OSDev.org

Hi,



Yes, hardware manufacturers (and software developers) make mistakes, and this is inevitable. My job is to ensure these failures never reach the end user. If I have to spend 6 months researching, designing and implementing a system for handling displays just to ensure hardware manufacturer failures don't become my failures then that is what I will do.

But there are hardware failures that are beyond my ability to prevent from reaching the end user. For these cases the answer is not the acceptance of failure (end user configuration); the answer is the rejection of failure (clear and unambiguous dissociation - "This piece of hardware can not be supported due to manufacturer error").


Cheers,

Brendan

Hmm... a working display with a configuration fix, or an unusable display that works in every other OS? I know which one I want.

Yeah, this thread is getting kind of silly.

Hi,



Ah, the old false dichotomy. Nice.

My turn: A working display where you can't quite get the native resolution and where the admin is warned exactly why (before they purchase hardware via. compatibility lists and/or if the hardware is used); or an OS that is merely an excuse for a festering pit of configuration options that never works because it's too confusing and is always mis-configured.


Cheers,

Brendan

Neither. An OS where the display works at a non-native resolution for those who couldn't care to re-configure it, but which gives the user to override the auto-configuration and set the display to their choosing if they have the knowledge, time, and/or willingness to do so. I know for one thing that before I went blind if I'd booted up your OS and it had tried to run my buggy CRT at that awful blurry flickering resolution that Linux always defaults to then that would've been the last I'd have seen of your OS - if that's your attitude towards auto-configuration then I can tell you right now that you're never going to get enough users to make the handling of buggy displays even matter.

Hi,



The reality is that 90% of people will give up on the OS as soon as they find out they need to touch configuration (note that software developers tend to be an extremely poor statistical sample, as they have knowledge that existing users lack and are more likely to persevere).

If the auto-configuration is so broken that you ended up with a flickering screen, then auto-configuration is broken. That is the problem that needs to be solved. I have no doubt that the auto-configuration in Linux is awful because the entire OS is awful (Linux developers tend to have a "get it working for the easy cases, then let the user deal with all of the problem cases themselves" mentality, where the existence of hundreds of configuration files is used as an excuse for incompetence).


Cheers,

Brendan

Nobody will find out they "need" to touch configuration, only that they can if they want their display to use its native resolution. Given the rarity of those problem cases, 90% of people won't ever encounter configuration, and the others will likely already be aware of their display's weirdness from using it in other OSes.

Hi,



OK, so here's my compromise...

For the few people whose minds are so diseased by prolonged exposure to very poor quality OSs, who mistakenly think an OS should provide a way to for users to mis-configure things that the OS should've already auto-configured by itself; how about I implement an OS that actually works properly (as planned); but then also implement an old-school "text adventure" game where the player is given an antiquated command line prompt and has to navigate their way through a maze of cryptic "help" files and configuration files (in a simulated "/etc" directory) using virtual tools (e.g. text editor, man) that do unexpected and undiscoverable things when you press various keys; where the only goal is see how long you can continue playing before losing your sanity.


Cheers,

Brendan

How do you determine what will be considered a flickering screen?

CRT monitors always flicker, and the best you can do is set a refresh rate that makes the flickering too fast to be visible. Some users will find 60Hz acceptable, and others will require a minimum of 75Hz to prevent eyestrain, headaches, and immediate pain upon looking at the screen.

Of course, you can't know what the user thinks is acceptable without asking the user.


Funny you should compare it to a game, because that's exactly what it is. I wouldn't be where I am today if I hadn't discovered how much fun this game could be.

Hi,



You're mixing up completely unrelated things in an attempt to spread discontent.

There's normal flicker; like the flicker you get from things like fluorescent lighting and monitors (when working correctly) that happens so fast normal people don't realise, that is only noticeable when there's movement and the graphics system doesn't have motion blur.

Then there's abnormal flicker. There are only 2 cases where this occurs - either a fool screwed up video mode timings (e.g. tried to push a VGA into 800*600 despite not having a fast enough pixel clock and ending up with < 40 Hz refresh rate); and IBM's dreadful "1024*768 interlaced" video mode (where frame rate, not field rate, is only 43 Hz).



The same applies to colour depth, and horizontal and vertical resolution.

It's not a question of what is/isn't acceptable to the user; it's a question of finding the best compromise between multiple factors (refresh rate, colour depth, horizontal and vertical resolution, monitor capabilities, video card capabilities, renderer capabilities, etc); while taking into account mitigations (dithering, anti-aliasing, motion blur, ...).



Yes; for some people the more "challenging" (unusable, obnoxious, worthless) something is the more fun it is to try to "win". My OS is not designed for these people.

Most people see a computer as a tool intended to help them (not as some hideous Rube Goldberg puzzle). These are the people my OS is intended for.





Cheers,

Brendan

Most people can't see CRT monitors flicker at 60Hz; I guess it's safe to assume you can't see it either? I can, and I've met others who can. It gets painful after a few minutes.


Nor should it be! I was only trying to point out that I like your metaphor.


I will try to avoid making any comments that aren't relevant to an ordinary user.

Hi,



I can't remember ever noticing flicker on (correctly configured) CRTs; but I haven't seen a CRT for 10 years so I can't be 100% certain; and suspect people that did have problems back then were using CRTs designed for faster refresh rates (e.g. where the phosphor coating is designed to glow for 1/90th of a second) and running them at 60 Hz (where 1/90th of a second isn't long enough).

For anything that's actually still relevant (e.g. LCDs and not long dead memorabilia) there's virtually no flicker as it depends on the back-light not the refresh rate.


Cheers,

Brendan

How do you plan to find the best compromise when your heuristics don't have access to the best information?

Here's a compromise- expand your OS's official database of video modes to include the "bad" devices, and when the user goes looking for ways to fix their blurry screen, just give them a list of possible displays rather than assuming there's only one per VID/PID. All the mode info goes back "behind" your curtain, and the user only gets to pick based on the device name.

Hi,



For about 99.99% of displays that have ever existed, and 99.999% of the displays that are still actually in use, I will have the best information. For the remainder, it'd be close enough that the user probably won't even notice anyway.

If you want to continue whining; how about having a big "4+ page" whinge about my intention to not bother with ISA cards, 8-inch floppy disks, Itanium, and everything else that will be on my "never going to be supported" list that is all far more important than this?



You've got this backwards. When the admin sees the OS's "The monitor attached to head #2 on node #123 is a worthless pile of puss that should be replaced" message, chances are they'll replace the crusty old monitor with something that's 10 times better and the user won't even know why (but will be grateful).


Cheers,

Brendan

If you're going to use that defense for not supporting old monitors, then rather than complain about ISA/floppy/Itanium, I'll just ask why it's not a good reason to avoid supporting old fixed-function graphics hardware that can just as easily be replaced with something "10 times better."