
A business-critical SPF - replicate it - who would have thunk?
Welcome to Who, Me?, The Register's weekly dive into the murky waters of the reader confession pool. Read on to see if this week's tale triggers a guilty memory or two from you. Today's story from "Art" takes us back three decades to the closing of the 1980s. Knight Rider was a receding memory while Baywatch had only just …
This post has been deleted by its author
and in the 90s the engineering firm I was an apprentice at didn't realise the one A0 plotter we had was business critical despite everyone using it day in, day out. Until of course it packed in during a large print run.
The solution? Should at the lowly engineering apprentice who's somehow ended up doing all of the IT for the company..
Some things never change!
That's why the principle IT Team (Back Office etc) has just been replaced\supplanted by the field techs at my place of work (Who also seem amazed that they should have to do things like triaging tickets\emails & telling managers what new\returning employees AD credentials are & updating the tickets themselves is an optional task) .
Much hilarity for the outgoing team as things have already started to go wrong (In a Mischief Theatre style) three fifths of the knowledge has now gone from site.
Icon - Me getting me coat as I'll be gone by Feb end or sooner.
Icon - Me getting me coat as I'll be gone by Feb end or sooner.
I hope your new job brings you great new opportunities and many new stories, most of which of amazingly brilliant managers and oddly competent CxOs etc. May all your tales of Incredibly Incompetent Imbeciles be only of people you met before today's date :)
A customer of mine was a pharmaceutical wholesaler with a McDonnal Douglas mainframe controlling everything it could (at least everything early 90's style). Each of the wholesalers in the UK had a proprietary protocol to communicate with the chemist shops when they placed orders. I went to one chemists where they had 3 identical Epson PX-4 computers, one for ordering with each wholesaler they used. I quickly created one piece of code that would interface to all the wholesalers and reduce the number of computers in the (always) crowded chemists to 1.
Now, the wholesaler mentioned above, wanted a receiver to take orders from any of the remote order entry computers, and to increase the number of incoming order modem lines, and buffer orders when the mainframe was having its regular TITSUP. Just for fun some protocols where 8 bit no parity, some 7 bit even parity and one was 7 bit odd parity. The longest expansion bus I could find with sufficient power supply allowed 32 incoming modem interfaces. I got a 25MHz 486Sx and a 'large for the time' hard disk.
Within a week of first trials it was buffering all the orders and faking order reception when the mainframe was AWOL - all very mission critical. And all backed up, and the recovery procedure documented and tested. That should have been the end of a very successful project until the customer pointed out that I hadn't done the tasks on the back of the SOW - I had only been given a single sided photocopy!
Page 2 said that the computer also had to be used as a word processing terminal - "as it was only a backup for the main frame"! And the customer would not pay until that was demonstrated as well. They might have been brilliant at stock control, but they just could not understand that all orders were now passing through the PC and that it had to be dedicated. I, my employer, and the local university lecturer (who played golf with the wholesaler's MD) all tried to persuade the MD but to no avail.
So the order entry system was made into a Terminate-And-Stay-Resident program and the demonstration was made with creating and printing a Wordstar document at peak ordering time. At which point the glass door in front of the terminal was locked and not opened in the 5 years before I moved to another company. Just before I left, the wholesaler MD reported that the PC had never missed a beat and credited that with the software being in a TSR. He had apparently read in a magazine on the shelves of WH Smith, at the time he was writing the SOW, that TSR's had to be the most reliable software and that was why he insisted on it. (I can't work out whether he was a genius or a dick?)
My first PC was a Gateway. Over £2k of my recent redundancy cash (first of many due to the wonders of the .com 'business' model) blown in their Covent Garden shop on a shiny Pentium II and two - two! - Voodoo 2 GPUs, each with a whacking great 8MB of onboard RAM. Interstate 76 never looked so good.
... one of the most pirated small bits of code out there in the early 1980s ... for those who didn't have his earlier version that was shared through user groups, that is.
And who among us didn't have a series of bootabe floppies with the necessary system files for all the variations of MS/PC-DOS that were in the wild at the time? I still have that box of disks (and a couple releases of 4DOS). Comes in handy six or eight times per year ... you'd be surprised at how many pieces of decades old computer controlled equipment still runs on MS/PC-DOS. Bridgeport CNC machines of the era, for one example that is going to be around for decades to come.
I saved the original code I had been working on as .cobol and carried on making changes to the original .cob files, the direction of the code having been changed by management.
After 2 - 3 days of diligent work and testing, the new code was finished. I then went to delete the backup of the original code, with del *.cob;*
Hit return and then screamed!
The admins didn't have a backup, so I had to re-do all the work again. Luckily, most if it was still in my head, so I could re-type most of it straight away. The final result was actually little more elegant than the original and it was done in a couple of hours, instead of a couple of days.
I booked that code back into the core system, before I deleted the backups this time around!
For the uninitiated, that will wipe the current directory (the . ) and all descendants ( -r ). Depending on implementation, that may or may not include the parent directory ( .. ) recursing that direction too.
The safer way to reference dot-files is ./.*
Also, I always use tab completion to get an idea of which files the shell will interpret my scrawlings as before hitting enter.
In the 90s, I once had a call from someone who couldn't get their PC to boot. A few questions revealed that they'd deleted a sub-directory from the root using something like the following commands:
C:\> cd subdir
c:\subdir> del *.*
Are you sure (Y/N)? y
c:\subdir> dir
Volume in drive C is DOS
Volume Serial Number is D4F6-F60C
Directory of c:\subdir
24/12/199? 09:27 <DIR> .
24/12/199? 09:27 <DIR> ..
0 File(s) 0 bytes
2 Dir(s) 779,776 bytes free
c:\subdir> del .
Are you sure (Y/N)? y
c:\subdir> del ..
Are you sure (Y/N)? y
Still makes me smile.
Back in the days of 3.1/3.11 i needed to erase a floppy (stiffy) for re-use. It was something I did quite often and as windows file manager was a bit labourious I kept a handy "command prompt" shortcut (not sure they were called shortcuts back then) on the desktop.
My typing sequence was probably identical to Arts, except I was still in the c:\windows directory.
For the next 20 minutes or so it was fun to watch windows slowly fall apart until it finally gave up and bluescreened. Actually, I'm not sure Bill had invented bluescreens by then, I think it may have just froze.
Time to get the box of 21 windows install floppies out......
Hey, I still use the old TNI excuse (mostly to save very senior users' blushes. As in "No, Director, it wasn't your fault that you failed to save the document, it must have been a network glitch." I have the survival instincts of a cockroach). Don't give away trade secrets! ;-)
General Protection Fault
That seemed to cover every windows issue up until the NT kernel became more popular with Win2k
GPFs could only be issued by protected-mode Windows, so they didn't happen until Windows 3.0 running in "386 Enhanced" mode. While there were Windows/286 and Windows/386 versions of Windows 2.0, they didn't run in protected mode. Win/286 just allowed access to the HMA, and Win/386 ran Windows in a real-mode address space. (Maybe. Sources differ on this, and I don't remember for certain; the Windows kernel might have run in protected mode in Win/386, at least for the 2.10 and 2.11 releases.)
When real-mode Windows crashed, it just crashed.
create a Win311 folder and xcopy all the floppies into it
In a similar manner (and I'm sure I'm not the first or last to have done this) I got bored feeding floppies (was it 3 for DOS and 7 or 8 for WfW 3.11?), created a boot floppy which included a minimal DOS, NE2000 drivers and Netware client, and found that I could have half a dozen new PCs installing DOS and Windows 3.11 near simultaneously - if slowly - from the discs I'd xcopied to the server.
Simpler times.
M.
But not always easier times.
I still have nightmares of the time (mid '90s) that I was the new supervisor of a new branch office for an industrial service firm. Our regional administrator showed up bearing a slightly used SX386 with a minuscule hard drive and two 3.5" floppy drives. At least Windows was already installed. MS-Office was installed via about 45 floppy disks - nothing productive was accomplished that day!
from the discs I'd xcopied to the server.
Not too long back I realised with the Windows install (XP-7 at least) you could "interrupt" it at the 70th or 71st reboot (whatever # the last one was), clone the disk and then just copy that image to a machine. I used some very driver-laden images for that, though Snappy Driver Installer (and it's predecessors) worked wonders there.
I had a floppy "image" with a basic Dos (snaffled from HDD Regenerator IIRC), a pile of generic network drivers, and an old DOS-based Ghost. Boot the image from PXE (a small few seconds), clone the image across the network onto a new machine, reboot, fill in a few bits of data and deal with some nagging (windows really gets upset when it wakes up on new hardware!) and in sometimes as little as 10 minutes [1] you have a fresh install!
[1] We're talking Windows here. 10 minutes is exceptionally quick for an installation! Hell, 2 days can be quick to complete an installation when you include updates and the massive slowdown with 7's update process shortly after GWX!
As for your "XP-7": You could have completed the installation normally, set up most of your stuff and software, and then use sysprep. Google for it ;).
I may've come across that but the blurb notes on DDG (google doesn't work in my house) tend to refer to comments like "Sysprep is the Microsoft system preparation tool used by system administrators often during the automated deployment of Windows Server based operating systems. Sysprep is most frequently used in virtualized environments to prepare a system image which will be cloned multiple times."
Since I was using this on home computers, not servers or datacenters, that would've flagged it for "ignore". That's assuming I bothered to look. May've come up with this idea as an experiment while installing windows, captured the image at the right point and not even considered seeing if others had dealt with this issue before :) (I was often spending long hours there, working 12pm-4am wasn't unheard of!)
Another says "Microsoft also does not support the use of Sysprep to install an operating system from an image if the image was created by using a computer whose motherboard has a different manufacturer, or if the image was created by using a computer with the same configuration but from a different manufacturer." - if I'd come across that while planning for something where I can use it on any brand of mobo with any hardware...
Thanks for the headsup. If I'm ever doing that sort of work again (unlikely before the end of W10) I'll keep it in mind.
From memory and experience building PC's in a shop in the early nineties, Windows 3.1 was six floppies and Windows 3.11 for Workgroups was seven floppies.
That matches my recollection as well.
IIRC one of the W95 release candidates was >20 and maybe >40 disks. I took it as an omen when I tried to install it one day and found the first disk corrupted... :)
In the early 90s, I got a call from a 'colleague' who was having trouble finding a file he had created from a BASIC (I know, I know) programme (by using the dir command).
DOS (as it was) would not recognise filenames with spaces in them but it was easy to actually create a file with spaces in it from BASIC (a particular weirdness of DOS) to which one could happily write and read data from within such a programme.
I brought over my (invaluable) copy of Norton Utilities and used the file examiner (I do not recall the specific name), found the file and removed the space from the filename and all was well (the file now showed up in a dir listing).
I had hoped that he might have got the picture after I explained that filenames with spaces would not show up in a listing but the same thing happened a couple of weeks later; I could see the file with Norton but I did not change it and told him 'we must have been lucky last time' which did actually penetrate into his what passed for a brain as it didn't happen again (as far as I know).
Is HiRen still producing his utility CD? That rocked too!
If anyone deserves fame and fortune... It's all the people who made the utils that go into those disks!
(And Hiren himself? He can get an extra portion of fame&fortune for bringing it all together!) (poetry by error, sorry)
I had to write something once to work with an EDIFACT feed. What a viciously bad monstrosity of a file structure.
Hundreds of fields, most of them not needed but woe to those that missed or misinterpreted the dozens that mattered, even though they weren't properly documented and were frequently of the form 'whatever data happens to be in here'.
API by committee, designed by bureaucrats, never seen by anybody with technical skills before they were forced to use it. Utter abomination.
I used EDIFACT for a project - had to write some Kermit stuff to automate getting the files. I think EDIFACT taught me how not to do things but it did mean when I wrote an HTML standards checker it involved about 200 lines of code and the structures more or less cut an pasted from the standard.and I could check our web site for accessibility far better than the government recommended service that cost £1k a pop or something outrageous.
I work for a bank. I currently have a request in the queue for a new program to perform EDI encoding on certain loan types that need to be submitted to the US Dept of Agriculture in this format. The encoded files are to be manually uploaded to a website (yes, this post is being written at the end of 2019).
I contacted the USDA about possibly using XML and inquiring about the existence of a web service that we could connect to and push the data, and met with blank-faced silence and eye blinks. They did make some noises about "modernization", but made it clear it wouldn't be happening in my lifetime.
Not kidding, unfortunately.
Maybe I should look into FAXing them some hardcopy?
After doing del *.* without changing to the A: drive once too often (i.e. once), we ended up writing a batch file 'dela.bat' with the required commands in it, and then training people to use that to wipe a floppy/stiffie. As a bonus, it was easier to type than the series of commands previously used, so not too difficult to get people to use the batch file instead.
I did the same while sitting in c:\win31 on my dad's IBM 386 PS/2 P70 luggable he'd brought home for the weekend. Undelete saved me, but you had to supply the first character for every file (in DOS 5 anyway) so it took a lot of slightly educated guesses to get Windows 3.1 up and running again! This would have been 1993 ish and a very happy 13 year old I was when I got back into Program Manager!
A very common thing back then was somebody asking if they could check a floppy, will only take a minute.
The user of the machine would normally OK this.
In went the floppy, followed by dir a:
The contents of said floppy is displayed, the owner says "good, I can get rid of that and re-use the disk" --- next command is del*.* the floppy is removed and the owner walks away thinking they now have a blank floppy, completely unaware of the trail of destruction they have just left behind. I was normally called to sort it out when the user found their machine didn't work anymore.
Of course dir a: displays the contents of the disk in a but you're still logged on to c: and any further command will be carried out on c, not a, resulting in the root of c, or whatever directory it was in when the command was executed, being wiped.
"With a nostalgic twinkle in his eye he said, "These were the days of EDIFACT and TRADACOMS (do they still get used?)""
Oh yes. Most certainly.
In fact the company I work for uses its own personal TRADACOMS format that we translate to EDIFACT via a third party, a very common practice considering how much money they make.
We no longer use an AS400 to handle EDI and ERP, now we use Microsoft NAV, which natively supports EDI. The actual EDI transfers are handled by a couple of high availability Linux servers (running Ubuntu) in Azure.
"With a nostalgic twinkle in his eye he said, "These were the days of EDIFACT and TRADACOMS (do they still get used?)""
Oh yes. Most certainly.
I remember way back in the early 1990's when I'd say that "EDI Standard " was one of the greatest oxymorons out there, right up with "Postal service".
While EDI is a "standard" (the shop using it had even bought the dictionaries) it was also open enough that you could do whatever you wanted, and still be "within standard'. The half dozen partners we had were all doing the same business, but none of those streams were formatted the same.
So assume you have an in between bit of software to reformat,
That's a bit of an assumption for your typical LAN-based accounting software of the early-mid 1990's. Nope, it involved either manually re-entering the data at one company, or having custom-coded hacked-together import/export modules at another. EDI was nice in theory, provided you had small enough volume to re-enter the data or had a big enough IT department and were on systems that actually allowed you decent access to the data tables (and additionally your company was typically holding orders to do purchase allocations in a boatload of spreadsheets *before* allowing orders into the system). Oh, and if you were trading with more than one partner, expect to have to have your expensive subscriptions with multiple EDI interchange services. If your production lead-time was closer to 6 months, "just-in-time inventory" was not going to happen.
One can only hope the software has improved, and the interchange services have finally, grudgingly, agreed to talk to each other.
I've never done that.
I did write a deletion script for another department which escaped its playpen and wreaked nagasaki across a wide swath of production, high-volume directories.
I was able to put on a horrified look and say "Who changed my code?" before we got to the Blame Allocation Phase, fortunately.
I once gave a client a test version of our software to install on his laptop, unfortunately he somehow managed to get the installer to attach to their main production server and wipe the live database, instead of installing a new database server instance on his laptop.
I was in another client's office (they were both up North) and spend half the day purchasing I think RedGate tools to roll back the transations to before he'd reinitialised his main database.
Sweaty....
I was a Systems Engineer for a local IBM reseller at around this time. The company made TONS of money performing "miracles" with PCTools, Norton Utilities and Maynard parallel port tape drives.
These were the days when an SE's competence was measured by 1) who had the most Microchannel Adapter Description Files in his recovery toolkit (the most obscure, the better), and 2) who could cram the most drivers into Himem.sys and autoexec.bat. It was always tricky finding the space to get the drivers for the Token Ring NIC, external CD ROM, and the aforementioned Maynard units to load.
Now, where's my pipe & slippers...
I suspect this has possibly been mentioned when you previously used that stock photo, but... I don't care if it's bright red, who on earth is using a keyboard that has the "delete" key where the right-hand Shift normally goes?!
That's just asking for trouble...
Installs new application.
Install procedure cleans up after itself.
Reboots system.
System falls over.
What do you mean, SET TEMP=C:\DOS ?????
That resulted in a bus ride home and back to get some DOS floppies to re-install the C:\DOS directory, and savage editing of the AUTOEXEC.BAT file.
Been there, seen that. Quietly recovered floppies with PC Tools unformat and carried on.
Actually there was more to it. This person wasn't a misanthrope, but an engineer who had recently read about new class of "CMOS viruses" and was clearly afraid of them.
Others tried to explain him that common RTC chips had only about 70 bytes of CMOS memory. Furthermore, as the BIOS never tried to execute code from it, there was simply no way to use RTC as an attack vector. Yes, there were some viruses around which wiped the RTC CMOS memory, but that wasn't quite the same thing as propagating from there.
He, as a supreme engineer, obviously knew better and proceeded to format all floppies in sight. Which we quickly recovered once he had left the area.
Later he went on to become a fully blown PHB.
And 25 years later, with the advent of UEFI, flash-based viruses became possible.
And 25 years later, with the advent of UEFI, flash-based viruses became possible.
Yup. IT industry seems to have a lot of people who see a hoax and think "Hey, that's a great idea, we should make that possible!"
The Good Times virus hoax - claimed that just opening the message would let your machine become infected. Complete garbage as emails (and netmails etc) were just text, nothing executable. Then we got HTML and various scripts in email.
Another variant on GT was that just the subject line listed in your email program was enough.. Not possible as subjects were just text, then hey presto, a certain Redmond-based company released an email client that could be infected via enhanced subject lines (or was it a buffer overrun error?)
Virus that could destroy hardware or at least screw up your firmware and brick your computer? Not possible at present, but hey let's fix that! (wasn't that Redmond-based firm heavily involved in UEFI standards as well???)
If only these people could be put onto FTL travel, or teleports, or over-unity power generation, or....
That rings a rather discordant bell. One client used it fairly heavily, mostly because HMG Depts seemed to favour it. One of them took a few weeks to tell them that their machine - to which they were being sent invoices - was out of action. My client's FD wasn't pleased. They also sent orders over to another client of mine. I had lots of ?fun fixing up the problems with their "chief developer's" EDI "parser".
During the late 1980s and early 1990s, I worked mostly with Data General minicomputers. There the equivalent of "cd /; rm *.*" was
DIR :
DELETE #
The good news was the :PMGR.EXE, the peripherals manager, was fairly high up in the order of files in the root directory (hash-organized), and once it went the system would forget how to talk to the disks, halting the operation. If you were lucky, the slob who had run in the last update would have left :PMGR.EXE.OLD in place, and you could rename it, and start the recovery. If not, you would have to go to the "Systape". I first ran into this when a system halted, and a guy from another contractor called from the server room to ask where we kept the systape. I later did it myself, but I think only once.
SELECT FROM table. WHERE colname = some complex IN condition probably including a Count; to see what will be deleted. Followed by an edit...
DELETE FROM table WHERE colname = some complex IN condition; distraction from luser’s complex call just before pressing [Enter] so edited statement becomes...
DELETE FROM table (now has semicolon at the end of first line) ; no I don’t know how it happens either...
The scariest word in IT....is "Oops"...... followed by a scream.
I disagree. Followed by silence is scarier..
Followed by the emitter quietly starting to pack up their desk, not even a hint of bothering to try to cover their mistake or undo their error? Scarier still..
I (and Kevin) had to deal with a situation in the early 1990s where the culprit (Carl) had been the last person to leave on Friday (and by the time we arrived on Monday was abroad on leave). He'd left some sort of unfinished reformat and reinstall job without so much as a note about how far he'd got and what was left to do.
During his holiday someone phoned for him and when I said he was away they asked "Oh! How do you manage without him?".
Reminds me of having to placate a livid customer who received a message about having used up all the company's monthly bandwidth mid-way through a month.
I was very puzzled until I went on-site and deduced from the managed switch traffic logs that the company's creative designer was the culprit. Broke into his office and found he had left a torrent going, downloading movies, whilst on vacation, hoping that noone would notice.
I seem to recall "editing" command.com to disable del as a command., forcing use of erase. Somehow, typing the 5 character command rather than the 3 character command would give me just enough pause to dramatically the occurrences of erroneous delete, and reduced the workload on my pctools disk dramatically
(1) I cannot count the number of times Filer turned out to be a life-saver for my customers using Netware. Even if they had a recent backup I found it far easier to look for deleted files/directories using Filer.
(2) Another customer not using Netware: their HR manager wanted a backup configured on her pc. Anyway, prior to our visit to set this up, she thought she'd have a spring clean. She somehow ended up formatting the hard drive. Norton Utilities Unformat to the rescue. In those days people often used to store data in the root directory, but she'd used the data directories we'd set up for her. As a result everything was rescued.
I suspect "Art" may be South African. It was common, in South Africa, to make a distinction between 5.25" removable disks and their 3.5" siblings as being floppies and stiffies.
That time when. working for a South African owned company in the early 90s, when someone arrived in the office and needed to save some work, he went to the department secretary and asked for a stiffy? Priceless. I wonder if she ever got over the shock.
In the mid -70's I worked in an IBM shop and one day I was tasked with making a backup of the system procedure files. This was a Partitioned Dataset (remember those?) of JCL decks (remember those?) of regularly scheduled jobs. The operators of the IBM 370/158 could invoke these by name as directed by their 3-ring binder of things to do that day. In those days the "job scheduler" was a human being.
Ok, IBM provides a handy utility that copies such datasets (today we call them "files") from one disk to another. I typed up the necessary JCL commands on the 029 keypunch (remember those?) and ran my job. Unfortunately I got the disk drive names backwards and ended up copying an empty dataset on top of the live one. After the expected yelling at by my boss, I had to go find the last backup tape in the tape library and use that to recreate the whole thing. But in the meantime the operators could not run their jobs.
Many years ago I worked for a firm that had a modified PC set up to allow non-technical staff to produce the software distribution disks sent out with the hardware they manufactured.
The machine automatically booted into the program, and normally sat at an "Insert disk and press Enter" prompt. The software then automatically formatted the disk, copied and verified the appropriate files, produced a success or fail message, before resetting ready for another disk.
Both floppy drives on this machine had been modified to ignore any write protection tabs, as the custom printed 5.25" distribution disks produced did not have the usual write-permit notch.
There were warning notices taped to the machine, and screen text explaining what the machine did, but people regularly overwrote disks they tried to copy with it.
Ran this bad boi exactly once in my life. And spent an entire weekend using Norton Disk Doctor to rebuild the hard disk from scratch. Not a bad effort if I must say so myself, given that I had essentially zero prior knowledge of IBM disk structure.
Deleted it on any and every PC that I had anything to do with after this little lesson in life.
I had a yellow plastic utilities boot disk and a red plastic utilities boot disk, so the people knew how much trouble they were in. PC-Tools, XTree, Norton Utilities, Laplink (with a RS232 cable as few parallel connectors existed), various odds and sods routines and drivers.
I had legal copies of all the software but couldn't risk those disks in other peoples drives.
In the mid-80s, PCs were not the only desktop; the Victor/Sirius had morphed into the Apricot, and they were sexier than a PC.
So our MD gets one of the first hard-disk Apricots for his secretary, because only the best will do.
Unfortunately, on this machine, the hard disk is drive "A:", not "C:".
And the first time his secretary needs to format a floppy, she puts it in and types:
> format a:
like you do. It replies:
> "Confirm you want to format Winchester Disk?"
And never having seen the word "Winchester" in her life, she assumes it must mean the floppy, and she goes ahead.
Eek.
This was a very early lesson to me, as I was morphing into an allegedly user-centred software designer, that if there are places in your product where "things written by programmers" (as they were called) surface themselves, you better find them, before the end users do...
(And no, there were no backups, but then again, each backup would have taken you about a day.)
Takes me back to when I always carried Norton Utilities and PCTools with me wherever I went.
I remember trying to help out one poor user whose WP program wasn't set to do automatic periodic backups while running. I was the new IT help guy for the room, learning what had and had not been done, or setup. The user had put in a lot of work on one document, and was apparently used to saving on exit by choosing exit, then responding to the "would you like to save first" prompt. Unfortunately, she chose to save the document with the title "con"*. She, like almost all users, didn't know, and there was nothing in the OS or WP program to advise her, that "con" was already spoken for. I happened to be across the room, and arrived just in time to watch the end of the document scroll past on the screen, and explain to her what had happened.
I went through the machines one by one, and set the WP program (the only thing used on those machines) to backup every few minutes, and posted a notice on the wall with "forbidden" file names.
*a law student preparing something on "con (constitutional) law".
Except that the system was Xenix and I did:-
rm -rf *. lst (note the space between the *. and the lst....)(bad keyboard)(and I was at root and the root user).
I stopped it before it did too much damage, it really just took out the system files. It was also a critical system and "back in the day" (late 80s/early 90s) recovering a system was a long and tedious process. So we just left the system running while we finished the development we were using it for, hoping that there wouldn't be a power glitch in the next six weeks.
Lesson learned.
Admittedly deleting anything on a FAT FS under MS-DOS was a very long time ago for me, but didn't that process usually work by nuking the first letter in the filename, that you later had no way of retrieving unless you knew? No unerase tool could help you with that. I mean, *imem.sys is pretty obvious but who knows all the myriad others...?
but didn't that process usually work by nuking the first letter in the filename, that you later had no way of retrieving unless you knew?
That's how I recall it as well. Undelete would find the file OK, but you needed to provide the name.
I do have a niggling memory that there were pointers to the next part of the file or something that may also have contained the name, but I did often do data recovery for people so my recollection of this stuff is an amalgam of remnant bits of lost knowledge :)