Your mission, should you choose to accept it, is to save data from a computer that should have died aeons ago We have yellowing plastic, disintegrating drives and serial shenanigans to start the week in The Register's regular column of reader misadventures. Welcome back to Who, Me? Today's tale of derring-do comes from Aaron (yes, real name!) and, despite the age of the hardware involved, took place in 2010. Aaron had spent the …
One has to wonder whether they really needed all the old data. IRS rules would only require records going back 3-5 years, depending on the details. Only other reason for old records is unexpired warranties. Keeping old business records longer that necessary just creates liabilities.
... still waiting for your rebate (with interest)
Ah that's the beauty of back duty. If they owe you money, they can only go back 6 years (that was the case when I worked for the buggers), but if you owe them anything, they can go back to when the first lungfish crawled out of the primordial waters.
Awhile ago, but not too long ago, I was asked to erase some hard drives and help recycle some old PCs from church member friend who's husband had died.
Used DBAN on the newer PC but the older one with his business Quicken on it only had a 5 1/4 inch floppy and silly me I didn't have DBAN on one that size. So to speed things up I told her I'd take the hard drive home to clear it and we took the PC's to recycling.
Only at home did I realize the 5 1/4 inch hard drive (only one I've ever seen) needed the pre IDE hard drive controller we had just tossed.
I used to have a colleague who was a sniper in the Dutch National Reserve and he sometimes used old hard disks for target practice (never more than one at a time from a raid set) and delivered them back as proof of permanent decommissioning. The auditors were impressed.
Without wishing to blow my own trumpet (Ta-Raaa!!) in about 1983 I wrote a complete small business bookkeeping system in MBasic that ran for 11 years with one alteration in about year 7 to alter a hard-coded VAT rate.
Same computer, same printer, same software. Almost the same VAT rate.
I've got code running one of my greenhouses that I wrote around 40 years ago. For the first 30-odd years it ran on a Z80/S-100 system with no problems. That's one machine, non stop, except with power-offs for a routine vacuuming out once in a while. The code is pretty much unchanged after all these years (it ain't broke ...), but is now running under emulation on an old headless Slackware laptop. The whole kit & caboodle is going to be replaced with a dedicated system based on the ATmega328 RealSoonNow[tm]. Maybe.
When you bought a license for some software that never expired.
Shhh... don't let on of these Millenials will think you need to be put out to grass.
The software worked and worked and worked as users came and went over the years but the business functioned as it had been for years.
Now... unless someone is there to make sure that the subscription gets paid on time every month the business can keep on working at the whim of the software vendor. If they decided to can the product and stop verifying the license renewal then it is you the business that is in deep do-do especially as they VATman wants his tithe monthly these days instead of quarterly.
But hey, that's supposed to be progress 'innit'. 'cool'. and other words that have [redacted] modern use.
Software subscriptions ==> [see icon] IMHO
But hey, that's supposed to be progress 'innit'. 'cool'
No, that's supposed to make the software company more money.
Really the bigger question is why software companies ever sold their wares for a fixed price, instead of taking the razor-blade business model and charging perpetually, for everything.
My first naming convention for servers was based on Metallica album titles.
Master of puppets was a dc
Live, shit, binge and purge was the backup machine
Random supermircro black tower was the black album
Kill em all and ride the lightening were reserved for psu failure
To be honest. I'm guilty of still doing this with every system I manage with their hostnames or other settings, each one has a unique name.
Which is why the store is kicking out two wifi signals named:
Morticia
Gomez
From the router who's hostname is "Wednesday"
But there are also some systems running with fun names like Viktor, Lucifer, Dracula, Ophelia, Lurch, Selene etc.
It just makes it easier to instantly know which system In managing and has nothing to do with treating them like my children.
Also makes "ssh dracula" more interesting and easier to remember lol.
With IPv6 now in play, I also get to choose fun IPv6 names like "I feed all dead", "Feels good" etc. :-D
I like to think one of my legacies will be obscure knowledge based naming conventions:
I have been responsible for (all at different sites/employers)
- A Convention based on call signs and ship names from top gun
- A Convention based on vehicle, character and mask names from M.A.S.K.
- A Convention based on the roster of mortal kombat fighters
- A Tagging effort using paradise lost song titles (forever failure was an unreliable SAN unit natch)
And i inherited a system which had used trumpton as its naming convention, was quite neat really as there was a boot order to get things up and running, so as long as you could remember "Hugh, Pugh, Barney, McGrew, Cuthbert, Dibble and Grub" you had an easy time of it come reboot/maintenance/pizza on expenses night
Mine are generally mythology characters of some sort. Athena (a server), Daedalus (dangerously close to Dataloss...), Io (as in I/O), Phoenix (Io but with new mobo, processor, and memory after mobo failure, so "resurrected"), Janus (dual-boot), Fafnir (mini PC), Nyx (black laptop)...
One of my favorites is my current laptop. Only comes out at certain times of the year (vacation or when we have guests), so it's Persephone.
This post has been deleted by its author
I went on a training course in Bristol, we each had a trial copy of MS Server 2000 & asked to give our machines a name.
We were then asked to announce our server names.... when it was my turn I announced:
"Zen!"
Instructors response was "There's always one!"
Icon - What happened to Star 1.
The first 'networked' computers I ever came across were some 480Z machines at School. Ford, Trillian
The tape only 380Z was called Arthur
One of the BBC Micros was Agrajag, and possibly Zaphod
Then came a new machine, I don't remember what it was other than it was from Research Machines and called 'Deep Thought'
No Apple Computers so no Douglas ;-)
"so as long as you could remember "Hugh, Pugh, Barney, McGrew, Cuthbert, Dibble and Grub"
The snag is that there never was a Hugh - these were all surnames aside from McGrew who got his first name mentioned too. There were two actually "Pugh"s. I wish I could recall useful knowledge as readily as this trivia!
My first network was named after Greek heroes. We got in a MicroVAX that the original admin named Ulysses. So we later added Ajax (a 486/33, no bloody DX or SX) , and a couple of others with names I can't remember. When we got a Pentium II/200 box, I named it "Helen" because Helen of Troy was "fast".
VAXes named Hera and Zeus... VAXstations Eos and Io... terminal servers (remember them?) Calliope, Clio, Erato, Euterpe, Melpomene, Polyhymnia, Terpsichore, Thalia.
A subsequent generation was named after Norse gods, at which point people started saying that the IT department was staffed by Nazis.
Where I work, I think one of the sysadmins has/had a toddler and consumed too much CBeebies. The VoIP system has two border servers - Makka Pakka (it's rock solid), Upsy Daisy (failover) and the call queue server Pontipines. I might try to get my own back with Shaun the Sheep....
... Maybe not, that's a baa-d idea.
I once built a Windows domain called ATOMIC. The original idea was to have each server type named a different period of the periodic table. The two domain controllers were called Neon and Xenon, named after noble gases because they don't do anything. Then we decided to call the Exchange server Mercury for obvious reasons. We finally forgot all about the period idea to call the two terminal servers Titanium and Arsenic purely so we could shorten their names and say stuff like "Dom is logged into Tit, Brian is on Arse". Yes, we were much younger back then and things like this were considered amusing.
The Linux side of the business named all their servers after ales. If i remember correctly I think one of the DNS servers was called Riggwelter.
In my first job working in Ops back in the dim mists of time (OK, OK, mid 90's) we had a naming convention that all the Servers would be named after male Star Trek characters, and all the Routers, Modems, Mulitplexers, etc. would be named after female Star Trek characters.
Except for the router we had up in the North East, which for reasons of linguistic piss taking was called WAYAYMAN . :)
I still do every time i have to do an exam to renew competancies remeber my entry for my CCNA or P forget which looked like this
Fartknocker (switch) connected to buttmuch (router), via dumbass (serial link) talking to cornholio (decent router in lab) with a static route to the philosopher [ASA] (named after there awesome vid review of the song by death), to access sadam-a-go-go (intranet - GWAR song) and rev horton heat (internet - was there an episode without some psychobilly????)
"Death Magnetic" for destructive dev testing, 'cos you really don't care about it!
( I love James H to death, as a spotty 14 year old in 1986, my long blond hair and " big tongue sneakers" Hetfield was my idol and but after Load it all went downhill really fast for them. "Lawnmower Deth", now that's a band to be serious about! )
That would bemy choice for a secure data destruction box :)
Yeah Black Album was the end of it for me (for studio albums anyway still rate s+m), you can hear the exact point they run out of mustain riffs half way through the never...
Ahhhh lawnmower deth, how i miss john peel playing them at the wrong speed and moaning extreme noise terror were nether extreme or noisey enough to be terrifying, have a pint to go with your good taste in metal \m/
I have been guilty of modifying network diagrams and nmap reports to check if people were reading them, depressingly i have only been asked once in 20 years about why we have servers called:
pr0n
warez
s33db0x (used to be l33chf4rm but times change)
r00tl0lz
&
free-iphone-unlocker is the new one i added 10 years ago
One of our clusters was 'Kido Butai'; the machines in it were named for Japanese aircraft carriers, battleships, cruisers, and destroyers. Another was 'Force H'; British carriers, etc. One more was '3rd Fleet'; American carriers, etc. The last one was 'Snakepit', we were tied of ship names, named for assorted venomous snakes, mostly Australian, to the great annoyance of the several Aussies on staff, though one of them did suggest the name of a certain PM of Oz, as said PM was, and I quote, "The biggest, most dangerous, most evil, snake ever to slither across Australia". Management wouldn't let us use it, though. Pity.
Those of you who are not of a historical bent might look up the names.
Long ago and far away my employer was buying up "mom and pop" shops. These shops served the same market, but they all used different LAN-based software. The idea was to build a centralised service with a new single brand. The new spiffy, centralised AS/400 system was populated by using an old laptop and some printer capture software to get customer lists, product lists, order lists, inventory lists....and so on. Worked a treat! So did the AS/400 plus frame relay network. Pity that the conversion revealed considerable affection for the LAN-based software....and considerable hatred for "centralisation"....even though it allowed for each store to lookup inventory elsewhere. That's the problem with conversions....the technology can work perfectly......not neccesarily true of the humans!!!!
"the 99/4a inherited the 16-bit TMS-9900 processor from the 990 despite having an 8-bit architecture, making it (according to 99/4a guru Mark Wills) "probably the most bodged home computer to ever hit the marketplace"
Worse than the Sinclair QL? 32 bit processor with an 8 bit data bus, that was literally eventually sold with a box sticking out known as the Kludge because they couldn't fit all the necessary hardware on the board?
And not the 4a version (the 4 had a chicklet style keyboard whilst the 4a had a more normal key type). Nothing like a game of Blasto to get the arguments started. Interesting machine that had a choice to start in an equation calculator mode (not available in the 4a).
Only the first units needed that.
If only they'd not rushed it, not used microdrives, and perhaps put a GUI on it, it might have stood a chance in small businesses which were the target market. There was little choice in half-way serious computers in the UK that weren't PCs at that time (i.e. the BBC, not even the Amstrad CPC 6128 had come out yet).
The QL was repackaged as the ICL One Per Desk "an executive tool offering a screen, keyboard and telephone in a single unit with its built in personal productivity suite"
it still had the Microdrives which we renamed internally a WORN (Write Once Read Never) drive.
The device had in internal modem and a vt100 emulator so when a large number of us needed to start working remotely when on-call to support the new range of Series 39 mainframes we were issued with these wonderful bits of kit with severe warnings about taking backups and the ginormous investment the company was making in us. Needless to say trying to use the diagnostic systems or scroll through system dumps on the tiny screen was a complete waste of time, the fact that no backups were ever readable form the microdrives meant that the phone book of connection entries disappeared every time the damn thing was powered off and its general crappiness meant it was generally quicker to get up, dressed drive into West Gorton battle with security for access tot he diagnostic centre make a remote connection in reality the only thing ever used after the first couple of months was the phone (ICL had actually paid to have a second line installed so we could be on the phone whilst dialled into the failed mainframe). The OPD was also rebadged by BT as a the Tonto ' a special' device for configuring their rebadged Siemens ISDX PABX's. Something i wasn't aware of until I became a tech support manager and found the officious 'PABX manager' came under my control. It turned out I was paying a couple of thousand pounds a year in maintenance because 'its a specialist device and nothing else can be used to configure extensions' that piece of junk was replaced with a 286 pc with a vt100 emulator within a week. All in all the OPD was a disaster. I've been surprised to find that there are actually enthusiasts who are trying to keep this monster goinf in 2020, but heh ho there's also a Morris Marina owners club!
"Show creator Trendle grew up in Michigan, and knew members of the local Potawatomi tribe, who told him it meant "wild one" in their language. When he created the Lone Ranger, he gave the moniker to the Ranger's sidekick, apparently unaware of the name's negative connotations."
So quite unintentional. An awful lot of innocent words or names in one language are inappropriate in another.
The Sinclair QL was the business machine of choice for a while in Kenya. It was small enough to bring in a suitcase, avoiding the 160% duty and sales tax after President Moi had declared the only purpose of computers was to put secretaries out of work.
I wrote a 24 hour pedal kart timekeeping program on a QL - multitasking 4 68008 machine code programmes with a BASIC one for reporting.
> The Sinclair QL was the business machine of choice for a while in Kenya.
Interesting. We got a QL in South Africa at the time they came out, and they were not ineffective as business computers with the Psion software 4-pack that came with it. Add in the price point, a quarter or so of an IBM compatible at the time, and it was a wonder.
Fast forward about 10 years ago, when I found a box of microdrives, including two of them full of recipes my wife had collected and transcribed. I used Aaron's technique of a serial connection to a laptop to siphon off the data. And no data loss from microdrives 25 years old either, despite moving countries twice in that period.
There was something satisfying about the QL. I still have some in the loft, complete with colour monitor and those all important microdrives.
> There was something satisfying about the QL
It's worth remembering that the QL was highly anticipated at the time and specs were great. Then people got to use the microdrives for real. :-(
A real own goal: the Spectrum was a games machine and there was large gamer community ready to upgrade to something more powerful. But not being able to load games you've bought is a guaranteed way to kill the market stone dead.
Worse than the Sinclair QL? 32 bit processor with an 8 bit data bus ...
Worse than the IBM PC? 16-bit processor with an 8-bit data bus ...
Sometimes you use a cut-down component with a restricted bus to reduce the cost of interfaces to the rest of the system, because sometimes (you believe) your market won't pay for more. Sometimes you do it for other reasons.
[The QL] was literally eventually sold with a box sticking out known as the Kludge because they couldn't fit all the necessary hardware on the board?
That's not quite right ... the story I heard from Tony Tebby (Wikipedia link) was that the QL was sold with the Kludge because management type at Sinclair took a "daily" firmware image to be sent off to make the ROMs without asking whether it was stable. It wasn't, so the first batch of ROMs contained old and rather dysfunctional firmware. In order not to lose time to market these machines were sold with the Kludge, which was just a more up-to-date version of the firmware in EPROM. I still have mine, somewhere ... I was supposed to return it with the QL when I sent it back to have the internal ROM swapped (or motherboard swapped, or whatever it was they actually did) but I "forgot" and Sinclair didn't complain.
> Worse than the IBM PC? 16-bit processor with an 8-bit data bus ...
Well, yeah. People complained about whether the QL was a "proper" 32-bit machine, but it seemed to be forgotten that the 8088 in the original "16-bit" IBM PC only had an 8-bit data bus too.
Also, El Reg's own article on the QL includes some interesting details on the choice of the 68008 and how in hindsight they could- and probably should- have been able to use the "real" 68000 instead.
The Maxtor 120 Meg drive of that vintage would have been an ST506 compatible full height 5.25" MFM drive, probably an XT-1140 (143Megs unformatted capacity). In 2010, it should have been readable (on suitable hardware, of course) using Linux or BSD. I pull data off old systems like that several times per year. It's quite a lucrative side job.
I was working in Beeston on a corporate reporting system, which was based in Luton.
HP 3000 mini in Luton and me sitting in front of an HP dumb terminal. Connected at 9600 baud to the local board, to a central point at 19,200, to a modem at 2,400 to a Kilostream to the MUX at the other end, to the HP 3000... Serial really doesn't like bouncing up and down those different baud rates!
Using the line editor (think "vi"), it would take around 20 minutes to display a line of text. So it was enter the line number, Enter, count the positions in the listing for the first insert, n X space, "i" for insert, then the text to insert, escape, spaces to the next bit, "d" to delete, "i" to insert etc. All blind. Go away and drink a coffee, come back and see if you got the right number of spaces! Press enter, rinse and repeat.
It took them about 2 days to sort out the problems with the buffering and get everything running smoothly. As it was changing reporting templates, following a takeover, for the corporate monthly reporting system and it was 3 days before end of the month, there was no sitting around and waiting for the comms problem to be fixed, I had to soldier through with the long pauses and get the new templates finished...
Back in the mid to late eighties I worked on DEC VAX systems, coincidentally at both my first two jobs we needed to hook up plotters via serial ports. Due to the way the software worked we could only do this by tapping in to the terminal serial line (first instance) or via the secondary serial port on the VT terminal (second instance).
In both cases we got it working but I seem to recall on the second one we had to add a preface string in the printer driver to stop the VT accepting commands as otherwise it would lock up when a particular sequence was sent to the plotter.
I think I've still got an RS-232 breakout box somewhere!
The number of times I've used RS232 as a communication of last resort.
I remember using a hex editor to capture a CP/M 80 program using a BBC micro as a data capture device, and then writing a program to turn it back into a binary so it could be run on the Torch Z80 second processor which ran a CP/M 80 rewrite called CPN. (CP/M machines were notorious for not having a standard disk format that allowed data transfer between machines).
I frequently used BEEBs as data capture devices, and wrote my own DEC VT52 emulator, and Tek 4010 emulator (this was before Termulator was available). I used to capture the graphic instructions for sessions from the MTS mainframe onto the BEEB so that I could re-display it while not connected to the mainframe.
I later became known as the terminal king at several places I worked, as I could nearly always get a terminal working on a serial line, and home-made interposers made using DB25 connectors held together with long bolts and nuts, with soldered wires between the connectors. Together with a Tektronic logic analyser, I was unbeatable! DSR, DTR, DCD, CSR, CTR, chassis ground and signal ground, XON/XOFF, I pretty much came across it all.
I was also the UUCP/BNU SME wherever I worked, whenever they wanted to do mail exchange (pre-TCP/IP) I was almost always involved somewhere along the way. Add to that Termcap/Terminfo descriptor file writing and editing, and that rounded out my serial credentials.
Printers, X.29 PADs, modems, terminal servers, reverse telnet for calling out through modems connected to the terminal servers, I did it all. The most strange thing I got involved in was a weigh-bridge that was connected to an IBM 6150 Unix system. That was a little strange.
All skills long dead, but I do still find it interesting that in these days of xterm being the (very poor, there was no one single xterm type) lingua-franca of terminal emulation, Linux still has a full-blown terminfo database and ncurses implementation, containing references to terminal types that were long obsolete before the turn of the century!
Back in the early 90's, I was tasked with the job of hooking up a proprietary Unix box to a VAX, though the vendor reckoned this machine would only successfully talk with IBM kit.
I proved them wrong, but almost at the cost of my sanity... I learned more about Reverse LAT and $QIO than a mortal man is meant to know...
(Icon, because you need many of them after a long DEBUG session)
DECs RS232 setup was pretty conventional. IBM 6150 / RS6000 was less so (especially when you consider the unusual 10 or 12 pin connectors that IBM used in the built-in ports). If I remember correctly, IBM used a combination of DCD/DTR/DSR handshaking, but both systems would be DTE (Data Terminal Equipment) devices, so there would have to be some sort of null-modem involved. The secret was connecting DTR on one end to DCD as well as DSR at the other, probably both ways.
I never got to play at a hardware level with a DECServers, but I would expect that they were not dissimilar to the serial ports on KL11 and DL11 serial cards in PDP-11s, which I did a lot of work with. Should not have been too difficult to get working, even using off-the-shelf cables and null modems.
Configuration wise, it depended which way you were going, or whether you wanted both ends to initiate a connection. At the IBM end, you would have had to set the terminal line up as 'pshare'. The other end? Well, that would depend on what OS you were using on the VAX.
Already upgraded them from VT220s to Putty on the PC they had alongside every terminal. Now fighting the retirement of the Xavier 386 based server and collection of DM printers. The guy won't upgrade anything and has noe backed himself into a corner. Wants all the new bells and whistles however I cant find IP Cameras or VOIP phones that support 10base2. Wont pay for a rewire, wont pay for any IT soend and Im getting REALLY bored of fixing it. Not to mention media converters arent cheap. FTTP, into firewall at Gigabit, out firewall and BAM right into an MAU
Get some 10base2 to 10baseT bridges. I'm sure that you can still find them on Ebay, although only being 10Mb/s, you're probably going to have to plug this through a switch that can still talk down to 10baseT.
Heck, I think I may still have one sitting in a crate at home.
Bloody hell, someone is trying to sell an AUI to 10base2 transceiver there! Now that is seriously obsolete.
Get some 10base2 to 10baseT bridges.
Or even a co-ax adapter that had thick on one side and thin on the other. It worked on short distance if the network is not too loaded.
And I may even have some Eth *hubs* with one 10baseT connector and eight RJ45. Not sure if I still have the power adapters though.
Also had something similar happen to me.
Client had an ancient Un*x system that ran on an 286 (IIRC). But they want to transfer the records from this Un*x PC over to PC software. They used some database software, and had dumb terminals acting as clients/tills. It is so long ago I can't remember specific details.
Floppy drives was FUBAR, and tape drive.. yup, also shot. Stiffy? Given the fact that it was Un*x I was not in the mood to try out my weak Un*x skillz, and we did not had Google in those days.
That left the serial port.
Laplink cable in hand, terminal software running on an PC... data went over well, except for one specific record, everything on the Un*x box just stopped when I tried to export that specific record.
I stopped the transfer, rebooted the Un*x box, faffed around a bit and discovered that if I just skip that offending record, then everything works as it should.
So the export was done in two parts, first part before the corrupt record, second part after the corrupt record.
I was pressed for time, and to faff around with a database (or fsck tool) may have taken longer than anticipated, or even worse, messed up the existing data.
The data was all beautiful, no corruption at all. So the Un*x machine got switched off for the last time, and delivered over to the customer, and we massaged the data fit for their purpose.
Client was happy.
This post has been deleted by its author
We have a similar box near my desk. It's an AS/400 and no-one wants to decommission it as the original developer has left and the compiler is custom written.
Nobody writes compilers for the AS/400 except for the compiler team in Rochester. I am not sure Barbara Morris is still leading that team, but she used to.
I worked on an awful system that ran on IBM AS/400 kit and had it's own custom compiler. You wrote the software in a horrible stack based language similar to PostScript which was compiled in one pass.
For a few years after I quit that company I'd get offers to work on systems written with it as clients were given the source and a copy of the compiler. I declined.
It probably was a preprocessor, generating RPG (or COBOL or C) from that source to feed to the compiler. Theoretically, it is also possible it was processed to MI (the AS/400 version of assembly) and passed to the QPRCRTPG API (which is exactly what all of IBM's OPM compilers also do/did), but I would be very surprised if they did that.
There is a YouTuber (CuriousMarc) who covers a lot of stuff like this. Its truly fascinating. I watched something a couple of days ago where they managed to jury rig something in order to read data off of an old 8" floppy - lots of geological data I seem to recall. Worth a look!
They definitely deserve a beer!
I used to work at a University and the space planning department had some ancient plotter that printed serial. My job was to replace their Windows 3.1 machine with new ones running NT4. imagine a 21" CRT monitor and how much that weighed. Amazingly, their CAD sw installed OK all we had to do was get the plotter to print room plans. I got a mate in the ops team to make up a new one and ti even worked.
I looked after a few walking dead type systems and did my best to pretend they were indestructible, never to be touched. So whenever it came to the time to finally shut them down the offending department always had to pay for a very expensive specialist to disinter the vital data. I usually insisted on them being hired for two weeks, gave them instructions and logins then went on holiday. It always pays to have someone to mind your office and let you know if anyone was trying it on.
And “Yes” I wrote all the cruddy old systems including the “export all, then die routines”, but having an office sitter was vital in our department (and paid well for old specialists who needed the work).
seeing as how you also have another MTBF stories
https://www.theregister.co.uk/2020/02/24/the_wristwatch_of_the_long_now/
you appear to have a theme this week :o)
anyhow, it brings me to the point, not much of one, but the reality is who IS actually and aggressively trying to sort out formats that will hold data for not just decades, but centuries, maybe even millennia ?
we already have some old formats that would not be possible to read or recover, and it isn't just down to the youth of the techs looking at this, who here can recall the most pirated software from way back when - Lotus 123 - and its office stablemate, AmiPro ?
does anyone still use them, are there any data stores around stuffed with old shit ?
you may well say that old isn't worth keeping, that business only needs to hold onto forms for X number of years, but what's going to happen when .mp3 dies off ?
who is ready to lose their .mp4 movies ?
the future may be bright, but it will also need to be curated properly, we are bound to have issues and some formats will die, but it's more the BIG players we need to concern ourselves with
FWIW - I still have some 3.5" floppies with my AmiPro docs on it showing how to maintain / repair several [back then] VITAL machines in a factory I used to work in, I started to collate a lot of data / reports and repair notes into digital format, not to mention the Lotus 123 spreadsheets for maintenance schedules
work management never understood why I was doing this, TBF, neither did I :o)
but it meant I could also keep my Lottery syndicate details up to date, with who paid, who owed etc
how do I open them now I plead
they were my life ffs :oP :o)
"how do I open them now I plead"
You could try:
www.documentliberation.org
Haven't used any of their software yet but might be worth a gander.
I haven't done anything more exciting than connecting an Amstrad luggable with terminal emulator software to a microVAX to capture screen dumps for documentation - a long time ago.
of the Software written for the ti990 was COBOL based (usually MCBA)
The OS was either DNOS or DX10 (Yes I programmed on those back in the 1980's)
The systems were pretty much bullet proof and if you write decent code, never was problematic - except seg limits were restricted to 64K
Converted one of the afore mentioned setups to a Motorolla 66 series UNIX box. The var had some sort of conversion tool that could read the
tape backup off the TI and lay it down on the UNIX box.
Trouble is, herculean efforts usually result in "Oh, nobody bothers with most of that stuff anyway". Personally I'd have printed out the data and told the client to hand enter the bits they really wanted. It probably would have turned out to be the last year's data and that's it.
Old PC running some sort of custom software (I think it was called SNIP?) where the program author wanted £££ for a "module" to export the data as CSV. Company in despair. Eventually I get to here about it.
Sure, I'll give it a crack, I say.
Weird ass file format that seems scrambled in a way to make pulling out the data to be difficult. I didn't have the time as they loaned me the box for a weekend.
But, possible solution. The software would spit out detailed reports to a printer, the company sometimes complained that it output too much and wasted paper.
Enter my trusty Acorn A5000. A bodged cable and some software written in BASIC, my A5000 became a printer, at least, as far as the old PC could see. But that's not all, I interpreted the printer codes, built up pages in memory, then scaped them to extract the data that was translated into CSV and sent via serial (Xmodem) to a real PC in the other room. All of this was happening real-time.
Set the old PC to print, set the W95 box to receive a file, and leave the A5000 in the middle to work the magic. Nice weekend hack.
Old PC running some sort of custom software (I think it was called SNIP?) where the program author wanted £££ for a "module" to export the data as CSV. Company in despair. Eventually I get to here about it.
My first IT job, we had a very kludged/customized install of Real World Accounting System. So customized it was impossible to move to a newer version. The company needed a summary report of open orders with totals of each. The system was only able to print full reports of EVERYTHING, or summaries that merely said customer, order date and ship date. Would have been thousands of dollars for yet another bit of kludged customization.
In the meantime I had been cajoling them to get me a copy of Paradox database, for another tracking DB (incoming orders, at the time handled on a convoluted set of handwritten and rewritten sheets). Eventually they grudgingly bought it. Poking through the flat-file export of the order headers, I happened to see it showed order totals. I cobbled together an import of the order headers, and then massaged the data for the very report they wanted. Saved them thousands of dollars on top of not turning the accounting system into an even bigger pile of shit.
but I did use a TI Professional Computer way back when.
That was my first experience with a machine with a hard drive rather than two 5.25" floppy drives. A whopping 5MB drive, and it was stunningly nice to work on with (as I recall) WordStar and Lotus 1-2-3. Ctrl K D anyone!
My brain hurts.
I used to do data entry on that system and help them do the backups on it back in 1998 for a semester in college. That looks like the one a the downtown office. It was a beast and the owner asked back then about networking them to keep inventory and sales in sync.
I did have one project like that .
We were planning to move from Data General Nova's to spiffing new Motorola Racks, however we could not work out how to transfer the code.
someone came up with the brilliant idea of exporting it via paper tape reader, then reading it in, on a paper tape reader connected to the target machine. Easy enough, except the host machine paper tape was fan-fold while the target was a reel. So we had someone output the paper tape, i had then every day drive 10 miles to the facility to feed in by hand the paper tape into the reader, hoping it would not tear in the process.
I did this for 4 days, then 5 days later the project got cancelled, so it was all in vain
You tell kids that nowadays, and they won't believe you
I had a blind friend with a speech-output computer - an XT-286. (I don't know how many people have told me there is no such thing, but there was.) There was a special board that handled it all, and a special word-processing program that fed it. The manual (and the company that produced it, the program, and the board) were lost to the mists of time. I decided the easiest thing to do was simply to keep the thing running, and with the occasional part replacement it did. Worked for about thirty years, and in fact outlived the friend (may he rest in peace).
Once upon a time I was handed the job of importing data from a system that processed telephone accounting records from a circa 1975 system that used circa 1955 GPO telephone equipment that had been installed in one of Betty's former possessions in Africa. They even used the good old GPO data entry methods, photographing banks of those little meters that clocked up your usage, and then entering the data by hand from the photos.
All this was just strings of numbers. Format? We don't need no steenking formats. The people in charge were too busy practising nepotism to bother with technical details anyway. So I cobbled together a little app using MFC (oh how I miss it!) that let me choose how to chop up the data into records and fields and display it so I could see the patterns.
And patterns there were. It was amazingly easy to home in on the record and field sizes. And so the data was imported and we sold a whole new set of systems for them to neglect. Or, as it turned out, set on fire because somebody's little schemes were under threat. Sic transit.
Not quite right on the history. The 990 started life as a business computer built for Ramada Inn as a reservation system. So large PWBs with a bunch of DIP integrated circuits. That computer was commercialized as the 990. Most of them were rack-mounted, had a row of 16 LEDs and a row of 16 buttons below them. So the box in the photo isn't any of the 990s I knew and loved - It might be one of the Ramada-vintage boxes.
The TMS9900 microprocessor was designed with the same instruction set as the 990 minicomputer, rather than vice-versa. It had a symmetrical 16-register file rather than an accumulator-based architecture, which was useful for the real-time applications I was building. A complete interrupt context switch was one instruction which internally performed 3 16-bit writes to memory. The down side was the register file was actually main memory words, so register operations were memory read/write - No cacheing in the earlier 990 models.
The later 990 models had an extended 20-bit address. One of those supported a development team of maybe 20 programmers on 20 timeshared terminals (not batch), building a collision-avoidance system for the US Federal Aviation Administration. At a time when the DEC, Data General, et.al. programmers were using glass TTY command lines, the 990 DX10 UI filled the 80/24 diaplay screens. Text, not pizels, but 2-D UIs and cursor moves in the text editor, with none of the vi bletchery.
The 99/4 and 99/4A home computers did use the TMS9900 with its 16-bit data bus externally narrowed to 8 bits. The TMS9980 with an 8-bit data bus was introduced later, but not used in the home computer. Another part, the TMS9945, if I remember correctly, was spec'd out but never made it to production - It included a UV-erasable program memory. The 20-bit address of the later 990 minicomputers was also spec'd out in the TMS99000 microprocessor, which I think did have some caching. I think some of those did get deployed. TI had a somewhat-successful commercial systems business. This was sold to Hewlett-Packard who apparently saw value in the customer list, but the computers themselves and the microprocessor line died out.
I wonder if that is a TI930 video terminal. If so, amazing it is still working after all these years. TI930 and TI911 terminals are not serial, current loop or anything like that. It's their own protocol.
In Chicago 10 years ago I met a guy who had been implementing an ASP.NET rewrite of a customer's insurance system that was running on a DNOS TI-990.
he said the 80-something year old programmer would still come into the office a couple of times a week, and they had a young guy that was able to keep all the kit running, because you can still repair those machines using off the shelf TTL logic.
At some time, I have been toying with the idea of using a serial connection to make a secured backup for syslog.
Log can be pushed on a centralized server, but that server needs to be attached to the network and if your network is compromised, the central log server could be too and the logs could be tempered with. The central log server could print every log line to a printer, but that creates a waste amount of paper and would be almost impossible to search through is you have to do some serious forensic.
So instead of attaching a serial printer, I was considering attaching a small machine to the log server, via serial line and that would record all that is being send. Of course that small machine would not be on any form of network. But the logs would be on disk and could be searched and analyzed automatically if needed.
It never get implemented.
Core memory? Seriously.
I have seen some kludges, but connecting an MCM based unit via a 286 server, serial port etc to a more modern machine, airgapped and firewalled.
How that antique was still working after all those years is literally a mystery.
They told me that "there were no plans to retire the system". IN 1997!!!!!!!!!
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
