Why don't computers work?

[hedron]

I don't think I've ever owned a computer that didn't have multiple problems that kept reoccurring and preventing me from using my computer. It doesn't matter the software, OS, hardware combination I'm using, whether it's a laptop or desktop, linux, mac or windows, intel or amd, nvidia or amd, etc, etc. The only notable exception I've noticed is single use devices like video game consoles. I don't think my old SNES ever screwed up, well unless you count all that slow down. Now that I think about it, neither did my old C64. But C64's weren't able to multitask, were they?

Why should a computer that was working perfectly yesterday, suddenly start locking up today? Did underwear stealing gnomes find my drawers too rank, so they got bored and messed up my computer instead?

 

[Whatsisname]

Because they are really fucking complicated, and market demands on PCs don't emphasize stability.

 

[Mister Natural]

PC's and operating systems are quite stable actually. It's the crap that someone installs on them that leads to problems. If you have constant problems with computers then you need to review what you're putting on them.

 

[zero2dash]

Can't really compare a closed ecosystem product available in 1 configuration (like a game console) to a computer system that is an open ecosystem product with a billion configuration possibilities.

Either you can't build computers very well, you're buying cheap products, you don't cover all the bases on the initial installs, or you have the worst luck in the world (but that can't be the case otherwise you'd be dead and unavailable to post on [H]).

Judging by your sig, I see a few decent brand names in there....IDK man.
Out of all the PC's I've built, I've had relatively very few issues. I'm talking probably at least a couple dozen systems, 1P and 2P (desktop and server), tons of different hardware, variants of Windows, drivers etc. over probably 15-20 years.... other than a rare DOA hardware issue, I haven't had much in the way of problems on any of those systems. Windows today (generally post-2000) usually sets up right and runs fairly flawless, as long as you check for drivers and install when necessary. Keep an eye on your apps, make sure you'd using good programs and drivers.....keep all testing to a VM. That eliminates a lot of potential problems.

I hate to say it's kind of [H]ard to fuck a computer up nowadays without doing a really good (bad?) job of it.

 

[bcrisp82]

you're holding it wrong.

 

[Cheetoz]

A computer is a complex device that tries to get dumbed down for regular customers.

Most of my computers have always been really solid, and when I setup old computers as servers, they get over a year of uptime (unless the UPS runs out when the lights go out or something).

But we can all safely say most computers are more stable than that first batch of xbox360s.

 

[bbouzos]

I assume you're more frustrated than serious in this posting, but the guys above me got it right. Computers are complicated machines, it's quite a task trying to make them usable for everyone!

 

[hedron]

Computers are complicated machines

Yea, but a lot of things are complicated. SNES is a complicated device. But I've never had a problem getting it to work. OK, I haven't turned the thing on in more than 10 years, but I'm quite certain that if I dig it out of my parents basement it will still function properly. I don't think there is a computer on Earth that could boast such a feat.

I just can't understand how I can turn off my computer yesterday and have had absolutely no issues what-so-ever, well aside a few annoyances, but anyway, and then turn it on today to find out I have recover my entire terabyte hard disk from back up because it just doesn't feel like syncing today. Last time this happened it took me a week to get back to normal because it will constantly lock up and crash trying to copy the files. And other issues like my Razer mouse just deciding at one point that it won't enable my configurations one day, just because. I could understand it if like I removed my video card and couldn't get a signal on my monitor. I can understand problems like that.

 

[PTNL]

There's more moving parts in computers than the SNES -- both then and now. Add the fact that you had zero control of the software on the SNES, and you see a very tiny footprint of actual usage, interaction, and even content creation when comparing a 90's console to a computer.

More simply put: apples and oranges.

 

[devil22]

SNES is complicated, but a modern computer is hugely more complicated than an SNES. Additionally, SNES' have failed (a lot of my friends had to 'blow' on the cart to get their games to run.) Sometimes software fails because a lot of programs are a huge amount of lines of code and humans make mistakes. Hardware fails because it is cheap, or you got unlucky (sometimes the software/firmware for the hardware has a mistake in it). Sounds to me like you need to check your hard drive to make sure it is not dying or maybe invest in a dedicated UPS, I've had a lot of hard drives go especially the big ones (2TB+), but usually this hardly affects me since I just restore from backups. My OS drive is just a 500GB with a few programs installed, if it died I could get another and reinstall all my stuff in a couple of hours. So maybe you should organize better and plan for things like hard drive failures, I know you have backups but if it takes you a week to restore you have organization gaps.

 

[PTNL]

[...]Sometimes software fails because humans make mistakes.[...]

ftfy

Judging software quality and completion by LoC numbers is an exercise in futility.

 

[devil22]

ftfy

Judging software quality and completion by LoC numbers is an exercise in futility.

Where did I judge software quality and completion by LoC numbers? What I was getting at, is that it's easy to make mistakes when making huge complex code bases. The same principle does apply to small simple code bases, I never said it didn't, I was just trying to help the OP visualize the issue easier.

 

[Skripka]

What I was getting at, is that it's easy to make mistakes when making huge complex code bases.

Yup. When my Intel NIC on my motherboard needs a 100MB driver for the ethernet port, and my HP printer has a 300MB installer for its driver...

 

[PTNL]

Where did I judge software quality and completion by LoC numbers? What I was getting at, is that it's easy to make mistakes when making huge complex code bases.

Your judgement was done by your assertion that there is a correlation between odds of failure and an unstated number of lines of code:

Sometimes software fails because a lot of programs are a huge amount of lines of code and humans make mistakes.

You could make an argument about maintainability, knowledge transfer between development team members, the tail-end of the software development lifecycle, and some other more-focused topics. But the blanket statement you gave is completely lacking any supporting context, and leaves a muddled perspective for those that have not worked on development projects with (both) a significant weight and a wide user audience.

 

[devil22]

Your judgement was done by your assertion that there is a correlation between odds of failure and an unstated number of lines of code:

All things being equal, more lines of code = more failures = more fallout from those failures, in general. It seems you are arguing against a hypothetical "simple large code bases have more failures than smaller complex code bases" or whatever permutation of that you are imagining. I never said anything to make you assume any differences except lines of code. And none of this or anything else I said is a judgement of quality or completeness of the code in question.

You could make an argument about maintainability, knowledge transfer between development team members, the tail-end of the software development lifecycle, and some other more-focused topics. But the blanket statement you gave is completely lacking any supporting context, and leaves a muddled perspective for those that have not worked on development projects with (both) a significant weight and a wide user audience.

Uhm because I was talking to a guy who asked "why do computers fail" and not "is there a correlation between Nth order complexity, size and failure rates and what's the log of that yadda yadda yadda" Man, I hope we aren't going to do this all day...

 

[octoberasian]

I got into a discussion with a friend of mine from Louisiana a year ago with a similar issue. It started when he IM-ed me on AIM saying he wish computers were more like consoles.

"Why don't computers just work?" he asked me.

He's gone through three power supplies in one year, two motherboard replacements, and four reinstalls of Windows 7 in that same year.

This is what he is looking for in a PC:

• You turn it on, no loading screen to the OS.
• You pop a game in and it runs without installation or very minimal installation.
• Graphics settings are automatically set based on the hardware you have, and at the maximum possible.
• Unobtrusive and hands-free OS. software and game updates.
• You plug in any controller, keyboard and mouse and it works without fiddling with drivers or customization options outside the in-game's control settings.
• You upgrade any piece of hardware, the drivers are already there/already added without the need to manually download them. Driver updates are done automatically without user-intervention. And, the hardware just works right off the bat without issue.

He would also like the hardware be both simple and powerful enough to run current and upcoming games to the maximum possible settings that the system can handle.

In other words, a computer that is like a console in simplicity and much less hardware failures. He wants it to just work and work stably enough without or with very little issues.

You turn it on, it's ready for you from the get-go. It's why he prefers playing on his PS3 and 360 more than his PC because they've given him less headaches and less financial frustration than his PC has. He's tired having to fix hardware issues.

But, I told him, it's being a little unrealistic. As someone said above, there's a difference between a closed and open system. The desktop computer is a very open system. It has hundreds if not thousands of different hardware and software configurations that there will bound to be issues and conflicts later down the line. The video game console doesn't suffer from this a whole lot. It's just one set of hardware configuration other than storage and a single graphics API/driver set and highly customized OS. Games for consoles are designed and developed to take advantage of every single ounce of graphics and processing power of that system. You also cannot upgrade a console like you can on a PC. But, you plug in a controller, it just works. You pop a game in and it's ready to be played immediately.

I joked at him that if he wants a closed system or a semi-closed system, he can go to an Apple computer. It's a single OS, single set of hardware configurations aside from CPU, video card/GPU and storage options, but is considerably more expensive.

The closest thing I can think of that may come close to what you want will be the rumored Steam computer or set-top box: http://www.pcmag.com/article2/0,2817,2401119,00.asp

Posted here on [H]: http://hardforum.com/showthread.php?t=1677457
My thoughts on it here: http://hardforum.com/showpost.php?p=1038455765&postcount=66

===============================================

Side-note:

If you don't know what WinRT is, it's supposed to be the replacement compiler/API to Windows API, or Win32. Microsoft is intending that future programs and applications be programmed using WinRT (or, Windows Runtime). It'll be very similar to mobile apps on your smartphone and tablets. It will require very minimal to no installation. You can run the app/program right away once downloaded. By moving to WinRT and simplifying it for the end user, that should lessen a lot of the confusion and any complications some PC users have in installing a program on a new computer. Only Windows 8 and higher will be compatible. It's solely intended for Metro style apps for Windows 8. I have no idea about Windows 7 but most likely it should still be compatible. Windows 8 will still be able to run Win32-based applications.

I think a very good hint at this and the direction Microsoft is taking is Office 365, the downloadable Office 2013 Preview, cloud-connected Microsoft Office. You download the preview from the Office 365 website, run the installer, it downloads the necessary files and you're up and running in mere minutes. One person on [H] commented that he was able to run Word right away while it was still downloading the files. I was able to do the same thing.

Apparently there will be two versions-- Office 365 and Office 2013 if going by the forum posts in the Office 2013 thread. The Office 365 version seems to be the least obtrusive and simplest to run. You choose which one you want-- Home, Premium (I think) or ProfessionalPlus. Download the installer, it downloads the necessary files and you can run the programs right away. Easy.​

 

[zero2dash]

Yea, but a lot of things are complicated. SNES is a complicated device. But I've never had a problem getting it to work. OK, I haven't turned the thing on in more than 10 years, but I'm quite certain that if I dig it out of my parents basement it will still function properly. I don't think there is a computer on Earth that could boast such a feat.

No it's not.
Does the SNES have moving parts? Optical sensors? Magnets? Does it put out heat?

It's a single circuit board with flash memory, no moving parts and (basically) no software that doesn't require anything other than a power cord to function.

That's like saying the car on the Flintstones is complicated like a Subaru, or a hang glider is complicated like a 737, or even a transistor radio is complicated like an iPod.

 

[k1pp3r]

Why Don't computers work?

I always tell people "A Computer only does what you tell it to"

 

[Whatsisname]

Don't hate on the SNES. It's not as complicated by today's standards compared to what we have now, but it sure was when it was built.

The reason for stability in stuff like the SNES isn't because they are simpler, but because the market demands stability in gaming consoles, it doesn't for PCs.

No it's not.
Does the SNES have moving parts? Optical sensors? Magnets? Does it put out heat?

It's a single circuit board with flash memory, no moving parts and (basically) no software that doesn't require anything other than a power cord to function.

That's like saying the car on the Flintstones is complicated like a Subaru, or a hang glider is complicated like a 737, or even a transistor radio is complicated like an iPod.

 

[zero2dash]

Don't hate on the SNES. It's not as complicated by today's standards compared to what we have now, but it sure was when it was built.

I'm not "hating" on the SNES; it's one of the top 3 game consoles ever created.

However, it's a simple design compared to a computer. Lumping in any game console along with a computer as being "complicated" that in of itself is foolish, but especially even moreso when you're talking about a game console that has flash memory and no moving parts.

The reason for stability in stuff like the SNES isn't because they are simpler, but because the market demands stability in gaming consoles, it doesn't for PCs.

....and because it's simple. Ignoring the electronics aside, it's simple. It's literally a circuit board with a cartridge slot, 2 controller input ports, a video out and a power in. Oh, and the "accessory" port on the bottom that was never used for the planned CD device by Philips.

Compared to a motherboard, ram slots, a CPU socket, expansion slots, power slots, external connectors of all sorts and sizes, input data slots, etc. that do absolutely nothing without an external storage device, boot code, and thousands of lines of code on top of that no matter if you're going for console only or GUI.

The market does not demand stability in game consoles. Never has, never will. No one in their right mind spent 100% of their NES, SNES, and N64 time without blowing on a cartridge at least once. You call that stability? Stability would be if blowing on the carts was not required from time to time. If stability was a requirement, Sony would be out of the game console business after people had to play their PSX units upside down and standing up. Not only are they past that, look at all the other failures they've had (Ps2 lens problems with DVD, Ps2 format games, sometimes both; Ps3 problems with YLoD; PSP dead pixel/screen problems, flying ejected PSP discs). Microsoft RRoD failures rate being rated 23.7% as of 3 years ago by Squaretrade! Hey, NES/SNES units have survived floods and worked afterwards; I'm not saying they're not built like tanks (compared to a 360)....but what I am saying (again) is that comparing a game console to a computer in terms of complexity is just really silly.

You can't just say "well it's got a circuit board like a computer, it's complex!" Doesn't work like that. In that case, a clock radio or a microwave is as complex as a computer.

 

[PTNL]

All things being equal, more lines of code = more failures = more fallout from those failures, in general. It seems you are arguing against a hypothetical "simple large code bases have more failures than smaller complex code bases" or whatever permutation of that you are imagining.

More code does not equate to more failures. Poorly written/tested/accepted code does. That distinction is key, as I'm disagreeing with your earlier blanket statements.

Man, I hope we aren't going to do this all day...

There is at least one consensus

 

[PTNL]

Side-note:

If you don't know what WinRT is, it's supposed to be the replacement compiler/API to Windows API, or Win32. Microsoft is intending that future programs and applications be programmed using WinRT (or, Windows Runtime). It'll be very similar to mobile apps on your smartphone and tablets. It will require very minimal to no installation. You can run the app/program right away once downloaded. By moving to WinRT and simplifying it for the end user, that should lessen a lot of the confusion and any complications some PC users have in installing a program on a new computer. Only Windows 8 and higher will be compatible. It's solely intended for Metro style apps for Windows 8. I have no idea about Windows 7 but most likely it should still be compatible. Windows 8 will still be able to run Win32-based applications.​

The WinRT stuff does look intriguing. It is a wrapper on top of the Win32 API, and does expose a lot of what used to be done with existing methods of invoking with Win32 API calls. The advantage is that the WinRT library comes as part of Windows 8, and can be accessed from Metro apps written in .Net or HTML+JS.

Though I'm unclear as to why that side note was brought up on this thread.

 

[wonderfield]

More code does not equate to more failures. Poorly written/tested/accepted code does.

I believe you're failing to see the forest for the trees here. Every line of code introduces an opportunity for error. As a code base grows in size, the opportunity for error increases. That's all he was saying. I suspect you know that already, though.

 

[devil22]

More code does not equate to more failures. Poorly written/tested/accepted code does. That distinction is key, as I'm disagreeing with your earlier blanket statements.

Of course. Then why is there no large (100MB+) program that is not pure repetition (like one command repeated many times) and not just wasted space, that is as bug free as a hello world program I can write? Because people who write hello world programs are better coders, or do better testing??

There is at least one consensus

Yet you waste my time anyway. Imagine that.

 

[PTNL]

Of course. Then why is there no large (100MB+) program that is not pure repetition (like one command repeated many times) and not just wasted space, that is as bug free as a hello world program I can write? Because people who write hello world programs are better coders, or do better testing??

Your implied familiarity with large software installations seems rather shallow.

Yet you waste my time anyway. Imagine that.

LOL.... You can only blame yourself.

 

[devil22]

Your implied familiarity with large software installations seems rather shallow.

As opposed to the guy who thinks a 100MB program can be as bug free as my hello world program?

LOL.... You can only blame yourself.

Sure, I can blame myself, I can blame the cubs, I can blame Obama. I think it's more accurate to blame you.

 

[PTNL]

As opposed to the guy who thinks a 100MB program can be as bug free as my hello world program?

Apples and oranges.

Sure, I can blame myself, I can blame the cubs, I can blame Obama. I think it's more accurate to blame you.

That opinion is already well established.

And if there's one other thing we can agree on, it's that this thread has been derailed for long enough.

 

[devil22]

Apples and oranges.

At this point you are arguing for the sake of arguing, and really have nothing left worth anyone's time on the subject. My original claim that you took issue with is that larger lines of code equates to more errors (in general, all else being equal.) To demonstrate my point, I put forth that I can write an hello world program with 0 bugs, but that there is no such thing as a 100MB+ program with 0 bugs, or anything even close to that, that I have ever heard of. Pretty much Q.E.D.

That opinion is already well established.

And if there's one other thing we can agree on, it's that this thread has been derailed for long enough.

Feel free to stop replying, because I'm not going to reply to a post you don't write.

 

[Cheetoz]

The argument should be more tailored to 'general purpose vs application specific' hardware/software

 

[PTNL]

At this point you are arguing for the sake of arguing, and really have nothing left worth anyone's time on the subject. My original claim that you took issue with is that larger lines of code equates to more errors (in general, all else being equal.) To demonstrate my point, I put forth that I can write an hello world program with 0 bugs, but that there is no such thing as a 100MB+ program with 0 bugs, or anything even close to that, that I have ever heard of. Pretty much Q.E.D.

A "hello world" application is trivial, and provides no value for anyone using the application -- bug free or not. Your argument is still hinged on an "apples and oranges" comparison, no matter how many times you apply the "all else being equal".

 

[heatlesssun]

You know why computers don't work? Because they don't have jobs.

 

[devil22]

(In general, all else being equal) A "hello world" application is trivial, and provides no value for anyone using the application -- bug free or not. Your argument is still hinged on an "apples and oranges" comparison, no matter how many times you apply the "all else being equal".

Trivialness, and value to people using the application are irrelevant and subjective. I can find a hundred examples of non-trivial but small apps that have value to their users that have 0 or almost 0 bugs, I'm betting you can't do the same for 100MB+ apps, many of which have no value to many people.

 

[octoberasian]

The WinRT stuff does look intriguing. It is a wrapper on top of the Win32 API, and does expose a lot of what used to be done with existing methods of invoking with Win32 API calls. The advantage is that the WinRT library comes as part of Windows 8, and can be accessed from Metro apps written in .Net or HTML+JS.

Though I'm unclear as to why that side note was brought up on this thread.

I was giving an example in my post regarding that rumored Steam settop box PC and how something like WinRT might allow easier software installation. Because, as I said in my reply, I gave an example of my friend who asked me something similar before regarding why don't computers just work when you turn it on.

Re-read my post. My friend is interesting whenever I get into discussions with him on AIM. He asks me a lot of random and sometimes inane questions. But, his computer issues are not made up. He's even had to ship his entire desktop PC tower to me to fix it for him.

 

[PTNL]

I was giving an example in my post regarding that rumored Steam settop box PC and how something like WinRT might allow easier software installation. Because, as I said in my reply, I gave an example of my friend who asked me something similar before regarding why don't computers just work when you turn it on.

Re-read my post. My friend is interesting whenever I get into discussions with him on AIM. He asks me a lot of random and sometimes inane questions. But, his computer issues are not made up. He's even had to ship his entire desktop PC tower to me to fix it for him.

Ah, I see your point. While WinRT would likely not be used in that case, I can see the high-level analogy being made. And I had no doubt that your friend was having computer issues.