riddle me this

[lauralh]

why is it that windows has to fucking install a new driver for something if I plug it into a different USB port than before?

in this case, it couldn't even install it for some god-only-knows reason. So I tried other ports till the one it had already been installed on beeped.

which is to say, pics are up.

Comments

[skiplogic.livejournal.com]

http://www.apple.com/switch

:P

[herbaliser.livejournal.com]

do I look like some kind of douchebag to you?

[skiplogic.livejournal.com]

i'm just sayin'

[tyrven.livejournal.com]

I think OS X does that as well - but it's more transparent to the user.

[skiplogic.livejournal.com]

i've actually had a ton of trouble hooking up flash media readers and getting iPhoto to use them without complaining. I just thought it was funny. I think it's pretty funny to post that link to anyone complaining about using windows because it always gets a big fuck you.

[tyrven.livejournal.com]

Totally. And, that said, as much as I hate Mac Zealots - I almost never use my faster Windows machine that I just built becuase I much prefer working on my Powerbook (although that's partially a matter of portability).

[skiplogic.livejournal.com]

I dont think I'll ever own another desktop PC. Notebooks are just too convenient. I might own another tower to run some services in my house or something at some point... but probably not.

Mac Zealots are a funny bunch. Brand loyalty of that severity is a beautiful beautiful thing (for Apple).

[skiplogic.livejournal.com]

In particular I have to re-install iphoto if I accidentally disconnect any of my readers while they're still mounted. Otherwise iPhoto will open but not admit there's a media card plugged in anywhere when you plug one in. I havent searched for support or anything, but it's totally reproducable (accidentally or intentionally).

[crumblingchaos.livejournal.com]

Its transparent to the user on Windows XP if the driver install is done RIGHT.

I would still recommend a MAC though because on a MAC the hardware for the USB ports is done well. Where on a PC they tend to have all kinds of issues. Power supply issues, etc.. USB2 is so freakin anal with the quality of everything involved. So its no surprise that people have all kinds of issues with it. Some stuff doesn't work on the front ports on machines (typically there not connected directly to the motherboard).

[tyrven.livejournal.com]

Windows has always notified me that it's installing drivers for a device when I plug it into a different USB port (even though they're already installed); it doesn't require user interaction, but it's not totally transparent.

I've never had power issues with USB on my PC, but then I've always had reputable brand machines or custom built them with reputable power supplies. The hardware modularity of PCs definitely opens up for a lot of issues. It always surprises me how cheap you can get UPSs, for instance; I don't think most people realize the impact that can have on power costs and hardware lifespan.

[stonecollector.livejournal.com]

this is why i love you.

[tyrven.livejournal.com]

I've always wondered about that. I'm pretty sure it's not actually reinstalling the driver, but it's just recognizing the device and mapping it to an existing drivers. That's really peculiar that it wouldn't recognize the device on a different port, although I've seen that happen before. Regardless, I think the UI for driver installation is pretty piss poor.

[zorbathut.livejournal.com]

Short version: It does that because the situation is surprisingly difficult.

Long version:

Device uniqueness is a pretty nasty subject. It turns out there isn't actually any way to predict what the user wants. Sometimes it's "obvious", and then sometimes you think it's obvious and it turns out to be much, much less obvious.

The problem is that often you can't tell which device is which. If you plug two identical keyboards in the computer probably can't distinguish between them. If you turn the computer off, swap the keyboards, and turn it back on, the computer probably has no way of knowing what just happened.

This gets worse because some programs *do* want to distinguish between keyboards, printers, and other devices. So Windows makes them unique the best it can - but if it can't actually distinguish between them, there's a limit to what it can do. And you could create some really weird entirely unexpected situations if it tried to "guess".

Imagine you have one printer, and a program set up to use that printer. You disconnect the printer and connect it to a different USB port. You'd want it to see the same printer, right?

But imagine instead you disconnect the printer and plug an entirely different printer into a different USB port. To the computer, this is the same situation that we had before - so if it does the "right thing" in the previous situation, it now does entirely the wrong thing in this situation.

Windows's solution is to go with predictability rather than trying to guess what the user wants. It might not get the answer right as often, but at least someone who knows what they're doing will be able to tell what the internal logic is for every case.

Note that it never actually uninstalls devices, just marks them as "not installed" - if you run Windows in safe mode and go into the device manager you'll see every device it has ever installed. Sometimes that can be quite a hell of a lot of devices.

If you read this far, you now know far more about device management than you really wanted to. :)

[tyrven.livejournal.com]

See, I thought each USB device had a unique ID (like a product ID not a serial number) and that it used this ID to identify what driver to install. Based on that my assumption is that if I unplug my mouse and put it into a different port that it'd wire it up to the existing driver and the process would be relatively transparent. Now, if Windows then reset the user properties for that port/hardware ID mapping that would make total sense to me (since it doesn't know if it's a new mouse or what). Basically this isn't that different from how it works now - just that it wouldn't notify the user that it's "installing" drivers.

Of course, if I think of the port address as an IRQ or hardware address then it makes more sense to me.

Does windows share this behavior if the USB device is dangled off another USB device (such as a hub or a USB device with a passthrough port)?

[zorbathut.livejournal.com]

I suspect that USB hubs and such create new unique IDs. The USB "ID" would presumably be the product ID of each device in the chain, including the last one, and that's what it uses to determine uniqueness of new devices.

I'm not 100% sure about that. It's just how I would do it.

[ashlin519.livejournal.com]

Like the boots

[herbaliser.livejournal.com]

THANK YOU FOR NOT BEING A FUCKING COMPUTER GEEK!

[ixzist.livejournal.com]

it could be a poor coded driver, rather than a operating system problem

[toastednut.livejournal.com]

open up device manager in the computer management console. check out universal serial bus controllers. likely you have more than one listed in there and as such, a driver will be installed for each controller (port).

[herbaliser.livejournal.com]

that's not exactly a solution

[toastednut.livejournal.com]

no, it's the reason. the solution is to not use windows, duh.

<-- os x

[herbaliser.livejournal.com]

Still not a douchebag!

[toastednut.livejournal.com]

you just don't want to hear it...