x86 complaints
The x86 platform, and all the hardware associated with it, is severely and depressingly rotten. When it was first put into service twenty years ago it was a good system, but I'm getting pretty sick of the continued desire to keep computers in the eighties.
Why hasn't the BIOS been updated? Why does my PC boot from an EEPROM running a piece of software written in 1987 or 1992 that has been hacked with a hex editor to add rudimentary configuration options for the anachronistic hardware it's supposed to be managing? Why does my BIOS still let me enable a "memory hole" and set options for OS/2? Why have these outdated options not been pruned? Why can't my BIOS run real memory checks, test system components, format hard drives, accept NDIS network drivers and provide a rudimentary web browser so I can download RAID drivers to make it possible to install Windows on my PCI ATA controller? There's nothing stopping this from being a reality except that nobody has asked for it, even though everyone would want it if they knew it was available. I know it's possible because I've seen some of these extensive special features on rare IBM laptops, and I know it would be easy to implement in practice because you could simply provide a customized firmware distribution of QNX, which is perfectly capable of doing everything I listed while booting from a 1.44MB floppy disk.
If you want to see more of what the BIOS could be, look at the Mac firmware. It supports high-resolution graphics, mouse and keyboard input, and a graphical, intelligent boot menu, among other things. At the very least we should have that on the x86 platform. If that last sentence is making you feel strange, maybe it's because your mind is trying to remember why that seems so eerily familiar. It's because Apple actually IS using their own, superior BIOS on the x86 platform, proving that the textmode Phoenix and Award crap isn't necessary at all. Text mode must die; we should never see it again during startup. It is the first step, and an important psychological one, on the path to a better, more "enlightened" PC.
Why are edge connectors still being used on expansion cards and RAM? These weren't a good idea at the outset except that they saved on assembly costs (one less connector to solder) which are certainly moot at this point. The only expansion card system that ever attempted to use a real connector was NuBus, which of course failed because it was only used on systems with tiny market share. I can't think of any reasonable explanation for PCI Express maintaining the same style of connector as the fifteen-year-old standard it's replacing, the designers of the standard should have taken advantage of crossing an incompatibility boundary to improve on the old standard at least by putting a real connector on it. VME is actually a very nice connector, NuBus had it right.
For that matter, I don't understand why PC hardware never comes in an actual finished package. Is it clear to anybody else that static damage to RAM simply wouldn't ever happen if it was packaged in the same kind of protective case as we would expect from any other modern, well-made electronic device? Why are memory sticks and PC expansion cards the only electronic devices we're comfortable buying as a raw printed circuit board with no case? To put this in perspective, the game cartridges for the Nintendo Entertainment System had a case with a plastic shroud around the connector to make it harder to brush your fingers across the contacts. Even the Atari had that. Why are we making faster and faster memory, adding all sorts of bells and whistles, adding heatsinks and fans, then not making any attempt to protect the stuff from being destroyed the moment it's taken out of the box?
Hard drives have exposed PCBs as well, which has never been well-explained. It only made a little sense when hard drives were new, now it makes none. It's just asking for trouble. For a short period Seagate was adding the "SeaShield" to their drives, along with a massive sticker (it covered the entire bottom side of the drive) explaining what it was for, but they dropped that after a few years. This truly confuses me; why is there an exposed circuit board? It's completely pointless, yet the only explanation for it being there is that the hard drive manufacturers are making an effort to expose it. At this point it is so easy to design and build a device that obscures all of its sensitive internal electronics that I can only assume that the hard drive manufacturers want people to be able to see the PCB.
Why are PC cases still the sheet-metal crap they've always been? We wouldn't accept this kind of cheesy, flimsy, slapped-together construction in anything else. Consumers want accessories to snap smoothly into their parent devices with an audible "click" and without any screwdrivers. Why doesn't this expectation of well-designed mounting standards extend to PCs? Why do PCs contain screws? Why isn't everything mounted with spring-loaded quick-releases?
I know there have been a lot of PC cases and rail systems that attempt to improve on this with hard drives and CD-ROMs, but - ironically - so far the only system I've seen that's even close to convenient is the system used in Dell package-deal PCs. You install four proprietary screws into a hard drive, then it slots into the rails in the system and clicks into place. No secondary locks or complex mechanisms, they aren't necessary. There's one plastic lever you pull to release the drive and that's that. And it works with CURRENT hard drives, which is nice for adoptability. But I argue that the entire screw-based mounting system needs to be replaced with a better design. Why are we still accepting this primitive standard?
Why does installing a brand-new hard drive in your PC take more than a minute? And before you respond, consider this: if there's a good reason that PCs are tedious to work on, why doesn't that reason extend to devices which are no less complex and no more standardized, such as game consoles? The hard drive on a 360 doesn't require any tools to install, and PCs aren't highly customized in general. The vast, vast majority of PC cases have the hard drive mounting bays in exactly the same location. There's no reason those couldn't be replaced with prewired, tool-free receptacles. And to avoid any possibility of compatibility issues and customer backlash, don't replace the current system, just augment it. Continue to use the same dimensions for drives, leave the standard mounting holes, just add new holes to accept spring-loaded tabs. There's absolutely no disincentive for hard drive manufacturers to do this. They could modify the molds for their hard drive chassis to add the notches to support better mounting systems and never feel any loss from the experience if nobody adopted it. Only everybody WOULD adopt it. The only reason it hasn't happened is because the manufacturers haven't added it to the drives, which has only happened because nobody is asking for it. And the reason for that is that everyone is convinced that computers have to be terrible.
That's the crux of this problem. It's the Unix mentality: The more miserable something makes you, the better it must work. The problem is that this principle is outdated. It was accurate twenty years ago, but doesn't need to be accurate now, and the only reason it is is because of this ridiculous catch-22. People expect good software and hardware to be ugly, and so there's no drive to make either more pleasant to use, so the powerful hardware and software that does get produced is ugly. And so on, and so on. If people would wake up and realize that we've progressed enough in the past couple decades to make computers easier and nicer to use without sacrificing any functionality then manufacturers and developers would be falling over themselves to improve usability. If manufacturers took the initiative and decided to produce computers that weren't unpleasant to use then people would learn to trust usable products.
It's too bad this won't ever change. Sometimes I feel like I'm the only person who realizes how much things suck and how much better they could be.
Why hasn't the BIOS been updated? Why does my PC boot from an EEPROM running a piece of software written in 1987 or 1992 that has been hacked with a hex editor to add rudimentary configuration options for the anachronistic hardware it's supposed to be managing? Why does my BIOS still let me enable a "memory hole" and set options for OS/2? Why have these outdated options not been pruned? Why can't my BIOS run real memory checks, test system components, format hard drives, accept NDIS network drivers and provide a rudimentary web browser so I can download RAID drivers to make it possible to install Windows on my PCI ATA controller? There's nothing stopping this from being a reality except that nobody has asked for it, even though everyone would want it if they knew it was available. I know it's possible because I've seen some of these extensive special features on rare IBM laptops, and I know it would be easy to implement in practice because you could simply provide a customized firmware distribution of QNX, which is perfectly capable of doing everything I listed while booting from a 1.44MB floppy disk.
If you want to see more of what the BIOS could be, look at the Mac firmware. It supports high-resolution graphics, mouse and keyboard input, and a graphical, intelligent boot menu, among other things. At the very least we should have that on the x86 platform. If that last sentence is making you feel strange, maybe it's because your mind is trying to remember why that seems so eerily familiar. It's because Apple actually IS using their own, superior BIOS on the x86 platform, proving that the textmode Phoenix and Award crap isn't necessary at all. Text mode must die; we should never see it again during startup. It is the first step, and an important psychological one, on the path to a better, more "enlightened" PC.
Why are edge connectors still being used on expansion cards and RAM? These weren't a good idea at the outset except that they saved on assembly costs (one less connector to solder) which are certainly moot at this point. The only expansion card system that ever attempted to use a real connector was NuBus, which of course failed because it was only used on systems with tiny market share. I can't think of any reasonable explanation for PCI Express maintaining the same style of connector as the fifteen-year-old standard it's replacing, the designers of the standard should have taken advantage of crossing an incompatibility boundary to improve on the old standard at least by putting a real connector on it. VME is actually a very nice connector, NuBus had it right.
For that matter, I don't understand why PC hardware never comes in an actual finished package. Is it clear to anybody else that static damage to RAM simply wouldn't ever happen if it was packaged in the same kind of protective case as we would expect from any other modern, well-made electronic device? Why are memory sticks and PC expansion cards the only electronic devices we're comfortable buying as a raw printed circuit board with no case? To put this in perspective, the game cartridges for the Nintendo Entertainment System had a case with a plastic shroud around the connector to make it harder to brush your fingers across the contacts. Even the Atari had that. Why are we making faster and faster memory, adding all sorts of bells and whistles, adding heatsinks and fans, then not making any attempt to protect the stuff from being destroyed the moment it's taken out of the box?
Hard drives have exposed PCBs as well, which has never been well-explained. It only made a little sense when hard drives were new, now it makes none. It's just asking for trouble. For a short period Seagate was adding the "SeaShield" to their drives, along with a massive sticker (it covered the entire bottom side of the drive) explaining what it was for, but they dropped that after a few years. This truly confuses me; why is there an exposed circuit board? It's completely pointless, yet the only explanation for it being there is that the hard drive manufacturers are making an effort to expose it. At this point it is so easy to design and build a device that obscures all of its sensitive internal electronics that I can only assume that the hard drive manufacturers want people to be able to see the PCB.
Why are PC cases still the sheet-metal crap they've always been? We wouldn't accept this kind of cheesy, flimsy, slapped-together construction in anything else. Consumers want accessories to snap smoothly into their parent devices with an audible "click" and without any screwdrivers. Why doesn't this expectation of well-designed mounting standards extend to PCs? Why do PCs contain screws? Why isn't everything mounted with spring-loaded quick-releases?
I know there have been a lot of PC cases and rail systems that attempt to improve on this with hard drives and CD-ROMs, but - ironically - so far the only system I've seen that's even close to convenient is the system used in Dell package-deal PCs. You install four proprietary screws into a hard drive, then it slots into the rails in the system and clicks into place. No secondary locks or complex mechanisms, they aren't necessary. There's one plastic lever you pull to release the drive and that's that. And it works with CURRENT hard drives, which is nice for adoptability. But I argue that the entire screw-based mounting system needs to be replaced with a better design. Why are we still accepting this primitive standard?
Why does installing a brand-new hard drive in your PC take more than a minute? And before you respond, consider this: if there's a good reason that PCs are tedious to work on, why doesn't that reason extend to devices which are no less complex and no more standardized, such as game consoles? The hard drive on a 360 doesn't require any tools to install, and PCs aren't highly customized in general. The vast, vast majority of PC cases have the hard drive mounting bays in exactly the same location. There's no reason those couldn't be replaced with prewired, tool-free receptacles. And to avoid any possibility of compatibility issues and customer backlash, don't replace the current system, just augment it. Continue to use the same dimensions for drives, leave the standard mounting holes, just add new holes to accept spring-loaded tabs. There's absolutely no disincentive for hard drive manufacturers to do this. They could modify the molds for their hard drive chassis to add the notches to support better mounting systems and never feel any loss from the experience if nobody adopted it. Only everybody WOULD adopt it. The only reason it hasn't happened is because the manufacturers haven't added it to the drives, which has only happened because nobody is asking for it. And the reason for that is that everyone is convinced that computers have to be terrible.
That's the crux of this problem. It's the Unix mentality: The more miserable something makes you, the better it must work. The problem is that this principle is outdated. It was accurate twenty years ago, but doesn't need to be accurate now, and the only reason it is is because of this ridiculous catch-22. People expect good software and hardware to be ugly, and so there's no drive to make either more pleasant to use, so the powerful hardware and software that does get produced is ugly. And so on, and so on. If people would wake up and realize that we've progressed enough in the past couple decades to make computers easier and nicer to use without sacrificing any functionality then manufacturers and developers would be falling over themselves to improve usability. If manufacturers took the initiative and decided to produce computers that weren't unpleasant to use then people would learn to trust usable products.
It's too bad this won't ever change. Sometimes I feel like I'm the only person who realizes how much things suck and how much better they could be.