OS/2 expert channeled a higher power to dispel digital doom vortex The end of the working week brings with it magical possibilities for fun and frolics, which is why The Register celebrates each Friday with a fresh incantation of On Call – the reader-contributed column that tells your tech support tales. This week, meet a reader we'll Regomize as "Rohan" who told us of an incident from the …
However, I do remember the UserFriendly cartoons, which have sadly all but disappeared from the interwebs..
A few still remain: https://nuless.org/comics/2016.08.03/User%20Friendly-2016.08.03.gif
I fondly remember the "Oracle DBA" cast as some sort of satanic monk, who would fix one's database in exchange for souls of the innocent.
Haven't really checked it out fully, but archive.org has a not-so easily findable dump of UserFriendly strips that appears to be complete (alternatively, WaybackMachine seems to have the entire site crawled).
Also, I'm 36 (coming in hot on 37) and very much old enough to have (correctly!) noob-level fiddled with autoexec.bat on DOS (and possibly config.sys, but I don't remember that). It was on a Pentium 3 PC with Windows ME[1], "famously" incompatible with DOS programs (about as incompatible as XP). There was a way/"hack" to mostly re-enable DOS, which I had done to try and play some DOS games that wouldn't run on ME, and I would fiddle with autoexec.bat (possibly config.sys) to load/unload TSRs at boot time/fiddle with the amount of free low mem.
[1]: It's probably the least hardware-compatible Windows version ever, and among the worst for software compatibility, but
a) With supported hardware it worked just as badly as 98/98SE
b) User Experience wise, it was still better than Vista/8[2]
and I will die on both hills.
[2]: An absolute shame, as both Vista and 8 brought actual, good internal improvements that "we all" misattribute to Windows 7/10.
I always thought archive.org and the Wayback Machine were one and the same thing.
Whatever, it's not rocket surgery, the first and easiest thing to do - if you remember the URL you are looking for - is simply to type
web.archive.org/<your remembered URL>
into your browser and if there is anything there at all, it should just pop up. For example, typing
web.archive.org/userfriendly.org
into Firefox caused the thing to load
http://web.archive.org/web/20220228234721/http://userfriendly.org/
automagically :-)
I read Userfriendly from start to finish while it was still going, and then again when it was repeated and was also disappointed when it finally disappeared from t'interwebs. Currently I'm enjoying - once in a blue moon when it is updated - Sequential Art.
M.
"Also, I'm 36 (coming in hot on 37) "
Sorry to hear that. "Coming in hot" usually means an aircraft coming in with unspent weapons, too much fuel, too fast, damaged or some other situation which implies a risky landing likely to end in an accident, carnage and maybe even death. Or did you mean it some other way? (I hope so!)
I meant it as "too fast" (as in somewhat soon-ish, I still have nearly two months left). I will admit I mostly heard the phrase from the FSX era of AirForceProud95 :), usually in videos where he was playing ATC, meaning either "airplane is coming too fast" or that the pilot seemed to be operating in "I'm landing now, good luck everybody else!" mode.
Yes. For OS/2 2.0/2.1 or earlier a boot floppy is required. For 3.0 you wait for the OS/2 blob in the top left, press alt F1 then C, for 4.0 it's alt F1 then F1.
It boots to a command line with no networking, TEDIT can be used to edit files (for OS/2 2.x you'll have to provide your own text mode editor).
It's not a single user mode, because OS/2 is inherently a single user OS (there is a sort of multi user add on, but it should only be used in very specific circumstances)
I’m fairly sure this was standard practice for most folk - either remembering stuff or knowing the phone number of someone who could!
I remember talking one customer through doing something over the phone, and when they realised I was travelling they wondered how i was driving while using a laptop - and I explained it here was no laptop, it was all in my head. There was a moment or two of silence, and then a question: “so how come you can remember all this, but can’t tell me what day of the week it is?”. Although now I’m older, I’m useless at days of the week, where my car keys are (last time was the pan drawer) and the computer stuff…
So far, I can still remember my own name though, so it’s not all bad ;)
Same here. I got so well acquainted with a specific DOS-based POS* that I could tell customers what exactly to do over the phone.
This also helped to restore some deleted data as well.
*POS in question was Ulti-Sales.
At that time it was all character-based, being DOS...
Digital had a line of cut down PDP-11 type workstations called DEC Professional. I was issued with a PRO-350. As I recall, the OS was a version of RSX-11 called "Professional Operating System" or P/OS - Most of us thought that it was. They lasted less than 2 years before they were replaced by genuine IBM ATs.
Back in the dying days of the last century, I was trying to fix a very broken industrial x-ray machine, with management breathing down my neck to get it up and running again for production. By this time, the machine had been extensively modified by the manufacturer and was distinctly non standard. No joy, so called the manufacturers. Got through to the designer of the machine in Germany barrelling down an autobahn at some insane speed, who gave me a whole series of "check the voltage on the blue with black strip wire going to pin 22 of connector C42 is between 1.7 and 1.9V type instructions for half an hour, entirely from memory and a second language. Much impressed!
barrelling down an autobahn at some insane speed ... entirely from memory and a second language. Much impressed!
The first two I can believe, the third on top of the first two requires sufficient command of that second language to be classified as completely bi-lingual, pretty rare for Germans. If true, I am extremely impressed.
Uhm... me!
Though friends and family in Germany now state that I've been gone so long that my German is getting to be like a mistake-laden, second language behind English.
Generally takes me a week or so, when visiting, to get back in the groove. But I still read it on a daily basis and can generally insta-context switch.
Generally takes me a week or so, when visiting, to get back in the groove. But I still read it on a daily basis and can generally insta-context switch.
It takes me about the same amount of time too for German and insta-context switching takes about the same time. I am for all practical purposes completely bilingual Dutch-English, but I don't have enough opportunity to keep fluent in German.
I do reasonably well in French, and have some basic German. I was once in German-speaking Switzerland, and ended up explaining to a French tourist that he was not mistaken, and that the fries/chips at a take-away really WERE that expensive. Neither of us made a purchase.
insta-context switch
In the days when I used to travel a great deal, I found it took a certain amount of time on arrival at the next airport before I could decode the sounds coming out of the tannoy into a recognisable language - and knowing where I was didn't help. Now, when I spend a lot of time in one European country, I find if I've been watching British TV or listening to Radio 4 and then go out, I'm tongue-tied for 5-10 minutes until my brain catches up with the change of context. I knew someone who was a simultaneous translator and seemed perfectly capable of maintaining both contexts at the same time. I guess I've just got a single-threaded brain.
Bilingual Germans are by no means rare. I know several including my best friend, his sister I used to go to school with, and their mother.
However I used to be in a relationship with that much rarer thing - a bilingual English person! She had lived in Hanover for several years and unlike probably most Brit ex-pats she was completely fluent. I was busy learning German at the time and could tell that much of the time she actually thought in German, so I have no problem believing that little anecdote.
Mine’s the one with pockets bulging with 1990s Deutsch Plus VHS tapes and Berlitz CDs...
And correctly identify a pen, a wristwatch and a notepad, remember the name of the Prime Minister, draw two intersecting pentagons, count backwards by seven from one hundred to sixty five, read an instruction aloud from a piece of paper then carry it out, fold the piece of paper in half and put it on the floor, then remember the the items you were asked to identify at the start.
Back before the Internet, I had one of these things... came with blank pages of paper, which I would fill up with esoteric knowledge, notes, things to remember in the future, entire snippets of code.Pages and pages of my scribblings, from sendmail configuration to recovery procedures for SGI Indigo workstations to creating a static C library for when these fancy new "dynamic" libraries (they'll never take off) didn't link.
What were they called again? Notebooks? Can't be, that was a kind of laptop!
Way back, early in my career in the pre internet days, I was told to look something up in the timsboard. When I looked puzzled I was directed to a lever arch file.
It turns out that a now departed employee named Tim had originally kept all sorts of useful notes on a notice board. Once that was full the data moved to a folder but the name stayed as a piece of in-house jargon.
Agreed, it's not that different to knowing some standard formulae.
Fast forward to today and we have the Windows registry and systemd: now get out of that!
As an aside, it's easy to make fun of the way IBM handled OS/2 but we should not ignore how well engineered, documented and supported it was. Indeed it kept millions of cash machines, aiports and other systems running smoothly up until about a decade ago.
<IBM ... well documented</i>
I've always found the IBM documentation to be top notch, learned a lot from it. When I started as a programmer (a long time ago) on my first job, I would have some of the IBM documentation on my desk to leave through before starting the workday, during breaks and at the end of the workday before catching the bus.
>> > "you...can’t tell me what day of the week it is?”
>> What the hell are you on about? [looks at phone] "It's Friday, ya moron..."
These days most people cannot remember phone numbers and would be lost without the phone book on their phone. We need to go back to dialling out the numbers!
Came here to say this used to be me.
I worked at an ISP and we only had one OS/2 reference computer. We got it for a few days to learn then it was sent off for a manual to be made. Then someone broke the OS2 computer before they made the manual (dropped it down stairs)
I solved quite a few OS/2 calls while staring at a white wall and pointing at where I remembered things being. OS2 had a very specific UI and was well organized, plus there were very few things that went wrong, so it's not like I had an eidetic memory, but the other techs nicknamed me "Mentat" from then on.
"There is a sequence I learned"
I have a couple burned into my memory : One is 'RC, CA, CL' on a System-X PU CUC terminal. The other is '6031769' which was the 'cheat code' for Manic Miner on the Sinclair Spectrum. ISTR this was the phone number of the game's author, Matthew Smith. 1st one earned me money, 2nd one gave me fun:-)
It's a sign of the times.
Back in the day we used to have people who really knew what they were doing. People just starting their careers were actively encouraged to learn. I worked with amazing people (one was even asked to produce content for the Open University). With the advent of outsourcing, that all seems to have taken a back seat. The gaggle of wizards is growing smaller, and that's a real shame.
Question - do you think the proliferation in AI tools like ChatGPT, Bard etc will lead to people becoming more dumber, incapable of retaining memories of certain processes?
Back in those days you absolutely had to memorize specific steps in order to be able to do your job.
Nowadays it is just as easy as opening a web browser, go to google, chatgpt or bard (or whatever) and asking for assistance, and most of the time you'll have the answer.
Certainly more dumb I should think. :-)
It's already happened with phone numbers, I have no idea what my son, daughter's, or fathers mobile phone numbers are because I never learned them just added them straight to the phone memory.
I can remember my grandmothers landline number from childhood, the landline number I used for a BBS, my wife's mobile, all the landline numbers my mother ever had and the landlines I've ever had but that's about it. Oh, and my sisters mobile number before she emigrated.
I've also done the 'list out commands from memory over the phone' gig.
Used to work for a bank and did on-call support but it was reasonable in that they provided a mobile dongle along with the laptop so you didn't have to rush to somewhere with wi-fi which wasn't ubiquitous back then or go home or to office, or rely on a hotel network when travelling. Important fact here was that the dongle was used very infrequently.
Had gone out with the family to a National Trust place a good distance from home, which turned out didn't have very good mobile signal. Got a call saying we have an issue with some workflow in Exchange public folders, don't ask, we need fixed soonest or massive penalties, but no data signal.
Move to a location with some signal and try to dial-in, nothing doing.
Ok, drive about ten miles leaving family behind to find a good signal. Still nothing doing.
In desperation called one of my managers who was an AD guy not Exchange to see if they happened to be at home and could dial-in, while I'm sitting in a car on a 30C day, couldn't run AC so he could hear me, and couldn't open window because it was a busy road that I'd ended up on and SLA was creeping up.
Managed to talk him all the way through checking and fixing it though so was pretty pleased with myself.
Turns out accounting had killed the mobile contract because it hadn't been used in the last 3 months, there was no tag to say it was infrequently used.
But of course they hadn't said to me that they were doing so even though my name WAS tagged against it so they could bill my department.
Thanks so much. Words were exchanged shall we say.
I can remember my grandmothers landline number from childhood, the landline number I used for a BBS, my wife's mobile, all the landline numbers my mother ever had and the landlines I've ever had but that's about it. Oh, and my sisters mobile number before she emigrated.
Funnily enough met up with some old school friends last weekend and we could all remember each others phone numbers from 40 odd years ago. Haven't a clue what my Dad's mobile is now, although the landland hasn't changed for about 50 years, apart from an extra digit or two added to the start.
"I can remember my grandmothers landline number from childhood, the landline number I used for a BBS, my wife's mobile, all the landline numbers my mother ever had and the landlines I've ever had but that's about it. Oh, and my sisters mobile number before she emigrated."
I can't even remember my own old telephone numbers, personal or work and never could - once I move house or job I seem to discard them. There's one exception: my parents' old number. That's because it's now our own. As to my own mobile number, it's not so much muscle memory as a kind of rhythm memory; if I tell it to somebody they read it back to check with a different pattern, e.g. "Oh double seven one oh - ..." becomes "Oh seven - seven one - oh ... " it's gone and I'll not get it back for several minutes.
Same with car numbers. My first two I've still got. I've also still got the last one because I only changed car in April. As to the rest I can only remember the area letters of one of my N Irish ones and that because I was driving down a ramp in an English multi-story (the ramp included a hazardous footway) and as I passed someone walking up I heard an incredulous "EOI?"
Actually, I would suspect the proliferation of AI would drive people to holding the processes in memory. Because AI is crap and prone to hallucinating bollocks answers.
As such, as AI proliferates, more sites will be filled with incorrect crap, so people will be forced to remember how to make things work, rather than relying on google searches.
It's a shame really, I mean I have limited mental capacity, every time I learn something new it pushes something old out of my brain. Like that time I took that home wine making course, and afterwards forgot how to drive?! :P
(This joke was stolen from Homer, of course... Just channeling all those AI's out there...)
Yes, you had to learn. There was no web browser. You talked to people. You experimented. You found ways to IMPROVE what was being done. What I see now is people blindly following whatever they are spoon fed. On more than one occasion not even bothering to stop when there were errors. The question then becomes, what happens if you look for an answer (as you say yourself - most of the time you'll have the answer) but you *don't* find one? Or the one you do find is incorrect?
I'd rather have someone who has the skill and drive to go and work out the answer....
"Nowadays it is just as easy as opening a web browser, go to google, chatgpt or bard (or whatever) and asking for assistance, and most of the time you'll have the answer." I wouldn't trust ChatGPT to provide an answer that involved several steps and trying to use google search to confirm it, would probably be a self fulfilling prophecy whether correct or not. I'm reminded of: "Just because it's in a newspaper doesn't necessarily mean it is wrong!"
I generally don't trust it, nor do I use it. However, one of the reason why some people do trust it is that, if it's something that people asked often enough while it was being trained, it will generally get it correct. So if they're trying to figure out something they should be easily able to remember anyway, then the answer they get is likely correct. This will work until they ask something that's actually tricky or is related to something small or new. Then they'll get equally plausible-looking instructions that probably don't work.
I think I read somewhere that this has been demonstrated. ChatGPT specifically I think. It's prioritised to give an answer, so when it doesn't know, and can't get an answer quickly enough, that is when it's most likely to "hallucinate" (or lie as we would call it in everyday parlance). Ask it the same question some while later and it's far more likely to come back with a correct (or at least more correct) answer. It's as if the "reward system" is set to mark down significantly if the current answer it has is "I don't know". I wonder if that's a (work) cultural thing in the US where most of these AIs seem to originate?
It's also rather reminiscent of the HAL 9000 in 2001: A Space Odyssey where HAL has been taught/programmed to be fully open and honest and then has to keep a part of the mission secret from the crew and goes "insane" over the conflict.
The best analogy for how to trust ChatGPT is to use it like an intern. Sure, you can give it stuff to do. Sure, it will get the stuff done. But, you're going to check every word it writes because it's an intern and it will make mistakes, and you'll still own the mistakes. Over time, you may pay a little bit less attention as trust is built, but you'll still always check. It's an intern, after all.
I don't think anyone knowledgeable thinks there's any intelligence in AI (yet).
But ChatGPT is definitely a very useful tool in some office and programming contexts. We're using it frequently for first drafts of large volumes of code that would otherwise take hours to write...or somewhat less onerously copy/paste from Stack Overflow :-)
> first drafts of large volumes of code that would otherwise take hours to write...or somewhat less onerously copy/paste from Stack Overflow
But just like Stack Overflow it (and I was going to say could but...) will contain some absolute garbage so you are still going to have to go over it line by line , or are you paid by the line and thus always produce huge bloated redundant code ?
But ChatGPT is definitely a very useful tool in some office and programming contexts. We're using it frequently for first drafts of large volumes of code that would otherwise take hours to write...or somewhat less onerously copy/paste from Stack Overflow :-)
You really need to stop shovelling shit and get out of this industry.
I think the main problem here is that the world around us changes at such a rate that it's no longer feasible to become a wizard. By the time you earned the pointy hat the world has moved on so far that your knowledge is useless. I used to know Turbo Pascal (and even 8086 assembly) at intermediate magician level. What good does any of that do me now?
Now that I think about it: YouTube::squatch253. Toby there made the same remark a while ago about the rate at which car engine designs change. It's practically impossible to keep up.
>> Back in the day we used to have people who really knew what they were doing. People just starting their careers were actively encouraged to learn
Having been around since just before the dawn of the PC era I can relate to that, in those days everything was new, PC cards needed IRQ's, DMA and all sorts of stuff sorting so we got to know how a PC worked intimately. And not just it physical stuff, there was networking and the software, so were learnt it all.
These days newcomers to IT seem to be lost if there is not a simple GUI screen with a "click here" prompt.
of what that must have appeared like to onlookers. :)
When I think of the number of times I was consulted by staff and students on issues outside of my actual responsibilities (and far from the expected competencies of the role) and have responded with "that rings a bell" and motionlessly, silently scouring my memory for something I had read or seen then responding with an answer or "you didn't change...?" I was fortunate not to receive a visit from the Inquisition*.
* of course no one expects the (Spanish) Inquisition
Use to be the Guru on a line of aerospace routers.
The way I use to work out the code interactions was rather like a 3D map, with villages on the hills representing the protocols or processing blocks and the roads between carrying the data streams.
Apparently I would sit there with my eyes closed, but my hands and sometimes head 'flying' through the landscape over the data packets' travel.
I had a map of around 400 'villages', and yet struggled with the names of some people I'd worked along side for 20 years!
Icon: black helicopter as that was probably what I was 'flying'
Many moons ago, whilst onsite at a customer setting up some new software on all systems I was requested to update a few packages and windows service packs on the Finance Directors Laptop as on site IT had had some non disclosed issues.
The FD was last, turns out the FD was French and his Windows was setup as French, this was early NT 4 days, you know when the locale, and keyboard layout was US English on the login screen despite system being set to French or UK English or whatever.
So I rattled on installing patched, getting the system up todate, then installed the company wide softtware, then installed the additional things requested all without issue.
The FD had been watching for sometime, and during the process started to speak to me in French, I replied I'm sporry I dont speaak French, He said "oh I thought you did as you understood everything that came up on screen".
Nope just know what each box meant by looking at it and what each button was due to location, to be honest I hadnt even noticed it as in french.
If the UI is consistent, that works even for (some) non-Latin cnaracter sets. Then you end up installing in Arabic or Hebrew and have to reverse all the instructions so that you Accept rather than Cancel - at which point it's almost easier to swap hands and do it with your opposite hand :)
Same here. Many years ago I had to install a Finnish version of PageMaker. I was used to installing the English version and muscle memory saved the day. Our customer was impressed that I knew his language. We had a good laugh about it later when I told him I didn't have clue what the Finnish instructions were telling me to do.
Stood in the middle of Tesco on the phone walking a friend in India through a particularly nasty issue to do with internal security certificates on a WebSphere Application server NDM deployment, visualising the console GUI and even moving my fingers as if they were the curser to keep my head "in the zone". yep, you do get stared at and no I didn't care as I was on decent £££ for it ( that 20min call paid for the weekly shop and a nice bottle of whiskey to celebrate )
And where it gets frustrating:
Me: I want you to type c d, then space, then slash t m p.
Them: Okay.
Me: Did it do anything?
Them: No, it didn't.
Me: Do you have a new line at the bottom of your window?
Them: Yes, it's different.
Me: Type pwd and tell me what it says.
Them: It doesn't say anything.
Me: What does the bottom line of your window say?
Them: It says user@macbook:~% cd/tmppwd
Yes, part of that is on me for not being explicit about pressing enter. However, it makes me dubious about dictating commands that could do any damage. Okay, just execute the command df -h and tell me what disks you have before we repartition one of them. No, I think I want to be doing that one myself.
If I had my way, I'd prefer to do any repartitioning myself. However, if that is not an option and I really have to choose between talking someone through a repartition task using CLI or GUI tools, I'm going with the GUI. This is despite the fact that if I was doing it myself, I'd probably use the CLI tools.
The GUI has more convenience functions for slightly less tech-literate people. Instead of device identifiers, you get disk model numbers and volume names. That removes one step, the one where I try to get them to properly associate a name with an identifier and can only guess whether they've done it right. There are a few more sanity tests built into many GUI tools. This may identify a mistake before its made, and even if it doesn't, getting some warning messages might cause the user to be more deliberate than the no warning messages generated by the CLI tools (yes, there is one but I'm not sure they'll remember that fdisk told them to "Be careful before using the write command." at the beginning of the process).
Maybe this is due to bad experiences I've had reading things out. Only recently, I was reading a password over the phone and had to stop the user several times from capitalizing random letters for no good reason. I've gotten someone else to execute an rm * by reciting a command, though I was pretty clear that there was not supposed to be a space between the * and the rest of the pattern. After that occasion, I started being very specific about whitespace, which makes reciting commands slower "d d, space, i f equals sign slash d e v--no, there is not supposed to be a space between the i and the f, I spelled it out because it doesn't actually mean if. Hang on what does your command look like right now? Let's start over.". I don't know whether this is due to a failing on my part or just general confusion that everyone will encounter from time to time, but I tend to provide written instructions first and talk them through executing them rather than speaking the commands.
"if that is not an option and I really have to choose between talking someone through a repartition task using CLI or GUI tools, I'm going with the GUI."
I generally split the difference and have 'em fire up cfdisk ... but then I haven't supported Windows in almost a decade and a half.
In my example I posited a Mac, and when I managed to confuse someone into executing rm * it was Linux, but it makes very little difference to me what operating system I'm talking them through. Either way, there's a chance that they could badly misunderstand a CLI command if they're not familiar with the CLI on their own, and as the rm * guy should have been familiar enough not to run that, I'm cautious about that as well.
"I don't know whether this is due to a failing on my part or just general confusion that everyone will encounter from time to time,"
Back in the mists of time I did an amount of "teacher training" so I could be an instructor at work. One of the exercises was to get two sets of wooden blocks of various shapes for two people. One builds something out of the blocks behind a screen and then tells the other person what to do to build the same item/shape/tower. It's remarkably difficult the first time you try it. Most people make far, far too many assumptions when passing on detailed instructions. It's a lesson I have never forgotten and am frequently frustrated by when following other peoples instructions, especially in an area I may not be familiar with. Project Managers take note!!!
I have a radio amateur friend who is dyslexic - cue *long* phone calls with me dictating Linux commands down the 'phone in NATO phonetic alphabet with occasional "press enter now". Then him reading back to me what he had on screen, only for me to discover a mistyped command. Lather, rinse, repeat.
I now have a work colleague who sits beside me who is bright but can't type and relies on me as the Linux guru - same system, with fewer phonetics ...
"Yes, part of that is on me for not being explicit about pressing enter."
Even that can be a challenge. Does the key say "Enter" on it? Or maybe it says "Return". Or just has the right angle arrow symbol that is used for "Return" but may not be called that by the person you are speaking to :-)
In particular, laptops. Remember when we all still had proper desktop keyboards with a number pad and we had both a "Return" key on the main alpha-numeric part and an "Enter" key on the number pad? I can remember giving people instruction to press "Enter" or "Return" as most people used the terms interchangeably, but there were always some who would look for the specific key you mentioned, as if it mattered which one they pressed. Probably the same people who got confused looking for the "Any" key :-)
Back in the day I worked on a lot of T-carrier stuff. I can't tell you how many times an owner/client ranted about a shiny new (fractional) T1/E1 link being down, how the equipment was shit, the field guys were incompetent, and how pretty much everybody involved with the installation should be taken out behind the barn & horsewhipped. Most of the time[0], it was an incorrectly set loopback switch on the new node. Seems bosses in general can't resist flipping switches ... and can't read blinkenlights.
Sometimes I'd casually reached out and toggle the loopback switch, thus fixing the link and painting the boss's face an interesting shade of red when I presented him with the bill reading nothing more than "Call out. Flipped loopback switch. $1,000" on an official invoice.
But once in a while, after inspecting the node, I'd stand aside & motion the boss through the door before me. While he had his back to me, I'd flip the switch ... and we'd go off to his office for a chat about fixing the obviously broken machine. I'd let him rant on for several minutes, around 20 was the record, but always ending up with something along the lines of "so what are you going to do about it, then?". To which I would quietly reply "Oh, I've already fixed it. We'll invoice you for the call out". Sometimes the resulting sputtering reached epic proportions ...
[0] The rest of the time it was a cable that had fallen out of the CSU/DSU because it hadn't been screwed down properly. We always took the blame for that, even if it was their guys bolting stuff together. We've all done it, we're only human, I'll take the blame, no charge ... sometimes it's handy to have a friendly couple of faces in a client's datacenter who probably won't ever try to throw you under a bus.
I was on a training course and we were given a test. I sat in front of the test unit, set up the tester and ran a test... nothing wrong... reran the test... still found nothing wrong!
Turned to the instructor and said I couldn't find a problem. He said that I'd found it... and pointed to the switch I'd automatically flipped without thinking because I saw it was in the wrong position.
I used to work with OS/2, from V1,0. No Internet of course, just a huge and heavy set of printed books which were so awkward you just had to keep it all in your head. So now I could still just about remember how to write an OS/2 and (PM) program in C off the top of my head, but struggle to remember people's names. But I totally agree that sometimes you just have to stop for a few moments to think and it comes back. I once worked with someone who would go for a shower when he had a problem as this was the best place for him to think. Sometimes several times a day.
My moments of greatest clarity were cycling into work, going over whatever problem might be most prominent for the day ahead. An hour and a bit cycling on quiet near-deserted country lanes was a great source of inspiration and often led to a resolution being ready in my head by the time the office loomed into view. Riding home was never as productive - maybe it was end of day so work is over till tomorrow, with the morning ride in being the opposite.
IME it's not so much the thinking as the stopping that matters. Once what's directly in front of you is removed the solution seems to emerge from the subconscious. Leaving work in the evenings was usually the occasion. On one occasion it took no longer than the walk across the car park but driving and, therefore, having to think about something else was more usual.
Many years ago, when SQL Server 6 was US and Canada only and we were definitely not supposed to be developing on that version in the UK, we found an issue with the database that we were creating. We called Microsoft and after going through several layers of tech support they decided that it was beyond their ability to fix over the phone (this was many years ago, after all) so they would despatch a human being to investigate.
The next day a giant of a bearded man with poor personal hygiene arrived from Seattle to stare at our database and our code. After three days of staring, occasionally typing and even less occasionally grunting or asking for coffee, he said, "Got it, I'll send a fix". He flew back to Seattle and a fix was emailed to us a day later. Our stinking, silent, staring guru had fixed things and all was well. We went live before the UK & Europe embargo expired.
I've got no real stories of when I (or a friend) channeled a higher power to get something to work, but one thing I've found does seem to work, but only if the device is near a window, is this. When it fails, simply point out to it that it is very near a window (it helps if the window is high up) and that it will being going through the window quickly very soon unless it behaves.
You'd be surpised at how many devices start working when you do this.
Back in the Win9x days I worked for $SchoolName as an IT Tech; and a large part of my work involved fixing hardware that the students had sabotaged.
There were the usual culprits of a 120/240v power supply flicked to the "wrong setting" (which caused a loud bang and lots of magic smoke the NEXT time the device was switched on); and that time I unscrewed the security bolts of a screwed-down-and-cable-tied Desktop to find an open and half-empty Coke Tin standing upright balanced on the top of the CPU heatsink.
But the most "fun" part was upgrade time. Each summer we had a little bit of petty cash left that was used to fund some new RAM; which got divided up amongst the most-struggling machines. Because their RAM sticks were mismatched and their Motherboards were old and full of caked-in dust and their Operating System was more ornery than a bag of badgers; this resulted in a lot of BSODs. So quite often a "sacrifice" was required.
Gather all the misbehaving machines into the same classroom and arrange them in a circle. Then select the most obstinate serial-crasher; move it into the middle of the room; and BEAT THE EVER LOVING STUFFING OUT OF IT AS A WARNING TO THE OTHERS.
More often than not, this resulted in the rest of the machines settling down and remaining comparatively well-behaved throughout the upcoming new term.
Pre-emptive Empathetic Percussive Maintenance. A completely ridiculous and unscientific troubleshooting method I know; but very cathartic and surprisingly effective nonetheless.
I'm thinking of using something similar as a training mechanism for the new AI that $CurrentEmployer are foisting upon us...
It's standard practice to set it for an OS/2 install. Other options include SET RESTARTOBJECTS=STARTUPFOLDERSONLY,REBOOTONLY (to run startup folder items only on a cold boot) or RESTARTOBJECTS=NO to restart nothing.
Other typical CONFIG.SYS edits include :
For HPFS only systems, editing DISKCACHE to a minimum value. No sense cacheing FAT when it's not used.
Set MAXWAIT to 1 - more granular multitasking
increase THREADS to 256. A modern ArcaOS system has bumped this to 1024 as a default. A sample, not stunningly loaded Windows system here is currently running over 4,800 threads!
Of course these days systems have so much memory and are orders of magnitude faster than 90s systems, so it's less necessary to fiddle around to gain small amounts of memory and performance.
Other typical CONFIG.SYS edits include :
PRIORITY=ABSOLUTE
One of the first things I'd set after installation was that, especially for something that was running background server processes - else, the scheduler would give whatever the foreground UI process significant cpu time - enough to have noticable effects on the performance of anything that is not the active UI process.
I did manage to impart that knowledge to a group of seasoned IBMers (though, not as it turned out, OS/2 experts) gathered around a client demonstrating the performance issue of a retail application - typical installation was 2 worktations plus a server in the office which woud double as an extra workstation in the office - problem was, if there was an active foreground application on that server...
Once we set that on the server, followed by a reboot, it was no longer a problem
Well, yes, dynamic priority is generally more appropriate for a desktop system - you want the program you're interacting with to be responsive. If you're severely constrained on resource you can set absolute priority but the danger is low priority threads never get scheduled due to a lack of a priority bump.
In general it's better to fiddle around with MAXWAIT than set absolute priority. It's not a setting I think we ever set, the 486s (or better) we were using had sufficient resource to run heavy processes.
The product we were supporting on OS/2 did support decoupling the server from its admin console as a separate product, but the majority of customers used the interface on the server itself, and it wasn't really an issue.
Memory is a bit foggy on this but I recall the absolute priority helping with multimedia playback and other things.
Loved OS/2's performance with dual PentiumPro CPUs(BP6 mobo) and the SMP kernel with SCSI disks.
WPS had Workspace folders which would relaunch all open apps/doc when you last closed the root folder.
Such great usability built into that WPS system.
As a Notes/Domino person (plus some other email systems), I've done a few of these,. Other colleagues, seriously experienced people, looked at me like I'd performed some sort of Dumbledore level wizardry, but it was just basically knowing stuff after a couple of decades of using the software.
We took on a guy who claimed to be an 'expert' in finite-element analysis. A likeable chap, at interview, he was hopeless, constantly wandering off-topic. Even his PhD paper was simplistic and implied a plodder rather than an investigator. Some of the projects he had worked on were obscure, but interesting in their breadth of application.
However, he knew his FE and CFD and he desperately needed employment. We needed help on a knotty problem at the boundaries of knowledge, so we took a chance and offered the job on a six-month trial. On arrival he immersed himself in our particular technical issue. He had periods of 'deep thought' which, to an outsider, indicated a lack of attention. He would mutter as he went along. His head 'wandered' mentally and physically. He asked for a better laptop, at eye-watering expense which we bought...... more, better (expensive) software, which we bought.... He took the laptop and problem home in the evening to work in his own time, including weekends. We suspect (know) beer provided inspiration.
In a moment of sheer genius, he identified the problem and outlined a possible solution. An original thinker.
This lasted until a new boss found he could not deal with people who thought out-of-the-box. And sacked him.
I'm still in touch with him, waiting for him to come up with an idea that is so far out of the field that it's bound to be a success.
A company I worked for in the late 1980s made systems on DG Minis--16 and 32-bit Novas and Eclipses. A hardware guru there was sent up to Long Island, where a recently-sold system wasn't delivering any output. He pulled the driver board, sat down with the schematics and began to inspect it. Eventually the owners, man and wife, went upstairs to bed. Until they fell asleep they could hear him mumbling to himself. Somewhere in the early hours, he took an X-Acto knife, made a cut to interrupt on connection, and the output board then delivered.
I never met the guy. Somebody who had reported to him described him as the smartest guy he had ever met. But somewhere between the Long Island wizardry and my arrival at the company he had slipped into a depression (or so I infer) and did nothing for quite some time. The company fired him, needing somebody to do something, even if without the same wizardry.
"(But then I always ask myself: what would you want to do with all that old crap?)"
I suppose it's a bit like asking a mountain climber why s/he climbs mountains. Or why any athlete does whatever it is they do. There's no real practical purpose other than "it's fun" :-)
(And those who are good at whatever "fun" they do might even make a living at it because other people like to watch them having "fun"). Although I suppose with something like retro computers, it's a bit like being a museum conservator trying to keep history alive. We're already at the point where in some cases emulation[*] is now the only option as custom chips in particular die and can't be replaced because no one has the skills or resources to reverse engineer and/or decode PALs/GALs etc.
* If you've watched Adrien's Digital Basement and/or similar channels, you'll have seen there are emulations of some custom chips using microcontrollers on a tiny boards to replace the unobtanium custom chips.
I remember a college friend going to work at Plessey Swindon, now he came from a practical (TV repair) enviroment & was good at fault finding in that manner.
One of his colleagues, would take a few Voltage readings, then walk away to identify the fault using the circuit diagram & applying Kirchhoff's current law's.
A different take on hardware, a few years later...
The founder of Novatech in Portsmouth had a mail order business selling parts for Vauxhall Novas - which then progressed to PC components...
Advert for original Novatech...
The company today...
https://www.novatech.co.uk/about.html
icon: see above link
For me it was talking a colleague through X.25 tracing and telling himw what was wrong once he got told me what he was seeing in the trace.... from the Glacier d'Argentière Moraine ( with the harness, ice shoes, helmet and more still on ) after comming back from a mountain climb. ( Chamonix valley, French Alps )
He never really understood how I was doing that all from memory ( since I obviously didn't have any protocol description paper at hand, nor a laptop [ at that time they were way too rare and heavy to risk one climbing ] ).
Technically I was in vacation, but it appeared that there's no such things as vacation for wizards.
At least I got a nice pint once back at work out of that ( thus the icon )
I once upon a time, set up for my old man this simple proxy network program with fixed IP adresses, so he could share the dial-up on a machine accross all of them along with a printer. Just something to get this tiny network going without floppies going back and forth, in Windows 95/98.
This network used BNC connectors and thin coaxial cables, so no routers. To him, it was wizardry.
It's was a very deliberate design. The idea is that you shutdown, all programs save their data, and when you reboot it's returned to the same state as before. There were also 'work folders' where a set of applications/files ran from a specific folder and did the same thing (close folder, all applications save data and close), as the intention was that OS/2 was designed on a data first basis rather than application first (you create the data file, applications that can handle that data type then load it).
Problem was, not enough applications were designed to handle that properly, and it does rather rely on the applications working and not overloading the system. It was a reasonable idea back when the industry was still trying to work out effective user interfaces.
Right, there was a 'workspace' option in the folder settings which triggered that activity.
To this day I have not run across an UI which comes even close to the usability of the WorkPlaceShell.
When I was at university I'd do a lot of coding in the evening in the library. I'd sit there drumming my fingers while thinking about the code for while, and then start. Now I don't think I was fully entranced during the coding but I would only become fully conscious and aware of my surroundings until I was done, the latest was when a security guard shook me because he wanted to lock up at 3AM.
A friend of mine did a lot of low level programming, and one day realised it had got a bit much when he found himself counting his change in hexadecimal.
My brain doesn't work like that anymore, or my fingers for that matter.
Been there and done that. 14 hours of solid Z80 assembler coding. Apparently, never left my seat the whole time. I thought it was about lunch time when I finished, but it was already dark. Worst of all my desk was by a huge window and I never even noticed it getting dark, or the need for bodily functions (input or output.)
In my early PFY years, I once had our external consultant (who had previously told me a gazillion times to defrag my Exchange dbs on a regular basis to recover the space we were always in short supply of) spell for me, afterhours, on his personal mobile phone, with no Internet access of any kind, an endless Eseutil string, character by character - switches, drives, options and al. Which saved my bacon.
Later on this came up in a conversation, and his comment was "...Oh, I DO remember that time, we were at the restaurant with my wife for our wedding anniversary..."
Dave, if you're reading this, thanks again.
Are you new here Jake? What have you done with the old Jake? :-)
"Consultants" being brought in in the UK almost always means an MBA type as per the many stories and recounting of tales in both articles and the comments section. Technical consultants who actually know what they are doing don't get mentioned much because there's no "story" when things go well :-)
there be dragons.... well actually the likes of me (you'd be safer with the dragons)
In years gone by, we had a mangler who believed that us CNC guys did nothing apart from press buttons then stand around waiting for the next chance to press buttons.
Then he comes out with a panic, the geometry on the compression rings we were doing was wrong and hes all in a "It will take hours to sort out with the computers and new programming and QA checks" kind of panic. then there was a loud clang as his jaw hit the floor as we turned around and said "Already done... just had a calculate a new toolpath with the new angles.... its running now and QA will pass it"
Mind you its always been the way with the knowledgable types... I bet back in the day of forging new swords that the king would go "Surely the gods made such a blade as this" while his cheif armorer would be standing there thinking "what a wanker, took me and the apprentice 3 days....and 2 of those were sitting the inn leering at the wenches..."
Incantations of the contents of TNSNAMES.ORA should be familiar to many el Reg readers.
I am reminded of one in particular where, I deliberately held onto said file separately from my ancient NT4 'terminal'. In the subsequent "upgrade" to XP, said file had been "rationalised" by to only refer to servers they were able to ping. Who ever said PING was a qualifying test for whether a connection can be made or not...
Needless to say, anyone who had dependencies on the legacy connections described in the original TNSNAMES, had them broken.
Now, backup/copy/paste is hardly rocket science but when a significant safety function of your operation depends on databases that were relying on that connectivity - enough to cause thinking about a roll back from XP to NT on thousands of systems - that's quite the save.
In other, not so intelligent uses of the command line, this el Reg hack was caught out once by issuing deltree *.*. Which, contrary to other DOS conventions, does not apply to the current directory...!
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
