(Subtle) Hardware Defects / Software Defects
One thing that's been bothering me recently is the fact that software errors and instability can be caused by slightly damaged (usually old) hardware. As I've noticed, inconsistent PSU voltages can have detrimental effects on motherboards, hard drives, etc. At least with a PSU, one can sort of determine it's reliability by just actually measuring the voltage compared to the specifications.
But I've seen some things like explanations on how the microscopic aluminum or copper wiring in microprocessors can fracture over time due to heat fluctuation and ESD. There is actual evidence to back that up. But the theory that if the computer hangs or an application crashes, then the hardware's at fault, I either don't buy, or just don't want to admit to.
Because then, if every fatal exception dialog box was really due to ESD corrosion in the transistors of your processor, then how would we distinguish between hardware error and software bugs? I know we can still do so for the most part but the mere existence of that ambiguity bothers me.
I've even seen some more ridiculous claims. At the library, I got distracted and began reading "Repairing your PC for Dummies" for a nice laugh. I found an interesting term in the book - "tired RAM" - which exists nowhere else (Google it). The narrator explains that as old PCs get worn over time, it's due to their RAM becoming "old". This claim didn't sit well with me. If there is real ESD damage to RAM, then it's bad RAM. That's why we have memtest86+.
But therein lies the uncertainty. There isn't a bootable benchmark to run all circuits in a PC to full capacity to fully eliminate the discrepancies on old CPUs and other PCB boards, expansion cards, etc., (is there?). How can we know if the IDE controller is damaged based on symptoms and not the power supply, or RAM, or a corrupt BIOS. (This was the case with my old Gateway that couldn't boot to a Windows 98 CD. It displayed random CP437 chars and beeped repeatedly. It was the motherboard, apparently.). I guess what I wish for here is: a true bootable 100% effective benchmark, which isn't going to happen, and mostly just more people's opinions regarding this, input, etc. What do you think? How could we draw an infinitely precise line between software defects and hardware defects?
But I've seen some things like explanations on how the microscopic aluminum or copper wiring in microprocessors can fracture over time due to heat fluctuation and ESD. There is actual evidence to back that up. But the theory that if the computer hangs or an application crashes, then the hardware's at fault, I either don't buy, or just don't want to admit to.
Because then, if every fatal exception dialog box was really due to ESD corrosion in the transistors of your processor, then how would we distinguish between hardware error and software bugs? I know we can still do so for the most part but the mere existence of that ambiguity bothers me.
I've even seen some more ridiculous claims. At the library, I got distracted and began reading "Repairing your PC for Dummies" for a nice laugh. I found an interesting term in the book - "tired RAM" - which exists nowhere else (Google it). The narrator explains that as old PCs get worn over time, it's due to their RAM becoming "old". This claim didn't sit well with me. If there is real ESD damage to RAM, then it's bad RAM. That's why we have memtest86+.
But therein lies the uncertainty. There isn't a bootable benchmark to run all circuits in a PC to full capacity to fully eliminate the discrepancies on old CPUs and other PCB boards, expansion cards, etc., (is there?). How can we know if the IDE controller is damaged based on symptoms and not the power supply, or RAM, or a corrupt BIOS. (This was the case with my old Gateway that couldn't boot to a Windows 98 CD. It displayed random CP437 chars and beeped repeatedly. It was the motherboard, apparently.). I guess what I wish for here is: a true bootable 100% effective benchmark, which isn't going to happen, and mostly just more people's opinions regarding this, input, etc. What do you think? How could we draw an infinitely precise line between software defects and hardware defects?
Comments
You get headaches right? I get loads. Sometimes it's due to tiredness, dehydration, a cold, or a million other things. Sometimes I just get a headache and never know why -- errors with machines can be similar.
I could see RAM "going bad" over time -- Possibly in the same way HDDs do, all HDDs are shipped with "more" capactity than they report to the BIOS/whatever, so that if some of those go bad, they can be marked as so and the disk won't lose any space. You only notice that a drive has issues when there are *hundreds* of them.
There are so many variables when it comes to computers -- new and old, that sometimes shit happens. And sometimes you can't explain it.
Like all things sometimes shit just happens and we try and give a reason for it even though really it's probably a million things factoring into why.
As such there can never be a 100% true hardware diagnostic tool bar from the ones used during the building phase, and the part may pass the basic test but fail down the line.
I guess I just have a hard time accepting the "shit happens" philosophy with PCs. Just the idea that an error that a year ago I would consider a pure software defect might actually be bad RAM, a slightly dead HD, and maybe a bad PSU slowly baking the rest of the computer. :P well maybe not to that extreme, but it's an interesting thought.
As you know, I love old PCs. Old Pentium I through IIIs. They're just more fun than new ones, in a way. And slot loading CPUs was/were really convenient. And a key part of my disbelief in obsolescence lies in that an old PC should perform as well as it did when it was new when the software is configured/reinstalled correctly. I've said over time that "the physical hardware doesn't just get slower over time. Most of the time it's the software, which is inefficient and disorganized." have you *seen* an old Win95 PC? Oftentimes, depending on whose it was, it has like 8 useless startup programs, loads of malware, etc.
I always believed that if you simply wipe the drive, all software issues disappear.
People have a hard time accepting that with anything.
"Dammit, my Coke exploded all over the floor."
vs.
"Sorry bro, drove your BMW into the lake. But shit happens, right?"
hehe
Lol, what is worse is you lamborgini is in the lake
I figured someone would comment on that. I have no idea why I posted it.
Lamborghini = Audi.