You’d think this headline was a joke, but sadly, it’s not. It’s the real headline of an article posted on SeattlePI.com in the blogs section. The core of the story is that a man couldn’t get his printer to work with Windows Vista, and ultimately, with the help of a Microsoft test manager, solved the problem warranting a follow-up article. The comedy here, of course, is in reading what went wrong and wrapping your brain around why engineers didn’t forsee such a thing happening.
Man Gets Windows Vista to Work With Printer
About The Author
Adam Scheinberg
Technology Executive • Web Developer • Father • Foodie • Music Snob • OS enthusiast
Follow me on Mastodon @[email protected]
1. So the first time I installed Vista x64, Visual Studio decided, half-way through installation that it turn off wow64 folder-redirection. Resulting in a horribly broken install that there was no hope of un-installing. The vista folder-redirection schemes sounded a bit dubious to me, even before I experienced this.
2. My Dell printer is actually a rebaranded Samsung printer. The Dell drivers are about 5 times the [download] size of the samsung ones, and *reduce* the functionality of my printer, (the Dell drivers add extra margins to printed pages that are demonstrably software-imposed)
Got a laugh out of this. “Man gets printer to work with Linux” would have been a common story back in 1996 if the world had been paying attention then. Maybe someday Vista will catch up. 😉
You can’t blame an OS for badly written drivers and applications. It’s like blaming Chevrolet for a flat tire…
Also, if a user can’t get his printer to work in Linux, it’s not newsworthy – it’s pebkac (“Oh geez he installed the non-pthreads enabled version of the driver, why can’t people read documentation?”). On the other hand, Vista should have handled two different versions of the same driver smoothly, of course, because the user isn’t meant to know what OS he’s running (as long as it’s a Microsoft OS).
And you gotta be kidding me if you really think open source alternatives are decent printing solutions. Gotta love CUPS’ structure (or lack thereof).
Don’t you love such an open-minded, well informed and unbiased community?
About as much as I love your open-mined, well informed and unbiased response.
There is absolutely nothing wrong with the design of CUPS or it’s functionality. The Gutenprint project also produces some very high quality printer drivers that work well for a vast majority of common and not-so-common printers.
Now Foomatic on the other hand, and the general unwillingness of printer driver developers to standardise around CUPS? That’s a mess that I would have agreed with you on. As it stands it just sounds like you don’t really know what you’re talking about, though.
To be fair, he was installing XP drivers on Vista. You can’t guarantee that drivers written for an OS released in 2001 will work on an OS released in 2007.
Then again, how is the average user supposed to know that. Vista should also keep better track of the bloody files it misdirects!
Well, I agree with you here … the problem is the user. You can’t expect drivers for an OS to work on another.
But I have to disagree about the part where the average user shouldn’t know what OS is running. It’s one of the most basic things to know about your computer. Clearly, someone like that has no business installing drivers.
I can’t feel sorry about him. If he wasn’t able or interested to learn a little more about his computer he should at least have asked someone to help him instead of wrecking his install and then making a media circus out of it.
I can understand that a windows XP driver can’t be installed onto a Vista computer in many cases. However isn’t the point of all the bloat in windows is that it bends over backwards for compatibility.
Surely as others have pointed out the OS should have stopped the install and prompted the user to install the correct drivers.
I never had this problem moving users from win9x to Win2k/XP. The win9x drivers would be stopped from installing on a win2k system.
I think this problem lays at both Dell’s and Microsoft’s feet, dell for crappy drivers and Microsoft for not making their os more robust.
That only extends to software. A fifteen year old program can be run, but a fifteen year old driver can’t be used.
The Win9x drivers were totally differently then Win2k or XP drivers. NT drivers could be installed on Win2k, since Win2k was NT 5.0, and those could produce spectacular failures.
That’s really 1991-think. This isn’t DOS anymore where you’re SUPPOSED to fiddle with config.sys, autoexec.bat and other .ini files just to get a printer working. EVERYONE is SUPPOSED to be able to do this, the first promise for this was made with Win95, but of course never delivered.
For me it’s as simple as if it’s bad (the driver), don’t try to install it (Vista). This seems to me like deliberate fscking up of a user’s machine. Of COURSE he couldn’t blame the user, but believe me, if he would have had the possibility, he would’ve.
MS just epically failed here, and there is no excuse for it (ironically there really isn’t, excuse the pun).
Because people like car analogies, here’s one: You cannot expect a car driver to know what kind of car he’s driving. Furthermore, don’t bother him with traffic rules. “I don’t need no stinkin traffic lights, I see a free road by myself!” 🙂
But honestly, in reality your requirement would save much time and stop most of the usual problems from occuring. But sadly, it isn’t the case. Why? Because users don’t use an OS, they even don’t use a particular program (except it is their favourite pirated software) – they simply do things, and some of them even don’t remember what they’ve done a few seconds after messing up a system.
For entertainment, see this example:
Tech Support: “May I ask what operating system you are running today?”
Customer: “A computer.”
See more examples here: http://www.rinkworks.com/stupid/cs_os.shtml
On the other hand, there are users who want to know which problem occured, most exactly and quickly. But error messages provided don’t help them fixing the problem on their own: “Error installing driver. Error code 22. Exiting.” This isn’t very helpful except you have the vendor’s error code listing.
Sorry, but users don’t learn, they are not interested in learning. They want to do things without needing to know anything.
The situation, as you described, is a typical one: People are doing stupid things and start complaining. They are driving a car when they are drunk and / or on drugs, they cause an accident, and they do complain why the car manufacturer didn’t protect them or why the tree didn’t get out of their way.
Modern operating systems, to come back on topic, should deal with such “lazy” users. They should prohibit installing wrong drivers, or, mich better, they should include the drivers and load them automatically if the related hardware is detected (this is how UNIX and Linux usually handle hardware, in most cases), there should not be a need to do something manually. Another option would be to detect the hardware, automatically download the proper driver (if it isn’t there already), install it, and the work is done.
If I may add this: Problems like the one described seem to occur only on nonstandard hardware, meaning that the manufacturer does not try to build a device compliant to existing standards and better insisting on doing everything in a different, incompatible way, forcing different drivers for different OSes. This seems to be very popular in multifunctional printers (“electrical sheet feeders”), egg-laying wool milk sows that do printing, copying, scanning, faxing, peeping et cetera. Such devices are usually sold to home users. Corporate customers seem to prefer single devices for each job, e. g. a capable and standard conforming scanner, a PS or PCL compatible laser printer, a separate faxing machine… you surely get my idea. Of course I know that both solutions have their advantages and disadvantages, but…
(1) The worst solution always prevails. (and implications!)
(2) Users want cheap, they get cheap.
(3) There is no number three. 🙂
You say that, but you’ve obviously never tried to get an HP Laserjet 1018 working in Linux.
I only occasionally use a venerable okipage 6e printer bought for W2K. It used to work swell with Windows when first installed but last time I tried, I gave up on it after a few tries, no test page, no comprendi. Tried the same printer on OSX then Ubuntu, still didn’t work on either. Finally tried it with BeOS and it worked fine, ironically thats the OS that was famous for having no print support when it most needed it. I am sure I could get it to work properly on every OS I use but why bother, I won’t print anything for 6 months.
So stuff is generally too damn complicated under the hood and time is too precious to waste fixing it on one OS if another can do the job. Of course most user won’t have any option but to get their only OS to work.
and that will be a perfect title for the Onion.
I remember installing Win2K on a system that originally had Win98. Having failed to uninstall the Win98 drivers from the system (I mistakenly assumed that Win2K would update any such drivers), I discovered that finding, uninstalling, and upgrading these drivers was a nightmare. After a few months I discovered that a usb port wasn’t working and other hardware had issues. As the machine was a dell, my first thought was hardware failure but a linux cd showed that was wrong. It turns out that I had missed a dial-up modem/sound driver. This particular piece of hardware was not actually used anymore but it’s undead corpse was still there ready to rise up and suck the life out of any driver that came within it’s reach. The only way to put a stake through it’s heart was with a driver update I’m still not certain I found all the corpses in that box.
A shitty driver causing problems? That’s such an extraordinary event! Never seen this before, shame on you Vista!
This story is so lame it isn’t even funny. Drivers are pieces of software, software has flaws and bugs, patches solve some showstopper flaws and create a few new ones, life goes on.
Has a printer driver ever been good? there are a few categories;
mostly works.
sometimes works.
will work but will keep randomly break the queue.
requires zen-like restart skills.
disappears randomly from the dimension.
you really do get what you pay for with printers.
This all is only funny if you take a step back and look at it. I mean if this was linux-related story, usual opinions would be an average of told-ya, yet in windows cases, even with vista, it’s more like so-what-move-on
I ended up having to switch off Ubuntu because — among a number of other bugs in the latest release — I couldn’t for the freaking life of me get my printer to work without mangling the margins and such on double-sided printing mode (which I need).
Fedora had no problems.
Moral of the story: printer drivers suck everywhere, Vista is not at all special or otherwise worse off because it has problems with drivers.
I thought hardware drivers had to be certified to install in Vista (not running it may be totally wrong). Surely when the installer tried writing the driver files, or modifying the Drivers section in the Registry that Vista could have said NO instead of SILENTLY putting them somewhere else.
If Vista had said no and the installer then errored and backed out, the guy would have looked for new drivers then and there. The OS (Vista in this case) IS at fault, it is the silent redirection of DETECTABLE driver files and config that caused the issue.
Probably at the point of installation, Vista did not know that this was a printer driver in particular that was installing and not some other piece of software. For most general end-user software (i.e. the MyPrettyAccountingPackage’s of the world), this redirection will make them work perfectly fine without requiring admin rights on the box. A combination of unlucky circumstances made it do the exact opposite on this customer’s machine and that’s unfortunate.
It’s hard to be sure something will work unless it’s actually tried. Bob Colwell, the chief engineer of the P6 and Pentium IV, had the extremely applicable motto, “If it ain’t tested, then it’s broken.” By the time WinXP went through two years of its life, things were pretty well tested throughout the Windows ecosystem, except for those hardware items that were discontinued. Vista has been generally available for 18 months and I’d be fairly confident about buying a new piece of hardware and having it work on Vista today. Perhaps the main failing here was the lack of great diagnostics about why the install was failing. Maybe such diagnostics were available and the Windows Test guy was able to use them… there are just some areas where novices can’t quite solve the problem without help.
Why don’t MS do the wise thing and actually port cups to Windows… Then there is a standard way to print for all the major operating systems, that doesn’t require kernel changes to print to a device.
Printer drivers work in user mode. There isn’t much kernel code involved.
Also, what benefit is there to Microsoft or its customers to switch when pretty much all printers that are created work fine with Windows and printer manufacturers are pretty used to the current model?
Maybe other OSes would benefit, so perhaps their proponents can do the work to make a shim driver between WinPrint and CUPS since the printing system is extensible like that. Maybe manufacturers would be happy to create cross-platform CUPS drivers in that case.
I love the line at the end of the article.
Said Microsoft’s White:
“We probably could have done a better job here — by ‘we,’ I mean the royal ‘we’ of the software industry…”
Right. The whole software industry. We all dropped the ball on this one.
It’s kind of true. Driver manufacturers had plenty of time to write proper Vista drivers for their hardware, but a lot didn’t. Dell’s installers are always crap, nonstandard, don’t use the Windows Installer service, they are just crap. If a driver manufacturer does not follow the guidelines, is it MS’s fault or the driver manufacturers? I tend to lean towards the driver developer.
It wasn’t for the lack of an existing Vista driver that the problem occured, it was because Vista allowed to install an incompatible driver.
I’m going to say it again because it seems that it doesn’t stick out enough for most of the people:
Vista allowed to install an incompatible, detecable as such, driver.
now with some emphasis:
Vista ALLOWED to install an INCOMPATIBLE, DETECTABLE as such, driver.
How is that ANYTHING but Vista’s fault?
Edited 2008-06-19 13:35 UTC
An installer is just a program. A program can be written to do just about anything. Now, Vista allows older drivers to install because SOME OF THEM are written properly, and will still run, but Vista has no idea which ones are which.
If Dell had written their installer properly, it would not have been an issue.
Let me repeat that IF DELL HAD WRITTEN THIER DRIVER PROPERLY, IT WOULD NOT HAVE BEEN AN ISSUE.
(see everyone can be condescending, it’s great fun).
If they had of written their driver properly, it would have used the Windows Installer service, which would have at least allowed the old driver to be removed properly, but because all these OEMs have to reinvent the wheel at every opportunity, they went their own way, and broke printing.
Vista, in it’s defense, managed to stay mostly functioning through the whole debacle, when older versions of windows may have BSODed.
There is no way to be sure that Vista could detect that the driver is not compatible, as we do not know if Dell’s install program gave Vista the proper info for Vista to make that determination.
As far as Vista knew, it may have been just a program that needed to write to restricted locations, and acted accordingly.
I’m fairly certain that Linux would also allow this behavior, as to install a kernel driver, you need to be root. If you are root, you can do anything, including installing an incompatible driver. It’s easy enough to do, try installing ATI’s drivers when you have an Nvidia card. Does Linux stop you? NO. At least Windows warns you that the driver may not be compatible with your hardware.
You’re right in as much as it wasn’t -just- Microsoft’s fault. It was also clearly Dell’s fault.
But to go from it being Dell+Microsoft’s fault to The Whole Industry’s fault?
No, that’s not ‘kind of true’. It’s blatantly disingenuous.
Edited 2008-06-19 15:14 UTC
It is the fault of Microsoft by virtue of the fact that they allow their name, and the Windows brand to be associated with substandard printers. The Windows Vista logo should not just be handed out simply because drivers exist – the product should have to go through a rigorous quality testing programme, from the product to the drivers. If it isn’t up to scratch – it gets sent back with a list of things that need to be fixed before it can be certified to be Windows Vista compatible.
Microsoft need to raise the bar; they can do it, but they refuse to do it. Heck, they could demand that all printers for Windows Vista must be XPS native out of the box – that is, you can send documents straight to the printer in native XPS format, and it’ll print without the need of drivers. The fact is, they choose not too.
Sure, it won’t stop crappy printers being sold, but it would mean that consumers, when looking for their printer, will avoid those products without the compatible logo on it. Consumers will then in future associate those products with the Windows logo as meeting a certain level of quality. Those that don’t have the logo, the consumer who avoid like the plague.
* XPS is a royalty free XML format which Microsoft is marketing as a replacement for PDF.
Well, that is another argument altogether, and I don’t disagree, the requirements are rather badly enforced, but in this case, I see it more as a problem with the installer, and not the driver per se.
That’s the thing; when it goes in, Microsoft will say, “this installer is crap, use the msi one – or not sticker for you!”; there will be some moaning and groaning by vendors, but when they find their sales drop because it doesn’t have the special sticker, you’ll find that things will change.
The ‘review’ for the sticker would be for the whole experience. From the installer to the software to the instructions etc. Like I said, the only way to improve things is through the branding, or there lack through the with holding of compatibility certification by Microsoft.
Edited 2008-06-19 18:35 UTC
Well, see the problem is, that the driver in question was certified for XP, (if it was certified at all) not Vista. So I don’t know if the installer is not up to snuff in XP, but generally, Dell’s installers is crap, so we can assume.
I ran into the same thing with a Xerox printer/scanner/copier/fax at work. Since the printer is shared through our small business server, when I tried to connect to it Vista asked me if I wanted to install a driver. The driver that was installed was an XP driver and didn’t work. Vista never complained though. The drivers just died with an out of memory error once you tried to print something. Once Xerox (finally) posted updated drivers that were compatible with Vista, those wouldn’t work either because Vista was still using the non-functional XP drivers.
Removing files and registry keys helped but took a while. Big pain in the ass. And we had 4 Vista PC’s that had tried to connect to that same printer.
*sigh*
I hate computers.
It seems to be that one of the hardest things for computers to do these days is reliably produce physical results on paper. I almost miss the days of noisy and slow dot-matrix printers, but they were some of the most reliable output devices ever made. There were no real ‘drivers’, you just plugged it in to the parallel port and printed – if there was paper in the tray and the printer was online, all was well.
But then the parallel port was replaced with USB and everything went straight to hell. Sure PostScript and lasers have given us razor-sharp printouts, but only when drivers are available and feel like working. It’s a trade-off, and if you look at the situation from a functionality point of view there are many times when photographic quality isn’t necessary and a blocky (but working) printer would do just fine. Now if you’re a photo fanatic and must have 2400dpi, six different ink cartridges, and 32″-wide paper, you’re stuck with the tragedy of modern printers. But if all you need is a copy of mapquest directions, do you really need vector graphics?
My ideal printer would have USB-to-parallel port emulation. A USB plug, but with a switch to enable simple LPT1:-like dot-matrix-like operation when drivers seem to be non-existant or non-functional. I’ve seen devices with integrated USB-to-serial chipsets (like FTDI), so a parallel implementation should be trivial. Or perhaps a simple chunk of block storage with drivers saved within the printer itself, after all USB storage is cheapo these days. A few hundred MB would be more than sufficient for Win9x,2k,xp,vista, mac, and even linux drivers to be available INSIDE the device instead of giving users a scratchable and loseable CDrom. If the memory is R/W, update software could even download and update the drivers to the printer’s storage when new versions come out (something CD’s certainly can’t do).
But I don’t build printers, so phoo….
Having drivers on the device wouldn’t help this guy, since his printer would never have had vista drivers on it in the first place.
On the other hand, Cups has the right idea… The print server has drivers, it accepts postscript data over the network, converts it to whatever the printer needs and prints it.
So how about a cups compatible print server built right in? Then you could plug the printer in to the network and have it work out of the box without requiring any drivers.
My printer at home does that, as do the printers at work, i can move between the two seamlessly and have never bothered installing any drivers on my laptop.
This almost works, but not quite.
HOWTO MAKE A POSTSCRIPT NETWORK PRINTER OUT OF AN INEXPENSIVE USB INKJET
The easiest way is to get hold of a cheap HP inkjet printer and an old laptop … even one where the screen is dead will do, but make sure it has an ethernet port and a USB port.
Install Linux, samba, VNC and CUPS from a liveCD on to the laptop (use another monitor temporarily if the laptop screen is broken), plug the HP printer into the USB port, setup the printer in CUPs, and enable the machine as a samba server and a VNC server, and leave the laptop running.
Connect the laptop via ethernet to your LAN, and on any connected Windows machine you should now be able to print postscript to the CUPs printer.
The only caveat … can you believe that Windows does not even have a generic Postscript printer driver?
You will still have to install a third-party generic postscript printer driver on any Windows client machines (Linux is fine, and Macs should be also).
Get a suitable Windows driver from here:
http://www.adobe.com/support/downloads/product.jsp?product=44&platf…
sorry, but windows comes with a lot of postscript drivers. pick any driver which has ps or postscript in it’s name.
don’t bitch about windows postscript support that was the easiest thing to get working. what a lot harder is is to tell windows the printer has only A4 and not letter. every app seems to go back to the default setting and now i simply print everything in letter format, it’s close enough to A4 to be usable.
Sorry, but you haven’t got it working yet.
I did not say that Windows had no postscript drivers … I said what it lacked was a generic postscript driver.
So if you use the generic postscript driver from Adobe, the driver will allow you to to supply it with the “PPD” file which will describe the particular printer features and characteristics for that printer.
You can even set up a second networked printer with different attributes using the same generic printer driver and a different “PPD” file.
http://www.postscript.org/FAQs/language/node22.html
http://www.adobe.com/support/downloads/detail.jsp?ftpID=204
http://en.wikipedia.org/wiki/PostScript_Printer_Description
“CUPS uses PPD drivers for all of its PostScript printers, and has even extended the concept to allow for PostScript printing to non-PostScript printing devices, by directing output through a CUPS filter. Such a file is no longer a standard PPD, but rather a “CUPS-PPD”.
http://osdl.org/en/OpenPrinting/Database/Projects
http://www.openprinting.org/printer_list.cgi?make=Anyone
http://www.openprinting.org/download/PPD/
Edited 2008-06-20 00:22 UTC
Crap I’m running XP but to use all the features for my HP 7310 all-in-one I had to install 700mb of rather flaky hp software. After dealing with having to reinstall it over and over again and the 100mb of overhead for 2 years. “Plus something in the hp drivers conflicted with my nforce4 drivers.”
I just hooked it up to a Linux box and set it up as a print/fax/scan server and ditched the hp software.
Now my system runs better and so does the printer.
PS. Sorry about the grammar.
There’s actually two issues here.
1. The user installs windows XP drivers on windows Vista. Sure its the user’s fault. Yet, isn’t it sad that the install program or windows vista doesn’t detect it is installing drivers for a wrong operating system?
2. After installing the correct drivers, it still wouldn’t work as the old ones were still there. Here I have more of a problem with Windows Vista. I don’t have the slightest clue how printer drivers work. Yet, some warning message would have been good (if this situation could even be detected)
I personally would have never published this story on OSNews, for a number of reasons – but hey, plurality is a very good thing, so I’m glad Adam posted it. Diversity is good.
My problems with stories like this is that they don’t really say anything. It’s highly anecdotal, but is presented as if it represents an experience almost every user has – which is of course complete nonsense. I have personally not encountered a single compatibility problem with Windows Vista (and I’ve been using the GM since November 2006) – but you don’t see me generalising that to the rest of the world.
I could just as well write countless similar stories about Mac OS X and Linux. I have a pretty decent Canon Scanner, and getting it to work on Mac OS X was a total and utter disaster. You see, the world’s most advanced operating system lacks built-in scanner support, so you have to fully rely on the manufacturer’s tools (instead of Windows which has the excellent and easy-to-use scanning wizard) – which are a total bitch to install. I nearly broke my scanner in half. And when I finally got it to work – well, let’s just say that I’d rather eat live piranhas than use the CanoScan Toolbox application.
Linux, by the way, generally uses SANE, which works out of the box. However, for some mysterious reason distributions insist on using Xsane, the world’s most incomprehensible scanning front-end ever devised. Live piranhas? I’d rather wrestle ten black mambas.
The same story for my built-in SD card reader, wireless networking (back in the day), ZipDrives – they all are or were a total pain to set up – countless hours, days, sometimes even weeks spent on getting them to work.
Yet, you don’t see me writing stories about it.
On the other hand, a complete reinstall (Format+install) would have solved his problems earlier, without the need for a Microsoft technician. That isn’t anecdotal, it’s been the experience lots of people have with Windows – even though you’ll always find some telling you repair works good enough, I’ve many times been able to solve problems technicians find insolvable (in these cases they accuse “the hardware”) with a complete wipe out + reinstall.
If you want hassle free printing on any O/S you only need to do the following:
1) Buy a printer with hardware Postscript (the standard), and
2) Buy a printer with ethernet attached.
It really isn’t that much more expensive, and your printer should last through several O/S upgrades. Anything else is a lottery, especially the mess that USB is with a fair amount of stuff done in software.
Examples?
A printer that works very well in my local office with Ubuntu as well as Windows is:
HP Color LaserJet 2605dn
The “d” part of the name indicates it is “duplex” (two-sided printing) and “n” means Ethernet networked.
It wasn’t that expensive. Like most things, if you do your “homework” and pay a little more up front, you will get good gear that you don’t have to think about again – it’ll “just work” and you can get on with making money instead of configuring printers
Professional office equipment seems to leave this trend, sadly. I’m glad I’m still able to use a HP Laserjet 4000 (and a Laserjet 4, more than 10 years old now, as backup printer) and a SCSI scanner (because I run some legacy SCSI hardware anyway). 🙂 Okay, nobody wants to have such monsters at home, but hey, do I invest 120 Euro once or 200 Euro every year just to be able to print? 🙂
And PCL is okay, too, if you don’t have PostScript. But still, PS is the printing standard, and it’s versatile and mature. An integrated ethernet port makes things much easier than connecting a printer locally to a machine (mind the USB max cable length).
My graphical I/O is running for more than 10 years now without problems, it has seen many different OSes (Linux and UNIX and MacOS) and it doesn’t even try to stop working. Must be quality. 🙂
I’ve seen printers rendering a computer unusable (nearly no response) while printing… And there were many complains from my family that their printer (which I told them not to buy) wasn’t support on a new “Windows” version, so they had to buy a new printer while the “old” one was still fully functional.
One of my objections to Windows, in particular, is that has a culture of massive drivers – a 400MB printer driver – ohh really!
Each driver seem to have to have a configuration application that for no obvious reason HAS to be permanently resident.
Not only do these elephants on their respective matchboxes take a tax on your RAM and CPU, they also want to update themselves on a regular basis … and lo and behold … a marketing opportunity implemented as a modal dialog box that you can’t cancel while it waits for their server to confirm that nothing has been updated.
This may be a rant … but sadly it is close to the truth. If you stop using Windows for a while then you really notice this sort of stuff.
I agree.. it’s just awful. Sometimes it’s possible to install ‘just the drivers’ and that would work well. It might be best to avoid the manufacturer installer entirely and instead go through device manager and pick the drivers off the install CD (or install download package). Or use a printer of the right model(not too old and not too new) to have the correct drivers alredy in-box on your Windows CD.
… to package managers. Apt-get & aptitude FTW
To be fair, it’s really not the fault of MS, or even Vista; this is a driver issue. In fact, I think if anybody attacked Linux driver support, there would be an onslaught of “OMG ITS 3RD PARTY SUPPORT. (and they would be completely correct, the Linux devs have to power over 3rd party drivers, specs released etc.)” Thought I’d just point out the double standard. It’s not fair to blame Vista or MS, in the very same its not fair for me to blame Linux because my atheros card has shitty drivers (madwifi is working on it
).
I love linux–don’t get me wrong; but we need to be intellectually honest here.
While it’s a funny story & ya gotta to sympathize with the guy, this is equivalent to getting a driver compiled for use in Ubuntu on a 2.4 kernel & trying to force install it on FC9 or SuSE 11 and expect it to work out of the box with no issue. Different OS(lx kernel vs xp kernel vs vista kernel), different package mgmt (dbpkg vs rpm vs setup.exe/msi vs UAC).
…still funny though.
To summarize:
1. Man attempts to install wrong drivers on Vista
=>(user’s fault)
2. Windows is unable to write the drivers to the correct directories, so it decides to place them somewhere else.
=>(Vista’s and/or driver manufacturers fault)
3. Windows is unable to uninstall the drives because of the alternate placement of driver files.
=>(Vista’s and/or driver manufacturers fault)
So, in a way, this was a simple problem that was exasperated by Vista’s and/or the driver manufacturers attempts to make things automatic and painless for the average dumb user, which ended up screwing things up when the “automation” went awry.
This story is really a non-story. Similar issues can happen on any OS in any situation where the system tries to automate without the users knowledge.
That has little to do with windows and much to do with hardware manufacturers. Windows has in-built a ‘add device’ mechanism – it is well documented, easy to set up, and how devices are SUPPOSED to install.
Was this his problem? No. His problem was Dell made their own goofy install program to load the machine up with their crapplets and ‘value added’ rubbish. As stestagg mentioned the drivers are filled not only with bloat that doesn’t do anything, in many cases it limits the functionality.
You see this with vender after vendor of certain pieces of hardware – the worst of the lot being wireless adapters and printers. They make it almost impossible to install the drivers ‘normally’ without installing their crapware first, try to override all the default windows dialogs, and in general end up being such a headache eventually you say “**** this ****” and try to figure out the chipset so you can grab the reference drivers.
If the drivers he tried to install were simply installed through the normal hardware detection mechanism and not the stupid malfing ‘install’ disk from dell – this guy wouldn’t have had a problem – be it 98, 2K, XP or Vista. At the WORST Vista would have told him it was a incompatable driver, instead of Dells installer ignoring the regular install mechanism and shoving it’s crap in whether it was valid or not.