve built countless pcs from scratch. Should I ever forget to plug in a keyboard on boot up the error ‘Keyboard error, press F1 to continue’ is displayed. WT? How does that help or even work…?
The BIOS errors you once got could be both multitudinous and meaningless. Fortunately things have improved a little with BIOS and now UEFI, but you still want to have the manual to hand.
Probably had a better chance of sorting it then - I imagine there are more details under the covers that will lead investigators to the guilty subroutine. I notice there is a full dot release b1 out …
Yes. It seems that error messages are now being written for users rather than programmers/maintainers. Less confusion for the user, but more work to figure out what went wrong.
Windows used to give decently detailed error messages - now it dumps a log file… somewhere.
I lost count of how many releases and years it took them to put all (only some? most?) logs in .\windows\logs. But then reading them to try to get information a mere mortal can understand has often been more challenging than even finding such logs in yesteryear.
it amazes me how much unhelpful and obvious stuff they log - I did battle with an old windows mobile driven barcode scanner a couple weeks ago, it claimed detailed information was logged which turned out to mean a rudimentary stack dump and the message “connection lost” - to where? how? timeout? reset? act of ‘god’? arrgggghh … I blame the ‘rona’ - hope the three Cathay Dragons that landed just now for storage weren’t carrying a fresh supply
Rehashing what I started this older topic with: My all time favourite error was from the NEC ACOS4 operating system, a fork of Honeywell GCOS. Like IBM systems of the time the JCL had to specify and size each file; the mentality was optimising a highly predictable transaction (banking) system; ACOS4 was considered significantly better than GCOS itself and ACOS4 crashes were unheard of (1980s-90s).
Unfortunately they tried to put it into a scientific environment whereby after days of execution a program would error off with the sole message ‘File error’ - not even its unit or name, just ‘File error’. Programs typically had scores of files, all left in situ at the moment of crash for manual cleanup. OPEN OLD FILE, MAKE NEW FILE, and DELETE OLD FILE IF IT EXISTS AND MAKE NEW FILE were all different, to be learned the hard way. If one tried to DELETE OLD FILE and it did not exist it would error
The JCL manuals were literally two stacks each about 1.5 metres high, most commands with only the absolute most simple examples regardless of how complex the command could be.
ACOS4 was subsequently replaced by a UNIX variant for the scientific world.
