One can only wonder... which pocket that money really has flown into before the "fix" and how much of it actually found its way back?
You know, skimmer by accident, the best possible accident accounting for future Friday beers ;-)
Welcome once again, gentle readerfolk, to the comforting haven that is Who, Me? – in which Reg readers share tales that show we're all just human underneath. This week we are once again joined by semi-regular raconteur "Bernard", who told us a while ago about a prank that he had played on the auditors at his place of work. The …
When the developers implemented a payment processor into the EPOS - what they didn't do is implement the refund correctly and basically whenever the system did a refund it sent a payment request instead - hence customer got billed twice, but as this was well before Online Banking it left a lot of unhappy customers come statement day. (Glad to say this isn't a who me as I was told when fixing the unreliable mess they left)
Back in the early 1992 the UK Pharmaceutical Wholesalers each had a separate order entry terminal that was supplied to Pharmacies wanting to buy from them. As was the fragmented supply chain at the time, I personally saw a Pharmacy with 5 order entry terminals to order the various drugs from the different wholesalers. (4 of the 5 were Epson PX-4 laptops and so they were stacked in a pile next to the PC used for the other wholesaler, in a very small office.)
I decided with one of the wholesalers to try and tidy this up a little and also deal with a mainframe that needed Viagra it was having so much problem with up time!
Cue a single program on the PX-4 to talk to all wholesalers and a PC front end to the mainframe to accept all the wholesaler's protocols. (Yes, it was hacking the protocols used over the modems, including different number of bits and parities (or none). Nobody minded a White Hat then.)
The solution had 48 serial ports on the 486SX to talk to the modems and buffer these (with stale stock status replies to the Pharmacies) when the mainframe took its regular excursions from working. The commissioning wholesaler also wanted the receiver software to run as a TSR (he'd seen the description of a TSR in a PC magazine and would not be persuaded), as he also wanted to use the PC as (effectively) a print server, (but with document transfer by floppy disk.)
Ahhh... the simpler times.
TSR - terminate & stay resident under msdos? Like Sidekick?
Hardly the good old days :)
I assume the serial port and timer interrupts were being hooked with mutexs to protect the non-reentrant bios and msdos.
Obligatory reference to the Python's "Four Yorkshiremen Script."
"...you try and tell the young people today that... and they won't believe you'.
Sidekick was brilliant in it's day. As a programmer, it was brilliant to use as a simple text editor and edit source code as you found things you wanted to change. OK, so you'd still have to quit, possible recompile and then launch the application, but in a single tasking environment, just being able to edit on the fly was a major breakthrough. Not like now, when you can edit the source code, stop your debugger, then when you click to restart the debugger, the IDE will re-compile anything needed.
"I assume the serial port and timer interrupts were being hooked with mutexs to protect the non-reentrant bios and msdos."
From what I remember of my DOS programming days, to hook you simply had to replace the handler in memory with your own, the call the saved handler after you did the filtering / infecting / bombing..
Like SEH in Windows
Yes, we knew about all those things.
In the 5 years before I moved companies, I never had one support call from the customer, and the only down time was due to BT or the power company disruption.
Quite a lot of the new fangled stuff has actually been around for a long time. In my subsequent job to the one above, I used <<predictive algorithms>> based on historic <<big data>> and current data, for all sorts of things; from predicting the wearing of jet aircraft engines, to underground carpark light replacement a day or two before they would actually fail. The use of feedback loops in those solutions would now be called <<machine learning>> or worse <<AI>>!
Just so last century!
====> collecting the flameproof Jacket
I recall being tasked, 35 years ago, with looking at the potential of an "expert system" to support some procurement decisions (in a large multi-national whose local office spent upwards of £100m pa on non-project purchases). My conclusion was that it would be a nice project (especially if we got the budget actually needed for the tech), but the money could be better spent on a couple of real experts. Yes, whilst the budget was big, a lot was spent on consumables and the system was to ensure off-the-shelf was used when it could, and bespoke when needed: two good engineers (with the right experience in the sector) would be far better value - not to say more versatile, responsive to changes and better able to recognise the inevitable odd-balls.
AI has come on a lot since then, but it still needs the human expertise to apply; if not, natural stupidity will trump artificial intelligence almost every time.
I think this mentality exists in numerous places. We have documented procedures that must be followed.
Curiously we do not have a procedure for dealing with cases when the actual procedures have been found to be bullshit or, in one case, dangerous.
It seems as if procedures cannot be rescinded (is this some sort of ISO accreditation nonsense?) but a procedure can be replaced by an improved one. It just takes endless meetings and studies and piles of paperwork in order to make that happen. Timespan? Evolution happens faster than the yoghurt weavers make decisions because that means somebody has to sign off and that means responsibility... though curiously (there's that word again) nobody is ever held to account for signing off on the crap we wish we could get rid of.
Icon because...my face hurts from all the palm impacts.
You see, banking data was all couriered about by people on motorcycles with sacks full of magnetic tapes. Much could go wrong.
Bank cheques were ferried around the streets of the City (London) by little vans/trollies in the middle of the night - once a bank had put the day's cheques through a cheque sorter that separated them by institution, they would be loaded onto the appropriate trollies and they would trundle off through the streets to be delivered to the correct bank (who would then sort all the cheques that were for them down to the branch level etc.).