Close
You weren't by chance given a lucky clover by a witch when you were born were you?
Welcome once more to Who, Me? where readers share their panic-inducing moments of tech support cock-ups. coffee beans A quick cup of coffee leaves production manager in fits and a cleaner in tears READ MORE This week, we meet "Adrian", who was consulting for a local weekly newspaper in the 1980s when he nearly destroyed 20 …
This post has been deleted by its author
“—Also, Stevie, ‘Gypsy’ is not the preferred nomenclature. Romany Englishwoman, please.”
But "Gypsy Wench" conjures up a vision of a beautiful wanton young temptress from a Smirnoff Vodka poster I had on my wall circa 1974 and Romany Englishwoman conjures up a vision of Lister's Gran from Red Dwarf
30 years ago I was in "build" on the mainframe which involved compiling source, and printing the listings of those which failed to compile.Typically I used to get a box of listings about 2 feet high each morning. Just before I went on a weeks holiday, I tweaked the process to "improve it". I came back to find a new wall in my office 6ft high and 8ft wide. I had to demolish it to get into my office and get to my desk.
Back in the early '80s when I worked with ICL 29xx mainframes, we ran line printers (AKA drum printers) that far exceeded that rate of printing. IIRC they produced about 300 lines per minute on 14 inch wide, striped, fan-fold paper.
We used to collect the output in boxes for ease of handling.
Used to work at a major North American totalisator company. The (duplex) transaction processing system had 900 line/min Dataproducts printers. They printed intermediate reports every minute, which were not needed unless something broke. Only the final-cycle wagering totals that printed after each race started, and the payout calculations performed after each race's result was determined, were kept; the rest of the printers' greenbar paper output went into 55-gallon trash cans. The operators generally left the printers' acoustic covers off, as they got fewer paper jams that way. Many operators seemed to be half deaf as a result.
I was working on CP/M systems that managed public information displays. I got tired of waiting for source listings on the 30-character-per-second Decwriter serial terminals that were available to me, so I designed and built a card to drive a Dataproducts printer when it was not required for its intended use. It probably took longer to swap the cable than to print one of my listings.
Dataproducts lost our business after they switched their ribbons from towel-style to a cartridge with rollers that attempted to stuff the ribbon in at the end after use. They kept jamming, and the repeated hammering would tear holes in the ribbon in short order. Centronics had recently bought the CDC printer line. While they used the same band character element technology that DP had adopted, they had kept the towel ribbon and that induced us to switch. I'm sure that that salesman ended up quite happy that he cold-called us that day.
Those impact printers probably stayed in service another 10 or 15 years. No doubt it's all laser now.
The line printer for my mainframe on my destroyer could chew through a box of paper in less than four hours straight use, which happened about once a quarter when I'd run my ginned up program so they could do a full audit of our on-board inventory of parts. That was one wild and crazy piece of software, written in a structured basic that took in five suspended prints, built its own database on the fly and then chucked out three separate sets of inventory audit sheets out by bin/storage location listing all the National Stock Numbers but only a open line for the quantity located there.
Now the first version, that was in a scripted line editor and pretty damned gruesome!
Errr, I think you lost the thousands digit.
The slowest ICL 1900 printer, the 1931/1, was good for only 300 lpm, with the 1930 managing 1000 lpm and the 1933/3 ICL193 being good for 1350 lines/minute with up to 160 characters per line on 14" paper.
IIRC the 2900 used more or less the same printers, though in orange cabinets and with the paper tape loops used to set the vertical page replaced with electronic analogues.
Aye, and how many knew that the venerable THROW TO CHANNEL TWO meant drive the tractors until the hole in track 2 of the paper tape loop was spotted by the sensor, and only worked right if the operators had remembered to put the general purpose loop back in the printer after they'd done the TAD and Payroll runs?
The printer I mentioned in my previous post was in the Series 39 range. Printer speeds started at 1000 Lon and went up to 2000 lpm. These were the real operational speeds and printers were used at roughly 80% utilisation 24 hours per day. Do t forget most information was processed on the mainframe and printed them delivered to departments. Larger installations would have peripheral operators charged with feeding the printers and I knew if several sites who would join the end of one box of paper to the beginning of the next to avoid any delay. It could be a real problem to unload paper fast enough
And even at 1,000-2,000 lpm, in the 1980s, that wasn't exceptionally fast.
The IBM 3800 could run an order of magnitude faster than that, at up to 20,000lpm (350 pages per minute).
When the thing[1] got installed at the shop I worked at (which was never knowingly undersold), the operators hated it, because it needed to be fed every 10 minutes or so. At 3,000 pages and about 12 inches per box, 6 boxes per hour, one of those beasties could produce the wall-o-paper in 8 hours, and data processing ran 24x7....
[1] I don't actually remember whether it was an IBM or the Xerox competitor that won the business; back then, most everything was IBM, but this was a massive investment and only made sense because we had a good distribution system in place, so all the centralized reports could be printed in London and sent to the outlying stores.
Amdahl had a largish printing facility just off Central Expressway in Sunnyvale (Cupertino?) in the 1980s ... Virtually every mainframe print-job generated within the company, world-wide, was printed at said facility, and then next-dayed to whatever office had requested the print job. They had several gawdawfulfast channel attached printers, and a fleet of trucks delivering & sending out paper. Was awesome to watch in full-swing, if you had adequate ear protection.
However ... it sounded daft then, and still sounds daft now :-)
Hmm, I recall reminiscing about fast band printers and being trumped in spades. Bloke used to work as an operator for ${big_four_bank}. This was back in the days when every customer got a printed statement monthly. I had sort of vaguely pondered how this happened and the answer turned out to be laser line printers. Continuous stationery, lasers a line at a time. Speed in the tens of thousands of lines per minute.
Anyhow he said that swapping boxes was a chore so they'd leave the back of the printer open, tape the last page of box one to the first page of box 2, etc across the room and press go. Drawback: A misfeed would fill the entire computer suite with paper before it was humanly possible to get the thing offline...
This is not IT related, but until few years back, we had outside movie theater in Thailand. They would install a screen on some premises (like a temple court yard) on some festive occasion, one projector, big speakers and show 3 or 4 (old) movies during the evening. you could rent their service is you organized an event, like a marriage and invite all the people from the surrounding village.
So one projector only. To make the change of reel faster, the operator would tape the end of one reel to the beginning of the next one, all while the movie was playing. You would still get the countdown synchronization at the change of each reel, but there was no interruption.
Since DVD and YouTube, they have mostly disappeared.
Bless.
We 'ad proper printers in them days, wi' duty cycles measured in days and throughputs y' can only dream of wi' a laserjet.
Of course y' damn near went deaf if you were working in the same room as one, and if management had cheaped out and bought a "shuffle printer" a long print run could involve two junior operators employed to hold on t' printer and steer it away from owt too fragile when it got the bit between it's teeth.
Client ran a few of them. Their business was printing, quite a bit of the work with those printers was with MICR toner. They were probably some of the cheaper devices in the plant.
One interesting fault on one occasion.- ink failed to adhere to the middle of the document. It turned out that the printer had been run a long time on one particular width of paper which had worn a groove of that width in a roller in the fuser unit, very shallow groove but sufficient for wider stock to sit clear above that part of the roller.
My IBM 1403 can do about 23 pages (~1400 lines) of 11X14 (132 columns) per minute. Can crank up to over 6 feet of paper per second if the printout contains a lot of blank lines. It will keep up this pace as long as you can keep it in fan-fold paper. Ink optional, unless you plan on reading it. It was made in 1963 (I have another one from 1960 for spare parts).
At that point in time: tractored (that's what it was called here) paper, you know, with the holes in the sides for the toothed transport wheels... Get a box, feed the first sheet into the printer, and it would spit out "endless", zig-zag folded paper. Into a box, or a pile on the floor.
Line printers then could print 2000lines per minute that equates to 33 pages a minute. If a page was only partially completed the page throw would take less than a second to complete. I seem to remember a box took less than 15 minutes to print. The paper quality required was very high, if it was not strong enough the sprockets driving the paper through would tear through the holes, if it wasn’t glossy enough the printer playpen would differ excessive wear and the manufacturer would charge a fortune to replace it. We used to have two pallets of paper a week delivered these were 5 foot high. That probably extend to approximately 100 boxes per week.
"you were probably using a pretty good printer."
Not all 1980s printers were FX80s.
in the early 1980s I watched chain printers printing phone bills - the fanfold paper would climb two feet above the printer's exit port before arcing down to the tray. Most telcos had 2-3 of these running all day every day.
Printers like the Epson DFX9000 (which still exist - at £4000 a pop - and with £1000 printheads incorporating a sapphire block at their core.) run at over 1500cps - I had one of these and rented it out extensively in 1999 - these are still popular in places like fulfilment centres due to their ability to make 10-part carbons.
Back in the mid-80's I worked for AT&T Philips Telecommunications based in Malmesbury where we did all the software dev for CCITT#7 BT for the DDSN. The system running the exchange hardware engineering ran on a DEC in Hilversum and as a part of my multiple day trips between both locations I used to bring the 'green and whites' back in specially constructed grey suitcases that held two stacks of fanfold paper. My record was seven suitcases in a single trip, LHR Customs took an interest in me that day - "Why is this your third trip this week Sir?"
The good old days when you could arrive at T4 an hour before a flight and still have time for a pint and sarnie, as a non-smoker I used to sit in Biz Class Smoking because it was empty, you were last on/first off the plane and beat the hire car desk queue in AMS.
Back in the day, whilst on a Cobol training course we came back the next morning to check whether our efforts had compiled or even, celebrate..., actually run... to find a six foot pile of system dump in the name of one of our classmates. (ICL29xx, i think)
She was very pretty, smart and all in all a lovely girl but really not suited to programming and papering the room with her fan folded output was not a very nice thing to have done...
She transferred to a marketing role soon after.
And, NO, I wasn't involved standing on chairs with the bluetac...
A few years ago, I noticed an option in our backup software to output a printed report after each job (I already got an emailed version). I clicked the tickebox, and went home for the night.
The next morning, I wandered up to the main printer to see if my report had been printed, only to find that the printer was stalled because the output hopper was full. Yes, the backup software had printed the full report, including the filename of every file backed up.
Fortunately the output hopper could only hold about 30cm of paper, and I'd got in early enough to chuck it all in the recycling before anyone noticed.
I stuck to the emailed report after that.
Some 15 to 20 years ago I was to check a malfunctioning PC belonging to a one person company.
Windows was stuffed and I fixed it (can't remember details), and in the end I let the PC run all MEM & HDD tests with our licensed PC-Check diskette.
One of the disk checks is destructive, filling the drive with zeros. Running this particular check asks for confirmation, and I typed YES when prompted. I still don't understand why... Some hours later I realised the mistake and got a rather ill feeling in the bottom of my stomach.
Turned out the computer was the ONLY computer at the client and everything went with it. Of course she had no backups whatsoever.
My boss was rather unpleasant about all this and we ended up paying and setting up a new computer set, wireless network and a color laser printer for the client.
Having something that can erase a disc on removable media labelled "PC-Check" should be grounds for a lawsuit. Just asking for confirmation is inadequate.
It would be like having a car with an extra position for the gear lever which, say, just required you to lift a small collar and then you could do something which, if you were travelling down a motorway, would cause the gearbox to auto-disassemble.
Ah. I seem to have spotted the attitude which creates this kind of situation. "It's the user's fault if they do it wrong."
"Having something that can erase a disc on removable media labelled "PC-Check" should be grounds for a lawsuit. Just asking for confirmation is inadequate."
How do you test a disk write function without writing to it?
I suppose you could read a track, save the data, write zeros, check the result and then re-write the original data which would be very slow.
AFAICR the SCO Openserver install disk offered non-destructve and destructive write test.options. You were expected to know what you were doing and take care.
I used to work for a company that had a knack for putting reverse logic questions in their GUI/command-line software, "Do you wish to not continue? Y/N" type of stuff. That, and same for config parameters, like
"disable_override=yes" kind of things.
Made for great troubleshooting sessions.
Pegasus Opera reverse ok cancel, so you tend to click cancel if you do not read, until you panic click on OK because you want to cancel.
What a shit idea.
I do not at all blame myself and this is what backups were invented for.
Mind you they were not fans of us, a customer with it had lots of index errors, (Pegasus said backup) so we reindexed using a software product of Extended Systems which happened to support a few variants of that file type. We knew that format trashed indexes so used the server engine we use to not get any index errors.
Mind you some was the natural animosity between Nantucket and Fox developers.
The trouble with repeated confirmations ("Are you sure?", "Are you really sure?", "Have you actually thought this through?" and so on) is that they don't actually induce careful consideration. It tends to be more a case of Y-Y-Y-Y-ohshit".
A better approach is something like "Please type 'I REALLY WANT TO FORMAT DRIVE A:' to continue", with appropriate mechanisms to disable command recall and copy/paste. Even better would be something that changes every time it's used, but that would be tricky to implement on CP/M.
when we implemented our first corporate Unix box (IBM RS6000) the PHB's suggested that the system operators should have admin rights as part of their role expansion. This was justified because the admin functions were initiated via a menu. I got one of my sysadmins to ad a delete system option (which did nothing) this asked twice if the operator was sure they wanted to continue before telling them that the system was being deleted. It took a whole week before we got a log entry saying someone had actioned it. This was prior to installing any production services on the box and I won my point. We spent the time in building the ui for the ops to perform backups etc in a less privileged user than root and the sys admins kept root access control;
""WARNING! ALL DATA ON THIS DRIVE WILL BE LOST. ARE YOU SURE YOU WISH TO PROCEED? YES/NO.""
You also change it so that you don't type "Y" to proceed, but the word "PROCEED" or "L" or something non-related to the command - so you're sure they definitely read the question.
Seem to remember the old partition manager on early windows builds required "L" as the confirmation key to delete a partition. DBAN needs a word typed.
If it had been me planning it, I would have had a second disk with the destructive write test and a suitable warning label.
I have to tell you this, though. At one company I worked for I was promoted away from a section which had a PC-CAD system. The component library was on the hard disk and also on 24 backup 1.2 Mbyte floppies (yes, this was some time ago, yes they had PC CAD then.)
A numpty, who not long after departed for other pastures for reasons to do with using company assets to run his club, managed first to erase the hard disk and then, in an effort to recover the data, somehow fed all 24 1.2Mbyte disks into the machine, erasing the lot. I never heard how he managed to do it.
Back "in the day" we had a DOS based program (i.e. pre-Windows) where one action was destructive. It required the input of a random character to proceed.
The disk wiping software I now use requires you to type "ERASE-THIS-DISK" before it will do so, impossible to type accidentally and sufficient to make you think before you enter it.
I'll confess my favourite trick was entering format A: on an ACT Apricot (A: was the HD, B: the FD). Norton Undelete was the saviour.
Automatically protect against? Nick, may I introduce you to dd?
Yeah, did that a few months ago. Working on a laptop, and SSHed into my main computer to look up some settings, copy files, etc. Then did a quick DD to zero out the partition table on the laptop to prep it for reinstall... Guess which window I was in....
Fortunately Linux keeps a record of the partition table, so I was able to re-create it.
pfft. I was thinking more in terms of what we've collectively learned about the dangers of working as root (and giving root to people). And of the fact that even windows now separates out different users and roles. And even of secondary nonsense like aliasing, overloading or renaming system commands to confuse users, or of nonstandard PATHs and LIBRARY_PATHs, etc.
Last Thursday, I was doing a Dell laptop from 2014, complete with a Broadwell Core i3 and, lo and behold, a hard disk.
It was unable to find /Boot/BCD (either on the "legacy" System Reserved partition*, or on the C: drive). So I whipped out my phone and booted the Windows 10 installation ISO. After switching the system's BIOS (firmware, for pedants) to UEFI mode and enabling Secure Boot, I ran MBR2GPT, had to delete a couple of empty partitions, then rebooted. Windows 10 booted successfully, albeit after a heeeuuuge wait.
Did I do the sensible thing and return the machine to its owner? Hell no. Pedant Me thought that Windows 8.1 would be a lighter load on the hard-disk-laden machine (which it actually is) ... inserted my Windows 8.1 DVD (because it's genuine media, or otherwise I'd be dealing in ISOs all the way), then formatted.
Surprise surprise, the machine's HDD was too crap to boot 8.1. I was left with a machine stuck in limbo, with neither Windows 10 nor 8.1. Had to waste hours that day coaxing anything at all to boot.
TL;DR: When you are on a tight budget with available time (deadlines in hours), trashing a machine (well, an OS) is really equivalent to trashing important stuff on the machine, in terms of time wasted.
You probably mean SSDs ...
Convincing plebs who own that kit that SSDs are superior is like convincing them that carpets manufactured using a nanoquantum polymerization process are better for increased oxygen circulation in the room ... they just flick to "Dummy Mode" then disconnect completely when they hear that money is involved (even with them being the equivalent of $25 for a 120 GB Kingston) ...
You have to demonstrate it to them - takes 8 seconds from a cold boot to login in Win10Pro (was 6 with 8.1 media center go figure) on a 2013 Panasonic CF53 Toughbook vs. a minute or so with a brand new stock Dell Precision...
I use Samsung PRO SSDs myself ;)
I do something similar ... I'd show them my own laptop (2014) vs. theirs (brand-spanking-hard-disk-new), or failing that, Windows booting from my phone via USB 3.(whatever it's called today) ... some cave in, some don't give a hoot*.
You use a Toughbook? I was thinking of buying a CF-C2 myself ... but I stepped back when I read that they weighed a ton (I use my laptop like how other people use their phones, can't stand tiny screens, so my laptop is a convertible 2-in-1 with a central rotating hinge, and is always, always on me).
However, I'm still fascinated by these, the mere thought that, as somebody once put it, you can slap-bang it on a table, with people scurrying to protect their fickle phones and Ultrabooks, is simply awesome!
How's yours? Heavy? Light? Can you compare its weight to some real-world object?
________
* nor do I give a hoot about them in this case.
P.S. Yay! I've got a silver badge now ... feeling proud!
Tried the tablet/cell phone route - turns out you have to carry two pounds worth of dongles, if available.
My CF53 has all the ports+ and yeah, it's heavy-ish.... ok, just walked around my house, it's a little lighter than a full gallon of Prestone antifreeze lmao...
However, it is a weapon, defensive and/or otherwise.
@404
Dongles? I abhor these. And tablet/phone processors ... no sir, my tablet/laptop (HP EliteBook Revolve 810) has a proper Haswell Core i5, 2x USB 3.(something), DisplayPort, Kensington lock (if that counts), barrel charger, dock connector, and most importantly, headphone jack.
And of course, Linux + Windows + Hackintosh + Android-x86 + (whatever legacy x86 or x86_64 I ever wish) ... impossible to do on a run-of-the-mill ARM device, yet important if one is doing serious stuff on them (I run hobby VMs under QEMU/KVM and the like).
The only terrible part is the battery ... wish it lasts longer. It gives ~ 3 hours without Wi-Fi and maybe ~2 with it, although it's a Li-ion. It's definitely not enough when your whole Internet life is on that thing ...
I Googled the weight of the CF-C2 (as it's a touchscreen convertible, the particular form factor I need) 3.7 lbs (1.6 kg). Actually, that's impressive, seeing that my current laptop is 1.3 kg, a 300g won't break my arms!
It's a great weapon, but border cops won't like the sight of it if traveling par avion ... especially if one gives it a startup chime that sounds like a buzzer!
I loved my ruggedized Toughbook CF-F8. It had a metal frame with rubber on every corner and a built-in carry handle. The hard drive was suspended using industrial Velcro to absorb shock. The keyboard had drains beneath to divert liquid away from electronics. Everything was recessed; even the DVD drive popped up from the keyboard so there wasn't anything to break off. It was wonderfully lightweight, maybe 2/3 the weight of an equivalent Thinkpad T-series. I'm not sure if replacement parts are as easily available as for my usual T-series, but that might be only what I'm familiar with.
I only wish they weren't so expensive.
I used to bring a white MacBook when I went to customer sites. I got some funny looks when I pulled that out to work on Windows, and one customer even asked if I'd like to borrow something stronger. I switched to the Toughbook. They had similar performance, but the Toughbook made rather a different impression!
@jtaylor
They really are expensive, but (if you're the one paying), why buy them new? They tend to depreciate very quickly (as they are still PCs) ... you can get a Toughbook CF-C2 (2014 model year) for about $200 on eBay now. Better yet, buy one good unit and one busted one for parts (though there is the eternal question of what if the same part fails twice?)
but the Toughbook made rather a different impression!
Two types of portable computer, in my book, leave a particularly James Bond feeling:
- UMPCs (ultramobile PCs) of 2006 ... used to see them when I was yet a wee lad, but the price ...
- And Toughbooks.
UMPCs are now considered underpowered, overheating, and overpriced crap (people on eBay are asking for $200 for a unit w/ a Core Solo!) ...
... But Toughbooks are readily available. Their only issues? Weight and parts (as you mentioned).
Parts are readily available for Toughbooks from Panasonic - even down to port covers.
This CF-53 was purchased a few years ago off Craigslist for $300, spent $200 at Panasonic getting the NATO-grade TPM password reset - had to send it off after 2-3 months of exclusive kitchen table use, parts spread across trying to do it myself lol.
There was a time when I worked for CSC as admin on some systems which a large insurance company had outsourced. My experience was in web middleware (Weblogic, Websphere and that sort of stuff), but this particular installation had a mission-critical installation using Oracle Forms. Which is tightly tied into an Oracle database.
Well, I was deploying a deployment of an update to the Forms application on a Friday evening, and was running short on disk space. So I took a look around and found a directory with a bunch of <whatever>.LOG files taking up a lot of space. Since logs were being collected on a central server, I assumed these had to be something which could be wiped and promptly did a "rm -f *.LOG".
Little did I know that an Oracle database keeps all pending transactions in a transaction logfile, aptly named <whatever>.LOG ...
So after moving the new forms files into place I tried starting the database and got a completely unknown error, but enough for me to recognize that those logfiles probably were a bit more critical than I had assumed.
My good luck was that the database had been shut down when I deleted the logs, so after a frantic bit of google'ing and some Oracle commandline magic the database finally did start without any loss of data.
I have to admit I kept this to myself, but from then on I insisted that there was a database admin on call for future deployments.
Yep, log shipping is a thang with database systems ;-). When I need to go spelunking for files to kill/trim, .log files are a good candidate but all my search tools have always had full path column and I sort by that first before block selecting candidates for deletion, then double-check that again before allowing the act. Come extremely close a few times doing the wrong thing there!
I've been caught on that one. Can't remember the circumstances.
But I recollect my natural caution kicking in. This is an unfamiliar system. It's not impossible the machine the Client has given me access to (for dev/testing) hosts something that matters. And no firm clues like timestamps that would indicate something familiar like regular log rotation. So instead of "rm", I started with "mv", on the directory containing all those "LOG"s. Server won't restart - whoops, rename it back to what it used to be (server now restarts after anxious delay). Find something else to delete!
Yep. One of the final steps to Admin Adulthood is learning to move/rename files instead of deleting them. You can get lucky deleting numerous times with no bad repercussions, but when your luck runs out, you're up shit creek. Always leave yourself a way to get back home.
Back in the days of multi-user CP/M and RS-232C terminals, it was rare to do stuff on site on anything other than the production system so I don't feel too bad for "Adrian". On the other hand, he knew about and expected to find "winform.com" but didn't recognise that it's output was different to what he expected is a bit suspect.
Is it that suspect? The output from the formatter was very unclear, simply asking for a drive number. It did not explain its purpose. If that was normal, it is entirely possible that the expected utility wasn't clear either. Maybe it would have asked for a drive on multi-drive systems before going to the interface with which Adrian was familiar, which he had not seen because previous systems either had one drive or had a utility that figured it out. It seems like a bad process for displaying information to the user, but given that, I can understand getting it wrong.
In CP/M and MP/M (which this likely was), numbers 0 through 15 would be user numbers. The files would all go in the same directory area, but you could hide files from others by changing the user number. User zero is where you'd likely find your executables.
Still, with the choice given, HDD numbers would make for a good guess.
Bothering to have the program announce itself as "Windows Formatter" or whatever it was really called, was too much to ask or something...
Good thing we don't suffer from such limitations in today's world of GB of RAM and TB of storage, so anyone who wrote a program that simply prompted for a location without any way to know it was planning to format a drive would be deservedly shot in the back of the head!
it was a great, if somewhat slow way to make the paper.
My wife worked at a bunch as a pre-press graphic artist before the constant closing of print facilities meant she had to seek another job (she now does tech support for major US tech companies). Her last newspaper job was at her hometown paper, 02-05, and it used that process. Got to see it in operation, and it was actually a really nice, professional setup, and because you built the pages physically, you could see a proof copy in your hands before it was printed. Many's the time I would go to meet her for lunch and spend a bit of time helping proof through the paper.
Alas, in 05, they closed the press, and outsourced it to a print facility 50 miles away, where they'd send the file and they'd digitally print, to save costs (they then sold the print building which became a gym, the offices became a restaurant, and they took business offices closer to downtown.
thing was, because they now didn't see a hard copy first, and the pressmen in that distant city didn't know or care, errors shot up. and often the papers would be late back, so subscribers plummeted, and 2 years later the paper (the oldest business in the county) went under, thanks to 'saving money', while one of the people responsible for this went on to head up the Chamber of Commerce, where she continues to try and kill local businesses to play favourites.
It sure is sad that this grand industry has disappeared. I worked for my family graphics business for a bit. From the mid-1980s until 1992, I was a typesetter. The business did well until the 1990s recession, and then customers failed. That was what I think the first strike on the printing and graphics industry because customers then started going in-house to save money. Laser printers and desktop publishing was first coming out in full force, and output from a typesetter wasn't needed anymore. Shortly after this, the lease ran out on the Varityper, and the leasing company didn't want it back. It soldiered on for a few years as we looked for a replacement, and ended up getting an ECRM imagesetter. We had to keep up with the times. Today, that setup is long gone, dad is retired, and my bro has moved on to other things. The need for film got to be so low that the chemicals would spoil.
In the end, I think the overall quality has gone down too in the industry because people don't care whatever the output looks like. Our local papers are rife with typographical errors. There is no care or pride in what's produced, and everything is about a quick splash and get it out.
When I was at Uni in the early 1970s, I was on the editorial team for the in-house newspaper. We used to gather news during the week, including photographs (on film), and get the college secretaries to type the text up on A4 paper. The night before publication day was spent cutting and pasting (literally) these sheets of text and photographic prints onto the galleys (A3 sized sheets of thick paper/thin card), and at 7AM I used to take the completed galleys in my car to the printers in Coventry (about 15 miles away). They would produce the print run, and I would drive the papers back to Rugby for distribution.
During this time, I also worked as an apprentice at Gants Hill office of Ford Motor Company (my uni sponsors) producing the workshop manual for the Mark III Cortina, using the same cut and paste system, only there we used Cow Gum to attach the text and images. The excess Cow Gum was removed from the galleys by rolling a ball of set gum around, which meant that the ball got bigger and bigger with continued use. There was a prize at the end of the week for the editor who could produce the largest Cow Gum ball.
My first career was in photolitho (mainly magazines and advertising), and by that stage we were working with mylar film for the layouts. Anyway, in 1996, we got Quark Express and similar tools. By 1998, out of a staff of about 30 people, there were 6 photolithographers left. I was not one of them. And of course no-one else in town was hiring photolithographers - they were getting rid of most of theirs as well.
Thus the IT career now. Which, thankfully, actually suits me better (just not one that many women were encouraged to pursue in the 80s, when I was at high school... although it's only improved marginally now).
Back in 1965, the Dartmouth Time Sharing system uses two computers - a GE 235 for computation and a DN-30 for everything else. The DN-30 was a real-time communications processor that handled ~40 teletypes and scheduled jobs for the 235. We would sometimes enter patches from the master teletype to the DN-30. Once the patch was in we would plant a branch instruction to the new code. If we got a line feed after the carriage return we were good. If the DN-30 proceeded to reboot then we knew we had made a mistake somewhere. 40 users would be instantly annoyed.
I don't know if modern programmers still have the skills to enter programs in octal (or hex, now a days) any more.
Some of them certainly do! Go read some of the white-hat hackers papers on breaking into things.
The most recent one I read was about defeating the whole-disk encryption passwords on various SSDs. Basically, the scheme is to use the JTAG interface to get to the firmware, patch the firmware so that the "check submitted password against the real one" function always responded in the affirmative, and wander around the newly-unlocked drive.
All you've got to do is find _where_ that function happens to be, and insert the bypass code. But all you have is the compiled version that you pulled out of the target device...
I did read the post, maybe 'salt' carries the wrong implication, if you really wanted to be secure you would not use such a function, you would use both the password and the stored key to generate the actual key for decryption. This way there is no explicit password check, only a generated key.
Generally the password is not a part of the encryption key - if it was then changing the password would render the device unreadable and not being able to change passwords is a bad idea.
At some point in the code there must be a "does this password match" function and "all" a hacker has to do is to change this code to always return true. The encryption keys must be on the device somewhere, however a small amount of private persistent memory is common in such devices and it will be stored in there and not accessible to anything other than the control chip itself.
This is avoidable; you use the password itself to reversibly scramble the stored key: new password simply re-encrypt the key. The protection on the key is admittedly only as strong as the password, but as the key itself should be random it is difficult to brute-force since the only check is to attempt to decrypt the data with the test key.
What's going on here is a well-founded effort to improve the key strength, but implemented with a flaw. The way encryption is used on these devices is that the user provides a key, which unlocks the "DEK", or Data Encrypting Key. The DEK decrypts the data.
The reason for this is that each of several different users needs to be able to unlock the drive.
A "clean" implementation uses a distinct subsystem like a TPM to store the user passwords (and DEK). The host-accessible subsystem can push user passwords to the security subsystem, but has no mechanism itself to push a DEK into the crypto engine, so the DEK cannot be intercepted and the password validation happens in a secure enclave (the TPM).
And this is why core security functions (such as password verification / encryption etc.) should be baked into the silicon, not flashed into firmware. Admittedly this means that if there is a flaw in your implementation then you can't fix it without replacing the hardware, but that's why you stick to reference implementations and don't roll your own...
The setup here sounded very familiar! Back in the mid-1980s, I repaired Ontel terminals when they became part of Visual Technology. Among their many OEM customers was Olivetti for whom they setup a word-processing system that ran CP/M, although I didn't need that for testing and instead used an Ontel O/S based machine connected to the terminals.
The terminals had what were called at the time word-mover-controllers. These cards did what we do today in memory except it was all shift-registers and static RAM for text buffers used for cutting and pasting text. These cards were a beast to work on too because of all the loops and clock dividers on them.
A few years later, I worked for the family graphics business. My bro ran one of those photostat cameras mentioned in the article, and I ran a Varityper Epics 20/20 typesetting system. I would "set" the type in galleys and the output was then covered on the blank side with wax or rubber cement. These galleys were then pasted down on Strathmore boards and my bro shot the photostats and sometimes films for the local print shops.
What was interesting is the Varityper had built-in filters to handle input from WordStar. In the 1980s, I had a Visual V-1050 that had that among other programs. One day we needed some additional help handling some coupons. Using my V-1050, my sister typed up the text in WordStar and I used the Varityper telecom package to import it from the V-1050. It worked perfectly and this was our backup terminal for a few years.
In the oughties I worked in a print shop that used a Riso printer. Physically looks like a huge photocopier, to the computer it looks like a laser printer, but you "print" to a printing master and then it uses that to print with. At full speed it would print and fill the output box just as fast as you could unpack fresh reams of paper to fill the input tray. Ear defenders were a must - especially when the folding machine was going as well.
Imagine you're installing a network that carries the traffic of an entire government, and you have the most important install (at the very seat of power) - what do you reckon your chances are of that going right?
Exactly.
On the plus side, at the time it was *really* new so we got away with it - we made it sound like the second site visit was all part of the Grand Plan :).
I had a much more minor incident, I was doing web and linuxy stuff for a more established consultant who resold my time out occasionally, and on this occasion he'd managed to sell me as capable of installing some line printers onto a Solaris based warehousing system, a operating system I made clear that I had no experience of to him at the time.
After goading and considerable prodding thought I'd very very cautiously give it a go, after telling Alan yet again I didn't know what I was doing really, and him insisting I tried anyway "as its not live yet anyway". So there's 4 temps furiously entering inventory data as fast as they can in four terminals to get the inventory system populated with the stock ready for the go live in two days time, and we arrived onsite mid afternoon and after a hour I had got a shell and felt about and installed the drivers by running the bundled shell scripts, but things weren't playing the game, so I decided one process called "printr" was the culprit and had to be shut down. It ignored a -HUP, so it got the big kill -9 shotgun, and that worked. Sadly I could tell it had because I heard the screams from around the room as a entire day's data entry went down the toilet as it cached all the data entries in ram until told to write it out, which they did at the end of each day. We fired it back up quick and it was devoid of entries...
On the plus side, at that point the printer started working perfectly, so I apologised profusely to the temps for ruining their work and made my exit. And the temps were paid hourly.
Given that a newspaper MUST publish in order to retain its registration (The Times produced a single copy every day during the very long strike years back) I am amazed at their lack of security.
I worked, some years back, at <national daily> (<ND>). The place had been threatened by the IRA, and had ultra-sophisticated physical security (they used to be able to detect pigeons landing on the roof).
I was in the team that ran the Sun-based pre-press systems. This was a laid back group to say the least: there was NO root password. None. And a modem on one workstation that ran a UUCP connection for private email, but could also be used for dialin.
When I left, they kept my account open and said I was welcome to dial in now and then to check email.
So, sitting at my desk at <new employer>, I'm logged in to <ND>. My new boss leans over, looks at the hostname prompt and says
"<host>? That's not one of ours".
"Nope", I said, it's my account at <ND>.
"Watch". Quick SU and I'm root. "Do you want to kill a national daily newspaper?"