Smart things are so dumb because they take after their makers. Let's fix that

Tech is a great leveller. You can drop £50k on a shiny Tesla and £1k+ on the latest iPhone 13 Max Grunt to unlock it. But if some netops drone located half the globe away misconfigured a server, you're walking home just like a peon with a scratched-up Android and a battered Peugeot who dropped their keys down a drain. Now, we …

[Pete 2]

What are error messages for?

> "Server Error 500."

ISTM there are two issues here.

The first is to appease irate users who have made the basic error of relying on technology that still needs another 5 minutes under the grill before it will be ready for everyday use by everyday users.

The second is that the function of an error message is to communicate state in an accurate (top priority) and concise form. Such that those in the position and with the ability to resolve it receive the information they require.

In the second case short numerical error messages do the job very well. Provided the error codes are unique and well enough defined that a single code points to a single cause of failure (and from there, hopefully a well defined fix). Their nature also makes it easy for a non-techy to report the exact message to those who need to know. Often without too much prompting.

However, they also provide a perfect platform for tech-haters to criticise systems for "gobbledygook" as their meaning is obscure to the average person. However, would a longer message, along the lines of "I'm so sorry but the computer is completely buggered up at the moment. We will be restoring normality just as soon as we are sure what is normal" be any more helpful than "Oooops, sorriy", or just :( ?

[ComputerSays_noAbsolutelyNo]

Re: What are error messages for?

Numerical error codes can of course be consise and precise indicators of status, but

1, they need to be sufficiently unique to actually make sense

2. they need to be documented somewhere, and this somewhere should not be secret knowledge of the illumiated few

While it would be nice to see IoT IT adhere to these points, not even plain IT follows them.

Trouble-shooting IT errors, which provide a numerical error, should lead via the first google hit to a knowledge base of the respective vendor which lists all error codes and the conditions which trigger them.

What you get instead is: you google "product error X", and you get hits for "product error Y", and if you ever find some info on "error X", you read that "error X" may be caused by A, B or C.

Yeah, thanks for nothing. I'll get my coat.

[Anonymous Coward]

Re: What are error messages for?

Or you will either get -

Other people having the same problem and asking for help

-or-

People who had the same problem and fixed it, but no info on how it was fixed.

[Version 1.0]

Re: What are error messages for?

The way it used to be done was to generate an error number and then feed the number to an "explanation" routine that documented it for the end-user. The routine needed (and I saw this happen) a subroutine that reported "The application error message is an error" when the error number did not match anything.

[LybsterRoy]

Re: What are error messages for?

I also get the case where things don't work and there's no error message, go online and google the problem and find people who haven't had the problem recommending solutions that do not and can not work, or just answering some other problem they find more interesting.

[MisterHappy]

Re: What are error messages for?

I like the result of "Thanks to Bob for the help, I fixed it as detailed in his Blog... Here".

Then the hyperlinked "Here" returns a 404 or a placeholder page because Bob moved it or stopped paying for it.

[Martin-R]

Re: What are error messages for?

don't forget the hits on 'error x' but produced by a completely different product :-(

[Boris the Cockroach]

Re: What are error messages for?

Quote

"technology that still needs another 5 minutes under the grill before it will be ready for everyday use by everyday users."

I would say that a hell of a lot of that technology needs more than 5 minutes under that grill

5 days would be closer.. along with the system designers and the marketeers who insist we need the latest and shiniest.

Lets see... in these days of systems on a chip , is there any need for your phone app to talk to a central server, which then sends commands back to your lightbulbs to dim down 50%? hell no.

The phone app should be paired with the SOC running your lights... send a command to that and your lights dim... without the need for anything to leave your house and get filed away/sorted and sold to a 3rd company making curtains.

Or am I just old and grumpy?

[batfink]

Re: What are error messages for?

You're just old and grumpy Boris. Get up and use the switch and avoid the whole problem.

[Neil Barnes]

Re: What are error messages for?

Tsk. "Oi, Junior, get off your backside and turn the bloody light on!"

[WolfFan]

Re: What are error messages for?

Some of us are old enough to remember when they were the remote control for all kinds of stuff. Changing channels on the tv, turning the volume up or down on the tv, the stereo, the shortwave (young ‘uns today don’t even know what a ‘shortwave’ is), and, yes, turning lights on and off.

[quartzie]

Re: What are error messages for?

The remote control in the above post is simply called "Junior" ;)

On the other hand, said remote control is rather expensive to keep, and its reliability can become questionable during the 10-15th year of use.

[Neil Barnes]

Re: What are error messages for?

Oddly enough, as a Junior myself, I used to earn the odd threepence turning the lights on and off on a Friday night for our observant Jewish neighbours...

[ThatOne]

Re: What are error messages for?

> is there any need for your phone app to talk to a central server

Obviously. The whole point is to 1. force users to pay monthly subscriptions, and 2. to collect valuable user data you can sell for profit. The initial product, the light bulb, is just a loss leader.

[bombastic bob]

Re: What are error messages for?

Obviously. The whole point is to 1. force users to pay monthly subscriptions, and 2. to collect valuable user data you can sell for profit. The initial product, the light bulb, is just a loss leader.

Sadly I can NOT think of ANY other reasonable excuse for making a "light blub" require a cloud server to operate.

Here is what it COULD do instead, if limited to IPv4 and a small amount of NVRAM (like an AVR processor) to communicate: You just need to listen on the LAN with IPv4 [assuming it is behind a firewall] and assign it a name via mDNS so you can query for and/or activate the right "blub".

for access from outside, you could still do cloudiness, if you really want to.

(this is NOT rocket surgery)

And if your always on PC had an application running with motion sensors and "blub" control, for an automatic house, there's NO need for cloudy-BS to be a "man in the middle"

[BenDwire]

Re: What are error messages for?

A lot of IoT stuff can be reflashed with cloud-free firmware. I use Tasmota for my sockets, light bulbs and remote control receiver. Admittedly it's hopeless for non-techie types, but at least some of us can keep things local.

[Dwarf]

Re: What are error messages for?

+1 for Tasmota, its a really useful project when coupled with a LAN based mosquitto (MQTT) server.

-1 for all the vendors that decided that over the air signed firmware updates were the only way to go, hence removing customer choice.

What was wrong with giving customers the choice in how they want to use the device - locally or Internet based. After all, the tooling is all there with the configuration app to allow both and its not as if they are limited on resources on the hardware to use a different MQTT broker, login and password.

[Richard 12]

Re: What are error messages for?

mDNS doesn't scale. I'm sorry to say this, but I've seen the tests.

It gets slower and slower as more devices are added, and Apple's Bonjour falls over completely somewhere between 100 and 250 devices.

A four bed home probably has 50-60 individual lighting points. It's already too slow at that scale.

Someone going all-in for IoT is going to be mighty irritated that it takes 30-40sec to find the reading light.

[Anonymous Coward]

Re: What are error messages for?

Do you live inside a Xmas tree? 4 Bed house and I counted 32, and that was being generous and counting the cooker hood light. OK, 33 if my 70s style lava lamp counts.

[John Robson]

Re: What are error messages for?

DALI for lights would be a far better solution, but that does need five cables to each position - so only any good in a complete rewire.

[ThatOne]

Re: What are error messages for?

> Here is what it COULD do instead

Yes, in a perfect world full of rainbows and unicorns... You people tend to forget those gadgets aren't made to make you happy, but to make money, as much as possible, as fast as possible, for as long as possible. In this perspective all considerations on how they could had done it better are exercises in futility: They wouldn't sell more if they those gadgets were more configurable, they would just make less profit per unit.

[LybsterRoy]

Re: What are error messages for?

Your response is incorrect, the correct answer, obviously, is that you need to be able to turn your lights on and switch the colour whilst you're still at the office!

[John Robson]

Re: What are error messages for?

Why would that need a cloud server?

VPNs are now utterly trivial (e.g. tailscale) to have in place, allowing you to access your non cloudy light bulbs from any connection.

[Richard 12]

Re: What are error messages for?

That's exactly what industrial IoT does.

No outside contact needed.

The trouble is that almost all industrial controls fall into two camps: Closed proprietary, must get everything from the same manufacturer, or are "in the clear". With bridges in both directionsof course. So if a miscreant gets onto the plant network, they have full control over everything and can do what they like.

The people on the line will notice pretty quick, of course.

[Mast1]

Re: What are error messages for?

Weren't the chances of restoring normality any time soon determined 40+ years ago as infinitely improbable ?

[DiViDeD]

Re: What are error messages for?

Microsoft, since Win 8.1 have perfected the art of the simple, easy to understand while simultaneously 100% impenetrable error message with their classic:

Oh dear, something went wrong. Hold on while we try to fix it for you

[Mike 137]

"Consumer IoT IT, in short, is the worst IT in the world"

The worst, probably because in most cases the hardware/firmware is largely non-profit making. It's the data stream that's the source of revenue. Consequently the kit is 'made to a price'.

[bombastic bob]

Re: "Consumer IoT IT, in short, is the worst IT in the world"

well if you use a tiny processor to control the thing, with an external (but on board) wifi or ethernet to talk between CPU and world, the cost of something that can support more sophisticated access (including security) may simply be too high.

(less sinister, more practical, without denying the possibility of "sinister anyway")

[Anonymous Coward]

Re: "Consumer IoT IT, in short, is the worst IT in the world"

I really really wonder how someone in China actually monetises the fact that i switch a light on in my garage?

I can't really see this being a market transaction down the local dodgy pub - 'Hey mate, wanna buy some juicy intel? "Mr Anon needed a screwdriver at 7:30 and went into the Garage - so maybe that's a good time for a burglary?"

Maybe they sell it to the people building power stations? "There's a surge in power at No 55 - quick - sell another 5 million options on Brent Crude"

[Anonymous Coward]

As a techie, I don't think "500 server error" is a bad message, because it gives me information that I can use in troubleshooting. I know what it means and if I didn't then I could look it up.

It's far better than the modern windows favourite of "We had a problem". What kind of problem? Disk? Network? Null pointer? Access Violation? No fucking useful information whatsoever. I just have to guess what it might be.

It isn't any better than the blatantly misleading "You may not have permission to access this resource", which has never in history been triggered by a lack of permissions.

You rarely need a java style full stack trace, but you do need some accurate and meaningful technical information to start troubleshooting a problem.

[Doctor Syntax]

If Server error 500 is a catch-all for errors that should have been caught and maybe automaticelly recovered some way back along the chain it won't help you very much.

[Neil Barnes]

(My) WhatsApp (and plenty of others from what I read) is currently unable to send or receive pictures. The message is 'unable to download, please try again later' which while not all that helpful does at least give one a faint hope of some future resolution. It hasn't happened yet...

When I try to send an image from the camera, it complains that the camera isn't working. Best of all, when I try and send a just-photographed image, it cheerfully announces that said image isn't on the system... right below a picture of the picture it's claiming doesn't exist.

I'm not convinced that a full-text version of 404 or 500 is necessarily any better to the uninformed than the number itself; it's a magical incantation that someone has mispronounced somewhere along the line and let the magic out. The user can't - as a rule - do anything useful with it.

[bombastic bob]

"You had a problem" - the OS equivalent of a "check engine" light.

It helps keep back yard mechanics and knowledgeable users from fixing their own schtuff.

[W.S.Gosset]

"check engine"

In future, lights will just cut to the chase:

"pay money"

[sabroni]

re: I don't think "500 server error" is a bad message

You're wrong.

[veti]

The problem is that developers don't like to think about failure modes. It's hard enough (they reckon) to get the bloody thing to work under ideal conditions. If they started affirmatively cataloguing all the ways it can go wrong, someone might tell them to fix some of them.

And they're long past fed up with this project and really want to declare victory and move on to the next shiny.

Their managers mostly feel the same way. And senior management, like sales and finance, just wants to ship a product. Any product is better than none.

Nobody in this whole stack is motivated to make sure failure modes can be correctly identified or diagnosed, let alone described to the poor users.

[Dan 55]

"Server error 500"

I'm not sure what a plain-English replacement for that error code would look like.

Well I am, as it's a consumer-facing product, it would be something along the lines of, "Hey there, something's not quite right :(, why dontcha try again later when it's right. :)" with a whimsical doodle next to it.

Fucking IT.

[DJV]

Re: "Server error 500"

Maybe it should be along the lines of:

Error code for techies: Server error 500

Translation for normal humans*: Hey, it's not my fault! Me (the bit you're holding in your hand) is working fine. I'm trying to talk to another bit a long way away out on teh interwebs. Teh interwebs seem to be working fine as I am managing to shout all the way to other bit. But the other bit isn't talking back in a way that I can understand. Until it does there's not a lot I can do at my end. I suspect coffee is called for at your end. If you're not at home should I try to find the nearest coffee shop (within walking distance**)? While you enjoy your coffee I will keep trying to talk to the other bit and I will ding at you when things are all happy again.

Error messages like this will certainly not add to the app bloat much, well maybe a little (ok, a lot then).

(* yes, I know what that is implying! :)

(** added if the thing the user is trying to get working is a Tesla***)

(*** other electric vehicles as amazing**** as Teslas are available)

(**** stop laughing)

[John Robson]

Re: "Server error 500"

I'm laughing at how many people seem to think that this was the only way into a tesla... there are several options for opening the car, and only one of those (probably the least used) was broken for a while.

Actually I'm really laughing at the non tech press who reported that "Approximately 500 people were locked out"

[Richard 12]

Re: "Server error 500"

Presumably up to 500 owners forgot their keys.

TBH, it's fairly reasonable for someone to forget the keys sometimes if they normally unlock it with their phone.

My car yells at me if I forget my keys. Much more user friendly.

[LybsterRoy]

Re: "Server error 500"

Mine's even more user unfriendly - it won't open up if I don't have the key <G>

[keith_w]

Re: "Server error 500"

My house will be even more upset with me as I would not be able to lock it without the key that is on the same ring as the car key

[John Robson]

Re: "Server error 500"

If you "normally" unlock it with your phone you probably do so over bluetooth, not via the cloud service.

[Ian Johnston]

Re: "Server error 500"

Laughing at the people affected by a problem does appeal to some people, I suppose.

[John Robson]

Re: "Server error 500"

Those affected didn't feature on my list of people I was laughing at.

[veti]

Re: "Server error 500"

"The server isn't responding properly. This is likely to be a high profile issue that the relevant people are already aware of and are working to fix as quickly as possible. If you can afford to wait an hour or so, it may clear up in that timeframe. Otherwise you can try calling this number and asking them about it."

[Ian Johnston]

Re: "Server error 500"

"... and remember that you can also unlock your car with the keycard."

[Martin-R]

Re: "Server error 500"

Oh if only... I was recently locked out of my business banking for 10 days due to a 500 error while they'd been telling other people with the same problem there was no problem... Posting the error message from the Chrome developer console to their Twitter support desk at least got them to admit there was a problem, but it took several more phone calls and them physical posting me a 'magic word' to get it resolved :-(

Hats off however to the Ocado developers who put their recruitment links in the console complete with Ascii art of the logo :-)

[Dan 55]

Re: "Server error 500"

The Grauniad does that too.

[batfink]

It's all going to be fine

I'm assured by the UK gummint that they're all over this and they're going to fix IOT problems by passing a law which outlaws default passwords, which should of course also somehow fix all the short- and long-term support issues.

[Filippo]

I have a semi-serious proposal

Some new legislation is required (isn't it always)!

First of all, now every company that makes consumer electronics has to provide 24/7 phone support, and people who call support must be able to speak to a human within 30 seconds.

Next, minimum wage for phone support staff is now 50 bucks an hour. You can outsource it, but then we'll tax you twice that much in duties.

I bet that everything will suddenly become so intuitive and reliable that a 90-years old granny will be able to figure it out unaided, and it will keep working forever.

[WolfFan]

Re: I have a semi-serious proposal

Depends on the 90 year old granny. I’m fairly sure that Grace Hopper, were she still with us, could figure out a whole lot more stuff than my late nana, born in County Kildare [redacted] years ago. A technical genius she wasn’t. She was, however, the main source for the bad attitude generally found throughout the family.

[Ken Hagan]

Re: I have a semi-serious proposal

A bit harsh on poor Grace there. Are your family all COBOL programmers or something?

[bombastic bob]

Re: I have a semi-serious proposal

when i deconstructed his sentence it seemed that some of it could have applied to his grandmother, and not Grace Hopper. It seemed a bit ambiguous to me.

I once saluted Capt. Grace Hopper when she was the RTC commander in Orlando, Fl. I was near her office for some reason (3rd class petty officer at the time), she came out, so I gave her a salute thinking "wow a lady captain - she must be the RTC commander". found out later who she was.

[Anonymous Coward]

Re: I have a semi-serious proposal

My 90 year old granny decided at about 75 that 'these computer things will be important in making my life easier" and applied her usual determination to get comfortable enough to learn what she needs to know. Hence internet shopping, emails and VC s to grandchildren and great grandchildren etc etc. Not going to debug some complicated TCP-IP protocol - but as capable as many younger people and quite capable of the essential fix to 98.7% of all IT issues- Turn it it off and on again, and if that doesn't work, unplug the network cable/power/usb/whatever and plug it in again."

[steelpillow]

Market forces

Polishing apps slows things down. Consumers demand the latest.

That's how people like Facebook grew big - they move fast and break things.

Similarly, most hardware these days has barely functional drivers when launched. The first thing the buyer has to do is download and install the latest update.

So as a supplier what are you going to do, polish yesterday's turd you have already been paid for, or rush out tomorrow's half-digested drivel ahead of sanity, and fill yer boots once more?

If you ever bring out a soft-fail standard, call it Polish or Gilding or similar. But nobody will stop by to adopt it.

Consumers have only themselves to blame.

[Twanky]

Re: Market forces

...polish yesterday's turd...

...tomorrow's half-digested drivel...

...fill yer boots once more...

Umm. I'm suddenly not feeling very well.

(have an upvote)

[jmch]

Re: Market forces

"as a supplier what are you going to do, polish yesterday's turd you have already been paid for, or rush out tomorrow's half-digested drivel ahead of sanity, and fill yer boots once more?"

This is actually one of the most compelling reasons for software as a service. As long as user data is not locked in, they can move on to another vendor if the current one sucks, and vendors have compelling financial reasons to support their products through their full lifespan

[Ian Johnston]

Re: Market forces

That's how people like Facebook grew big - they move fast and break things.

It's a shame Tesla applied precisely the same design philosophy to Autopilot.

[ThatOne]

Won't work. :-(

> Total absence of diagnostics isn't just a complete repudiation of the right to repair, it removes any motivation or ability to manage security.

Well, seems you explained yourself why it won't ever work: The commercial motivation to keep things that way in IoT is too strong, since adding diagnostics would not only increase the price without adding any marketable bling, it would also significantly lower the profits since people would repair instead of buying the next newest version, "guaranteed to have fixed these problems". It might go as far as people using the same kit for many years, without buying replacements!

The only way to force security and repairability to IoT stuff is to prune the market by making those features mandatory. 99% of stuff will disappear, the rest will be solid(-ish).

Now concerning error messages for Tesla users, that problem has no solution. Even if you sent a real human to explain, bow and ask for forgiveness, people would still complain. Because the issue here was not the message (they wouldn't be able to do anything about it anyway), it's the fact there was a problem which disturbed their everyday life.

[Charles 9]

Re: Won't work. :-(

How can they be mandated when the market forces are against regulation? They have the clout to influence the government, after all. Thus the bit about "tombstone" regulation, where it takes death and outrage to force the issue because that's about the only thing that can out-influence the market. And even then that's not always enough (like guns in America, which has found itself a counter-outrage to keep itself going--finding a reason to hate the only thing that can fix the problem).

[ThatOne]

Re: Won't work. :-(

> How can they be mandated when the market forces are against regulation?

Oh, I didn't say it will happen (or that it is even possible), I only said it is the only way I can think of to influence the IoT trash production.

[Spacedinvader]

Zengge Wi-Fi protocol

Gives 80,300 on my Google search...

[Gene Cash]

Re: Zengge Wi-Fi protocol

Interesting. I get "About 37 results (0.91 seconds)" and if I put it in quotes, I get just this article.

One result is "vikstrous/zengge-lightcontrol: A command line tool ... - GitHub" which is probably useful.

Another one is:

https://faqsys.magichue.net › privacy › policy › magicB...

Your privacy is important to us here at Zengge Co., Ltd., and we want to be clear ... local hardware identifier for applications using wifi and Bluetooth.

Which is absolutely hilarious.

[Ian Johnston]

Re: Zengge Wi-Fi protocol

You need to put quotes around it to search for the phrase. Then there are indeed only two hits.

[Anonymous Coward]

"in the closing stages of 2021 was "Server Error 500."

That's because people who never used a Commodore or a Spectrum think it is a good thing to use a web server for everything - even when there are far better technologies for device communications. The "web" mindset is crippling IT and making it a pile of layers barely glued together because the average millennial and whatever came later has been brainwashed by fashion- and marketing- driven development and it's all about colorful icons on your display, idiotic names, and underneath data hoarding. Everything built on top of an OS that with a design already outdated forty-five years ago. Plus languages like Go that goes backwards about error management. Blaming MS for its "Error 0x80002122" is fine, but any idiotic error message spluttered by a FOSS code it's because the developers are cunning, not lazy or incompetent.... and sometimes how could you get and handle the original error and message when the code path went through layers and layers of different types of code and languages without a common underlying error handling framework, maybe invoking here and there some command line script or utility? Applications are becoming brittler and brittler - and soon we'll see the full effects of that.

[Ken Hagan]

Re: "in the closing stages of 2021 was "Server Error 500."

Quite a good point right at the end here. For years we've been taught that it is good to abstract stuff and hide things. Yes, usually it is. It's great that I can bolt applications together and one library doesn't need to know where the other library is getting its data from. You end up with clients that can talk to any server and servers that can be used by any client, and no-one has a fucking clue about errors that are detected outside their own little abstracted world.

Exceptions are a fairly good way to pass around the (boolean) fact that there is a problem. They don't (in my experience) offer much in the way of end-user diagnostics. You'd need a higher level pattern than that. Are there any good ones?

[Richard 12]

Re: "in the closing stages of 2021 was "Server Error 500."

Exceptions work if they're actually exceptional, and the one throwing it actually puts the diagnostic into the payload.

Trouble is, too many devs leave them empty, and too many use them for normal failures - file not found is not exceptional, disk on fire is.

[Anonymous Coward]

Can't make it as a programmer...? Wiling to accept minimal wages for your awful coding?

Excellent! The BT Home Hub team has a job for you.

Are there any routers out there that work?

[BenDwire]

Re: Can't make it as a programmer...? Wiling to accept minimal wages for your awful coding?

Are there any routers out there that work?

pfSense has served me very well for the past decade

[Ian Johnston]

Re: Can't make it as a programmer...? Wiling to accept minimal wages for your awful coding?

Andrews and Arnold supplied my with Xyzel which seems OK, though the UI is tortuous.

[43300]

Re: Can't make it as a programmer...? Wiling to accept minimal wages for your awful coding?

I had loads of trouble with Xyzel ones - caused VPN tunnels to refuse to auto-reconnect when they dropped. Never encountered this particular issue with any other routers.

[a_yank_lurker]

IoT

Idiocy of Twats is something I try very hard to avoid. To many of the devices are for convenience but I have not yet figured out convenient for whom; the marketing slimes or very doubtfully me. And I am not very thrilled to share much with vermin who should be used something useful like target practice.

[yetanotheraoc]

Careful what you ask for

From the article: "undocumented infrastructures with no discernible diagnostics"

Would it make you happy if they documented the debugging port with the baked-in admin password? Sometimes obscurity is the only security there is.

[JohnG]

About that Tesla server outage....

These cars come with key cards or key fobs. If you choose to use the mobile app to unlock and start the car, a server outage is not the most likely problem. You are far more likely to drive somewhere without cellular connectivity for the car, your phone or both e.g. an underground car park. It's all fine when you are driving in but when you park up and leave the car, it will lock as you walk away. If the car and phone don't have cellular access, you won't be able to unlock it. Another problem is when your phone battery dies.

[scubaal]

youa re the weakest link

reflects something I have been saying since the start of IoT. The system is only as strong as the weakest link, and the weakest link is consumer grade no-name ebay IoT widgets produced in the million and never updated or patched

[jvf]

Who's running the show?

Ever wonder where Beavis and Butthead ended up? Check out the Board of Directors list of several major IoT companies. I’ll bet they’re on most of them.

[Dr. Ellen]

Lightbulbs and copiers

There's an old saying: "It's hard to make things foolproof, because fools are so ingenious." The internet of Things is one thing - but there is a whole world of Things that don't use the Internet. But they're designed by the same kind of engineers.

My particular object of tech-hatred was a copier that landed in our library some twenty years ago. It knew how you wanted to use it, and it was determined to do it for you. Only that wasn't always what I wanted to do. An 8-1/2 x 11 copy on 8-1/2 x 11 paper? Flawless. A life-size copy of something small? Surely you wanted the image to take up the entire sheet of paper. Nothing else would do! It could be commanded to do what you wanted, but as soon as you lifted the cover for the next copy, it set itself back to what it knew you wanted it to do.

The trouble was, the people who designed, made, and sold the copier made it work the way they wanted it to. And the rest of us. who used copiers differently, were left cursing in their wake. The makers and the customers weren't the same people. It works that way for light bulbs, too.

[Sir Loin Of Beef]

Humankind will never build anything greater than humans because humans are limited by our, well, humanity. We do not know anything outside of who and what we are.

[anonymous boring coward]

"Humankind will never build anything greater than humans"

Skynet begs to differ.

[Grunchy]

Server Error = Cloud Error

Anytime you don’t have complete and total control of something it’s because you either don’t fully own it or somebody is taking liberty with your property (you know, as if they stole it).

I would never buy a Tesla. The hell with that!

[Anonymous Coward]

Déjà vu

"or an industry that learns to look after itself"

Hmmm, haven't I heard this quite a lot of times about the financial markets ?

Advocates of this went bizarrely after 2008 :)

[anonymous boring coward]

Error messages?

"Something happened" should be all anyone needs, surely?

And an infinitely spinning thing, or a "sadly" smiley.

Just read what lusers write about their Windows updates. Things like "I'm on day three now. Should I wait longer for the update to finish?"

I guess we deserved what we got.

[Ian Johnston]

The best system for error messages was the BeOS browser, NetPositive, in which they were all haikus:

Click exciting link

Gossamer threads hold you back

404 not found.

The most apposite one for this article is:

Errors have occurred.

We won't tell you where or why.

Lazy programmers.