1

In short

Sometimes, people use computers in a way that taxes their hardware and can eventually lead to CPU or BIOS failure. When it does happen, there might not be any warning. Users can go from using a computer normally to a black screen almost instantly. So:

  1. Is there a way to tell users in laymen's terms that they will need to repair or replace their machine soon?
  2. How can machines guide users toward using computers in a way that is more sustainable for the life of the machine and the machine's intended use?
  3. Are there any studies about this?

More details

When a machine's BIOS fails or especially when its CPU fails, the machine has far less ability to give users helpful feedback about the current system status. Depending on what went wrong, users may be faced with the Black Screen of Death and unable to interact with their operating system's GUI. So users have to resort to documentation that is often not designed for them in order to figure out what is wrong. More often than not, it leads to giving up and buying a new machine - and possibly losing any data that they haven't backed up or stored in the cloud. This may be profitable for computer companies (a dark pattern in customer experience?), but it is bad for users.

(Note: I tagged this as "graceful degradation" not in terms of how that term applies in web design - e.g. using a site with older browsers - but in terms of fault tolerance due to hardware failure. On Wikipedia, "graceful degradation" redirects to "fault tolerance".)

What I would like to know

I'm wondering if any operating systems or computer manufacturers proactively help users to use their machines more sustainably, without making them open an external utility such as Task Manager and with using plainer language than Task Manager uses. I would also like to know if any operating systems or computer manufacturers monitor users' machines to let them know that parts are about to fail or need to be replaced, without making them look at documentation and guess after a failure occurs.

In particular, I'd like to know if there are any studies or research (not subjective opinions) showing how this can be done and if/why current ways of doing this are not supporting the interests of end users (i.e. supporting their goal of not having their machines die).

There are utilities that measure CPU heat and other metrics at a low level, but these tend to be geared toward expert users, who have to know that these tools exist and have a fairly strong knowledge of computer hardware. I'm asking this more in line with how more average users can more easily process their computer's "end of life" and avoid losing their machines.

  • CPUs and BIOS don't tend to 'wear out' due to 'email and facebook use'. I'm not sure the premise of the question is valid. – DA01 Sep 23 '15 at 18:25
  • As for why computers don't tend to have a lot of failsafes, it's the same as with most products: Consumers want a cheaper price point. So, if 'load' was actually a 'burden' on the CPU, then it's a cheap CPU. And if it's a cheap CPU, this is not a computer being marketed at people that want the additional cost of 'CPU is wearing out' warning systems. Those people just buy better systems...where you would rarely need a 'CPU is wearing out' warning system. – DA01 Sep 23 '15 at 18:26
1

Most systems do already have failure warnings in place - the trouble is that they can't tell is something is about to fail until it does and, even when they do know that something's wrong, they can't tell if it's going to be a problem to you or not.

The failures you described sound like the power supply has died. If that is the case then the CPU will have no clue that the PSU is about to fail until it goes pop!

Most systems will tell you if they are experiencing a lack of communication from non-vital hardware (something like: "Ethernet card is unresponsive") and, because not all systems are built/configured/used in the same way, it's up to you to decide if you need a new machine, if you can repair the fault, or if you can simply ignore it.

The amount of logic required to add a software layer to monitor and report every parameter of your machine's health and then decide, based on your set-up and usage, which measurements it should warn you about would make your system slow and difficult to use and that, in turn, is likely to trigger the warnings!

  • The power supplies actually seem to be fine. I've replaced one of the machines already. The new machine works with both the old power supply and the new one. – David Sep 21 '15 at 16:38
  • 1
    @David, I'm afraid the case still stands: there is no way the CPU can know of a fatal failure before it happens without a ridiculous amount of extra software getting in the way of normal usage. – Andrew Martin Sep 21 '15 at 17:25
  • Most of the Failures are Real-time. Whereas others can be caught as an Exception. The Real-Time failures cannot be predicted and a message cannot be shown as to what might have went wrong. For the Exceptions, an error code is shown. – Swapnil Borkar Sep 22 '15 at 15:16
  • I've changed the question to focus more on how UIs should educate users on using computers in a way that takes good care of the machine (in terms of the load placed on it). Still would like to see studies on this. – David Sep 23 '15 at 18:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.