One of the common problems when Windows Vista was released was that of missing or non-working drivers. Microsoft massively reworked many of Windows’ internal systems and frameworks, meaning lots of drivers broke, with most of them needing major work, and some even needed to be rewritten completely. Apparently, Microsoft didn’t communicate this well enough with its hardware partners – or the partners were lazy, who knows – because many devices failed to work with Vista during its early months of being out in the wild. Microsoft is trying to keep this story from repeating itself, saying that everything that works on Vista should work on Windows 7. To gain a little more insight into this problem, Microsoft gave out some very interesting figures regarding driver installation failure rates.
The below figures come from Microsoft’s Online Crash Analysis and Customer Experience Improvement Program tools, and show the success/failure rates of driver installations during September 2008, on Windows Vista with Service Pack 1 installed.
These figures are interesting, but at the same time, a tad bit on the useless side of things because there’s no way to make any comparisons. How does this compare to Windows XP? Windows Vista pre-SP1? Ubuntu? OpenSUSE? Please note that we are talking driver installations here, so drivers that are already part of the base OS are not taken into account. Microsoft has stated that anything above 3% is not good.
Then there is the problem of not really knowing what’s causing the failures. Bugs in Windows Vista? Buggy drivers? People trying to install XP drivers on Windows Vista? Crappy driver installation programs? Messy vendor websites, leading to people picking the wrong drivers? Simply the fact that there are no Vista drivers for that piece of hardware? There could be so many things going wrong here, that it’s really impossible to blame one single factor – these figures are too skinny for that.
The solution already exists. Microsoft has been frantically at work adding drivers to Windows Update, but apparently, this turns out to be difficult. Right now, they are adding about 50 drivers a week, Microsoft’s Chris Matichuk revealed.
Matichuck also stated that the company has a goal of making sure that 90% of systems are fully covered via drivers included on the Windows 7 disk and those in Windows Update. “We’re tracking this as a metric, by the time we hit RTM we want to have at least 90% covered. It’d be great if we could get 95% or higher, but 90% is the goal we’re going to go for.” For all those for-the-sake-of-argument “grandmothers”, let’s hope they succeed.
is good for MS too!!!
from my experience… the most universal pain in the ass to install are printer drivers… they’re terrible (to not say something worse) and, even if install correctly, they perform badly, slow controls, slow response time… and ugly as they can get!
fortunately printer drivers don’t give much problems as motherboard (specially IO) and video drivers that may prevent you from actually booting the system (or repairing the system…)… but for those I have to admit that the quality got a LOT better since real x64 drivers came (that includes ia32 drivers too…)
I hope open HW specs gets more popular soon, this means better drivers for everyone…
A quote from your fellow OSAlert reader, Kroc:
http://camendesign.com/blog/stop_writing_software
Edited 2008-11-10 16:02 UTC
This is the very reason why I refuse to buy a new printer. Mine broke like 18 months ago, but the amount of total and utter crap software and drivers the likes of Canon and HP put out simply stops me from buying their crap.
When I need to print, I either do it at university, or at my parents’. Until these guys learn to give me a driver download that’s under 5MB in size (preferably under 1 MB in size) I will not buy a new printer.
Does Windows not offer some sort of generic postscript driver? You could always just pick LaserJet 3, or LaserJet 1200 (PS) from the built in drivers, or something. You might not get one button, triple-sided, multi-duplex, ultraviolet printing with an automatic spin cycle every 20 pages, but it should work pretty well for what you really need.
And yeah, HP wants to get all intrusive in Linux, too. I just pick generic postscript and go on.
How? I can’t find anyway to install a printer with Vista and pick the wrong driver you a know a similar HP would work but Vista won’t let you use it.
I usually buy an HP and just use the Windows provided LaserJet 4 driver. I have never found an HP LaserJet that it wouldn’t work with.
I never by multifunction.
you can download the driver and not use the .exe program to install all of HP’s or Cannon’s garbage software. my print server for my work has all of it’s drives about 4 megs or under just because when i download the file frmo the site i just tell windows where the folder with the .ini file is and its dll’s that it needs. I instal the printer manualy in windows. when you do that its a rather small install.
You can take out the cruft out of the drivers. At least it used to be in HP’s Mac drivers case. Takes some time, but eventually you strip that onion down to just the bare drivers. Those even have their separate installer. You may try downloading the driver package for your prospective printer and see if this still applies.
Kroc is spot on.
Same with some Wifi drivers where there seems to be 2 paths…
1. Install the driver using the auto run application on the CD and have some crap installed.
2. Let Windows search for the ini on the CD and just have the driver installed without all the crap.
Why can they not have just have the simple driver and a PDF or RTF on the CD to tell you what to do.
I’ve been saying this for years…
I don’t need a CD full of crap just to get a mouse working!
And once I finally managed to remove flash from XP, Yahoo Messenger installed it back without asking…
The only time I saw driver installing in Vista it was like, the first one:
“This driver requires Windows XP or later.”
and the 2nd one even worse:
“The software has detected that the operating system is not Microsoft Windows.”
While I agree with almost everyhthing Kroc says one thing is not correct.
Once upon a time, in a galaxy far away (also known as late 80’s, early 90’s) Norton made some solid products, chiefly Norton Commander.
Of course, all that good software came to an end when Symantec showed up on the scene.
McAfee was also good back then, before they became Network Assholes or whatever the hell they called themselves.
Edited 2008-11-10 18:22 UTC
How does that knowledge help consumers buying a new Dell right now?
It doesn’t. So put your memories safely back in the box.
As it stands, Norton & McAfee are terrible, atrocious, unbearable products.
Yes, I think when comparing drivers issue one should keep in mind the result for the user:
1) disk driver, motherboard, video: the user can’t boot: very annoying.
2) network/modem driver: the user can’t connect to Internet: annoying especially when you know that the solution is probably available on Internet!
3) other drivers you just loose the functionnality of the hardware but the computer work and have access to Internet, so you can search if anybody has a solution.
What I want to know is how to get Vista Ultimate 64 to use my old HP 950C that is networked via CUPS. HP’s docs just say “it’s built into the OS”, but the OS doesn’t seem to detect it very well…
Any tips?
You either have to share the printer via Samba (ew) or use XP’s built-in IPP (Internet Printing Protocol) support:
http://www.owlfish.com/thoughts/winipp-cups-2003-07-20.html
When I still had an XP machine, I had it connected to a CUPS printer running on FreeBSD through this technique, worked great.
I have it installed on Win2k with IPP, which is what I want to use with Vista64, but it doesn’t detect the printer and I can’t find the driver in its listing.
Yet, HP just tells me that the “driver is included with the OS” so I can’t download a driver either or go the “Have Disk” route.
Eep, drivers might be your issue; I skimmed and didn’t notice Vista 64 before. There are a lot of missing drivers for Vista 64 still.
Vista should have a generic PCL or PostScript driver (depending on what mode your printer is expecting) that will work if you follow the IPP setup instructions…
What MS is essentially doing with posting these non-comparable figures and other publications about the drivers issue, is to blame others.
My perception is that MS is trying to tell “Our Vista OS is great, but was broken by third-party.” What MS has to ask themselves is, why is Windows driver quality that bad? Drivers in other OSes like Linux almost always just work fine, out of the box, w/o any hassle.
It is the same with user-space software. The Windows platform suffers most from its heterogenically trashing Application base. Trashing in a sense of DLL-hell, privilegue abuse, etc.
MS needs to realize that blaming others won’t help. The vendors of both drivers and applications are just trying to be nice citizens in a specific environment. MS makes the rules and gives example (with own applications), others just adhere to them.
One recent example regarding drivers: MS presented this new “my hardware” platform where for example a digital camera vendor could include his own crappy bling-bling interface and stuff. This is exactly what MS should teach them not to do. This bling-bling stuff is a real usability killer. The User should exactly _not_ be made aware of what kind of hardware he has. It should just work for him.
One thing that’s always really ticked me off is the double standard.
If hardware doesn’t work with Windows, it’s the hardware manufacturer’s fault for not writing a driver.
If the hardware doesn’t work with {Mac OS X, Linux, BSD, BeOS, Solaris, etc.} it’s {Apple, The Linux Community, The BSD Community, Be, Sun, etc.}’s fault.
What would be nice is if hardware followed specs (impossible for “new” devices, but those are rare) and you could just plug it in and get basic functionality with a built-in driver. Want to access advanced functionality or get the best performance? Then you install a driver from the manufacturer.
Its simple. From the users point of view, they don’t care WHY it is not working, or whose “fault” it is, they just care that the thing that they bought doesn’t work right on operating system x. From a technical point of view, there are many reasons why something could or could not work, so saying it is because of a vendors driver is a fine thing to say.
I don’t see how there is any double standard, it works the same for everyone.
The double standard is that MS is absolved of blame for hardware not working. Other OS vendors are not.
Absolved by who? Their marketing department? Popular opinions about vista show they aren’t absolved at all.
Microsoft requiring the hardware manufacturers to write drivers for their hardware for Windows (and quite often refuse to release Linux drivers for the same hardware) is a practice that Microsoft themselves introduced.
Some time ago the norm would be for the OS vendor to write software (including drivers) for the OS, from specifications published by the hardware manufacturer. This sensible scenario of course allows a situation where many different OSes can compete in the marketplace, since any software vendor may write drivers.
Microsoft turned all that around completely, I would suggest in yet another attempt to get “exclusivity” for Windows. “This hardware only works on Windows”. The idea is to try to force the consumers into buying this product over that one.
It is nice to see Microsoft’s own anti-competitive practices turn on them in this way, and leave Vista with a dearth of working drivers, due to manufacturers not writing new drivers for equipment they no longer sell (amongst other reasons). Why should Microsoft get to sell a new OS on the back of efforts of manufacturers to write new drivers for zero recompense, BTW?
Microsoft’s business model of selling closed-source binary-only copies of core OS and application software, and getting hardware OEMs to write drivers, is under quite a bit of strain here.
I feel that printers are amongst the worst in this situation because they most often remain in use long after the model is no longer made. Video cards (for example) are most often sold with a machine and never replaced, and the machine’s OS is also never upgraded. So in contrast the way that printers are bought and used hightlights the weaknesses of Microsoft’s business model more dramatically than other hardware.
It will be interesting when Microsoft tries to move exclusively to a 64-bit OS. They will be in such a mess with lack of hardware drivers it won’t be funny.
BTW: are there any doubters left about the article of a little while ago that claimed that “Linux has more drivers than any other OS ever”? For Linux, where the source code is available, a 64-bit driver is very often just a simple re-compile away.
Edited 2008-11-10 23:38 UTC
That was a very, very long time ago, long before microsoft. Someone has been feeding you bad history.
Not at all. All major CPUs, for example, to this very day, have a published specification for their instruction set. All software vendors of compilers use this information to write their compilers in order to target specific CPUs. It is not practical to try to introduce a new CPU into the marketplace today without publishing the instruction set specifications for compiler writers. Just not on.
http://en.wikipedia.org/wiki/GNU_Compiler_Collection
GCC is the premier compiler in the world today. It supports more architectures than any other. It arguably has compiled more of the code (in terms of unique lines of code, not binary copies of executables) that is actually running somewhere right now than any other compiler.
http://en.wikipedia.org/wiki/GNU_Compiler_Collection#Architectures
All because, to this very day, the manufacturers of CPUs publish their instruction set specifications.
As recently as the introduction of USB 1 in 1996, then again for USB 2 in 2001, and now (2008) yet again for USB 3, the hardware designers (in this case, largely Intel) publish the specifications against which OS authors can write drivers.
http://en.wikipedia.org/wiki/Usb#History
Some more examples:
http://en.wikipedia.org/wiki/AT_Attachment
http://en.wikipedia.org/wiki/Scsi
http://en.wikipedia.org/wiki/IEE_802.3
http://en.wikipedia.org/wiki/Vesa
I’m sorry, but these examples are not “from long ago”, nor are they in any way “bad history”. It appears rather that YOU are too enveloped by “Microsoft think” to be able to actually see the reality of things.
Edited 2008-11-11 22:30 UTC
I didn’t take objection to you saying open specs for hardware didn’t exist. I did take objection to you saying nothing but open specs to hardware existed before microsoft.
I also take exception to that. I’m not a microsoft guy, I’m a .net guy (or more specifically an ALT.NET guy) who doesn’t have the simplistic groupthink ideas about the company that tend to propagate places like this. I work every day with people who are enveloped by “Microsoft think” (which in reality is probably really different from what you think it is), and I spend a great deal of time both with them and in the greater developer community in my city trying to get people to take off the “redmond only” blinders.
I’m pretty passionate about .net, if you count the DLR it is hands down my favorite platform at the moment. I am probably more passionate about ruby, agile forms of software development, software architecture, and alternative operating systems.
Quote please?
Don’t bother … I’ll quote myself.
Gee, I don’t know about you, but to me there is a significant difference between an actual claim saying “the norm would be for the OS vendor to write software for the OS, from specifications published by the hardware manufacturer” and your allegation that the claim was “you saying nothing but open specs to hardware existed”.
Your resposne is a strawman … I made no such statement as you claimed.
Try again.
Edited 2008-11-12 01:35 UTC
I think I now see where this confusion has come from.
Here is (slightly more of) what I originally said:
OK … what the “completely” refers to is “turned around”. This means the situation now is the opposite to what it once was.
OK, it was claimed to be: “Some time ago the norm would be for the OS vendor to write software (including drivers) for the OS, from specifications published by the hardware manufacturer.”.
“Turned around completely” was meant to indicate that “the norm now would be for the OS vendor to expect the hardware manufacturer to write closed source drivers from secret, internal, unpublished specifications”.
“Completely” wasn’t meant to refer to all hardware … it was meant to refer to “completely opposite”. “Norm” means “normally” in both instances … not “in every case” but just “normally”. I can however see how my meaning wasn’t very clear, so I shouldn’t have snapped back at you.
In support of what “the norm” has become now, I would refer readers to the trouble that ATI seems to be having right now to let simple information about the registers on their chipsets to be published these days:
http://www.phoronix.com/scan.php?page=news_item&px=Njg0Mw
“In Matthias Hopf’s blog post he has stated that they are getting close and there’s just very few registers they now don’t understand. ”
…
“The R600 3D documentation has yet to be released due to legal troubles and such, but at least we have mode-setting support.”
It is ATI’s own hardware for goodness sake. Why on earth should there be any “legal troubles” associated with ATI publishing the details of the chipset register map of their own chips?
I find this quite annoying, because one of my systems happens to have an R600 series ATI graphics chip.
… I wonder what other company might be getting in the way here, BTW? Hmmmm?
Edited 2008-11-12 02:55 UTC
Consumer advocating and advice. Business and home owners. When you ask for a quote or you go to your local computer shopping center make sure the sales person can answer these two questions: Does it have an open source driver. Does it work with Macintosh, Windows Vista, Linux, Sun, etc. etc.
Why! You ask. Options. Potential.
1. Mac, Windows and Linux support gives you the option to switch from your initial choice to one of the others.
2. Multi OS support leaves you in better position to bargain on price.
3. If a driver is open and written for a Mac or Windows then someone will rewrite it for Linux. Open for Linux then re-written for Mac and Windows. etc. etc.
4. Got a good old piece of hardware that you’ve replaced. Give it to someone in need or send it overseas. Its really sad that so many good computers are scraped so Microsoft and its hardware buddies can replace your current setup. Open source drivers makes this old hardware really valuable.
Note: there are a lot of places in this OLD world where a years salary is no more than $200 US. Your old Pentium II would be a big step up from NOTHING!
5. You would think that old computers, printers and etc. that still worked after 10/15 years would be a great selling point. You’d think that instead of destroying them IBM, DELL and etc. would be sending their almost new returned for environmental reasons computers somewhere they would do the most good. Still working computers would be great advertisement if they just reached the home or business of someone that couldn’t even afford the shipping costs. New users, new programmers. New customers.
Yeah, collect information, MS! do it and ignore these bugs in the same time, neglecting to fix it.