Submitted by “One of the often cited reasons for not switching away from MS Windows is it’s breadth of commodity hardware support. The argument often goes something like this: ‘Since Windows supports all the hardware I have why would I want to risk it not working under Linux? It’s better to continue to use what I know will work.’ While that may be true in and of itself, I’m finding this argument to involve a limited outlook.”
Linux actually supports way more devices out-of-the box than any other OS in the world….most of the windows drivers came from 3rd party. Microsoft includes some in their install CD, but most of the new devices are supported through CD install drivers – Microsoft doesn’t add new drivers in their SPs AFAIK so for new hardware you always will need your CD, whereas many linux open source drivers are integrated progessively in the main kernel, and you get support out of the box for thousand of devices.
Also, because linux has the driver source of most of their drivers, it can give you support if your company goes out of bussiness or it doesn’t support your devices. As an example, most of windows hardware manufacturers are not porting their old drivers to 64 bits – even a serious company like HP has a (long) list of printers which will NOT be supported in the 64-bit windows platform. Most of the hardware manufacturers (there’re exceptions) seem to be porting to 64 bit only the drivers made for new devices. Linux continues supporting them even in 64 bits, which gives linux the oportunity of supporting hardware in certain platforms that windows can’t support.
The thing is, the very same developers making the Linux driver can make their own Windows driver as well.
The thing is, the very same developers making the Linux driver can make their own Windows driver as well.
————–
I will be reposting all comments that get modded down simply because some troll didn’t like the truth.
XP SP2 add a tab in the device properties dialog which allow you to get the PCI ID of the device that you get in lspci – it doesn’t gives you the hardware manufacturer’s name though, but you can look up it in internet or even in the linux’s pci id list.
it also allows you to see interesting info about the devices, like the filters installed in a given CD unit. Useful stuff.
if not the hard and soft vendors and microsoft wants to have available this blame game just in case they f–k up.
that way they can just point somewhere else and be done with it. kinda like the legal system in a way. defence lawyers try to either poke holes in the police’s methods or point out a likely diffrent suspect, preferably both.
same goes with politics. each case is tossed between diffrent departments on diffrent levels until the onlookers and media gets dizzy and tired.
black boxes, bad on all levels…
Edited 2005-11-10 19:35
This is a good example of an engineer writing a marketing brochure. Instead of addressing the key points that matter to the customer (the fact that the competition has better hardware support), he instead dwells on the technical details of a driver model implementation and why this makes troubleshooting a failed device easier.
Look, the user doesn’t care if the driver model in Linux makes it easier to troubleshoot stuff that doesn’t work. The user cares about if it works, and if it does, how well does it work.
Sound points are made in the article – that is not the issue. However, the article does not present a valid rationale as to why there is a “limited outlook” to using Windows because it supports more hardware. Instead, it kind of cedes the argument in its opening statement, takes a deflective turn and then argues about something totally unrelated to the original point.
sure, your avarage user may not care. but their tech buddy that gets used as free support every time something blows up, will.
when the friend gives up as windows cant tell you whats wrong and on the net there are only more people with the same problem and no solutions then the user will indirectly take note of the problem.
its even more fun if the tech friend boots up a live cd linux and the hardware “just works”.
its not like joe user reads os news. he is more likely to read the xbox review in the local newspaper…
Been there, done that. My friend had an ailing hard drive. He wanted to get his stuff off before it died entirely. I got called in. He couldn’t boot to windows (corruption on the boot sectors), and of course the windows “recovery CD” was completely and utterly laughable in its inability to do anything uesful. I brought in a Knoppix CD, booted from it, mounted his NTFS partition, copied stuff over the network to my computer, and surfed the internet while the stuff finished copying.
heh, i managed to overwrite the FAT of a FAT32 partition ones. windows was no go. linux had no problem what so ever. still, i didnt know then what i knew now. most likely i could have used some linux tool to rebuild that FAT.
Typical marketing response. I’ll tell you something, when I call up/go in/e-mail a company and I get response from either:
1.) Incompetant tech people.
2.) Salesmen
I write them off and find a competitor. I write off companies with overtly false advertising. I think there are a lot of other smart consumers who do the same thing.
Now why? There’s something “innate” in us that tells us not to listen to people who tell us what we do and do not care about. We know what we are interested in learning, and we ask: When someone says we won’t care, we believe he’s hiding something.
Yes, I’d happily see every marketing rep fired tomorrow. They’re a leech on society. They exist to sugar coat problems for consumers (i.e., deceit). They’re unnecessary, I certainly hope you aren’t one
.
The user doesn’t care that “the driver model in Linux does this or that” most of the time. But they DO care what the result of that is, and you can’t say the result and expect them to just believe you. You need to show them that you know why what your saying is true: So you explain it in the least technical terms possible.
You’re right though, the article would certainly fail as an essay. But it raises valid points, and I don’t think it was ever intended as formally correct work.
Look, the user doesn’t care if the driver model in Linux makes it easier to troubleshoot stuff that doesn’t work. The user cares about if it works, and if it does, how well does it work.
Your point is rather short-sighted. Of course the user just wants the thing to work, that’s self-evident. But the simple fact of life is that sometimes things *won’t* work. It happens with cars, with TVs, with just about everything. Sometimes things just go wrong. In such a situation, a technically minded person is going to be brought in to fix the situation (raise your hand if you’ve ever had to fix a friends’ computer!). When the operating system provides good debugging tools, that technical person can get everything working much sooner for the user.
This article is actually quite applicable to one of the computers I have at home. On that computer, I spent a year debugging intermittent networking outages. It turns out that the problem was the result of a “feature”, not a bug. Windows XP is “designed” to prefer SSID-broadcasting networks to non-SSID broadcasting ones. Like a good little user, I had SSID broadcasting turned of on my access point, for security purposes. My neighbor, like most people, had never read the manual and didn’t follow that recommendation. Since his network was fairly far away from ours, whether our computer could see it was somewhat dependent on ambient conditions, so that computer spent a year randomly jumping off our network whenever it saw the opportunity.
The most frustrating thing about this issue was not the inane “feature” put into Windows XP. It wasn’t that the OS was too stupid to realize that the other network was unreliably reachable. It was the fact that the problem was completely opaque to the user, and there was no way for the user to enforce what the computer should do. I couldn’t tell it what network to connect to, just what networks I “preferred”. It, of course, merrily ignored my “preferences”. There was no debugging output to suggest at what might have been the problem. In Linux, even if such an idiotic “feature” had been implemented in the first place, it would at least spew some output to the syslog that I could find via dmesg. In Windows? I was shit-out-of-luck.
This is a waste of time, just because this guy has a problem with a firewire card and can’t figure out what the problem is, Linux is better? I learned this lesson a long time ago with hardware, you want quality support you buy quality hardware.
Hardware that is well supported by the vendor will include drivers and utilities. This is not a fault with the OS, but the vendor. And if you are not sure whether something is supported, check the following:
http://www.microsoft.com/whdc/hcl/default.mspx
Although Microsoft does not maintain the HCL anymore, there is still the Microsoft Logo programs, which a vendor has to jump through various hoops before they can place the logo on their products.
This is no different with any other operating system and hardware, choose carefully and ask around if you want to avoid any issues.
would you consider ati hardware quality?
i recall helping a buddy try to get a ati card to work. windows could not find the required drivers on the cd that came with the card, and at the time we did not have access to a net connection.
end result? one f–ked up system that could only display 256 color vga…
Yes I would and was the card in an ATI box, or OEM? Because I have been using ATI cards since 1998 (because they are excellent with Solaris x86) and have had no problems.
from what i know it was ati. atleast the driver cd carryed the ati logo and all that.
The reason why I ask is this, when I went to upgrade the video driver in my Gateway laptop, the ATI driver setup would fail in previous versions that was not specifically written for the Mobility chipsets.
The Radeon drivers are designed to be compatible with all cards, and the 4 machines I have at home that have Radeon cards installed I have used the same drivers on all of them. ATI cards before the Radeon I can’t say whether this would work or not. In my case it didn’t matter since most of my machines at that time were running Solaris x86.
i was more or less guessing that they had fixed the “problem” at some point but the thing was that even a big brand name like ati can screw up when it comes to drivers…
Have you ever tried installing drivers under Linux?
Didn’t think so.
Go ahead and then come back and read again your comment
on a nice system i dont have to.
on a not so nice its a matter of getting the headers of the kernel im currently using and running ./configure;make;make install. then i modprobe the resulting module.
yes its a bit more “complex” then doubleclicking a exe and then hitting next a few times. but it rearly fails unless the driver is so old that it needs a total rewrite of the interface between kernel and driver.
and most likely, as the source, or atleast part of it, is available, someone else have done just that.
allso, failure to install the error is more likely to produce rich error messages that i can feed into say google to go look for fixes. windows often do a hal impressing (sorry i cant do that, (l)user) and thats that.
btw, cute nick. black metal and pertrol?
Edited 2005-11-11 00:25
Or you could just do (as root):
./configure && make && make install
and if you really know the name of the module, you could just add that at the end
something like:
./configure && make && make install | exec modprobe modulename
You could even script it if you really wanted to…
Dave
How simple. Even a monkey could figure that out right after he finished typing Hamlet.
make
make install
What’s so hard about that? As a prerequisite you need:
1.) A symbolic link in /usr/src/linux to your current running linux. You also need a /lib/modules/linux to your current linux modules.
2.) Of course, you’ll need any “source” package for your version of linux (good distributions make these easily available, or just offer you no choice but to have them).
Now, some drivers aren’t packaged with nice makefiles, but that’s the ditributors fault isn’t it
.
You forgot GCC and the binutils, which are not always installed on home-oriented distributions. Ubuntu comes to mind. Moreover, you must carefully read the README coming with the driver since you have to check the compatibility with your particular version of kernel, gcc and other parts1 of the toolchain. Oh, and after you did all that work, you should hope the folks who compiled the kernel for your distribution have enabled CONFIG_MODVERSIONS. Otherwise, you will have to repeat your work at the next kernel upgrade…
Let’s face it: Linux got better hardware support OOTB but it’s a PITA to install a module unless you have it in a repository package. It’s just easier to install a driver on Windows XP. Personally, it’s a non-issue for me since I can handle it but it would be a lie to say it’s easier.
Yes, I consider ATi’s hardware to be very high quality. It’s their driver department that consistently drops the ball.
Yes, I consider ATi’s hardware to be very high quality. It’s their driver department that consistently drops the ball.
I had a Radeon All-in-Wonder I used in an Athlon XP box with Windows XP Pro. Nice card, but after a few days, crap would start to appear on the screen, then the system would eventually crash. After restarting, it would almost instantly start with the crap on the display.
Reinstalling the driver would sometimes cure the problem… for a few days. I tried new drivers, I tried old drivers, I tried beta drivers… nothing would cure the problem for more than a few days.
After months of pain, I FINALLY found something I could install that cured the problem PERMANENTLY – Fedora Core.
and thats the point, even excenlet hardware may have crap drivers.
but as long as the drivers are open then they will continualy improve. not just as long as the company that makes the card think its cost effective to do so…
yes this most likely means a slowing down of the scycles of development but i must say that for me that would be nice. this rollecoster ride that tech have been for a while is just getting silly.
Its amusing to see someone pointing to the microsoft whdc site. Microsoft claims that drivers that pass its its Hardware Compatibility Tests (WHQL) only will be digitally signed (cataloged) as it confirm that nothing is broken there (of course vendors have to pay a heavy fees to MS for getting this signature). However, after working on Windows Device Drivers I have found that most standard drivers developed by Microsoft and shpped with standard Windows versions (like XP SP2) are broken and they do not confirm to the test standards. So either these drivers are buggy or the test suite itself has bugs
.
Most of the time the driver stacks are not properly documented leaving the HW vendors just to fend for themselves, which may or may not produce any stable drivers.
Definitely, open source has much better chance. The HW vendor knows exactly how the underlying driver architecture works and its much faster and easier to write a reliable driver for it.
My vote definitely goes in for Open Source here, and I feel its high time that hardware vendors stop supporting Microsoft OSes till Microsoft starts providing proper support and improves quality of its own stuff.
Get wireless cards working, either out of the box or hassleless, THEN Linux fanboys can claim better hardware support. >:(
Well, Mepis has good support for wireless, and I believe Kanotix has excellent support. And Ubuntu isn’t a slouch either I believe. Libranet does a pretty good job as well from what I read/hear.
Let’s take a step back here Ronald. How many Wireless cards does Windows support without the aid of 3rd party drivers (either downloaded from the web, or from a CD/floppy)?
The real issue is that wireless card manufacturers (including Intel) do not provide full specifications for the hardware. And, under the DMCA, it’s illegal to reverse engineer said drivers without the sole permission of the owner/manufacturer of the product! So, how do the Linux guys write a driver, on something that they don’t have specs on, and have no support from the actual hardware vendor itself, where reverse engineering for compatibility purposes is outlawed by the DMCA? Especially when the DMCA is meant to compliment the US Copyright Act I might add (and it supports reverse engineering…).
I’m a solid firm believer in that hardware manufacturers need to pull the finger out of their asses and support their products properly. They’ll write a driver for Mac OS X, but not Linux! Their argument – there’s not enough users to justify it! That’s bullshit. There are more users of Linux than Mac OS X worldwide. It’s hypocritical. I suggest you draw you energy and fire it off at the hardware manufacturers, who are only doing a half assed, cocked up job of supporting their customers.
Onto Windows, I’ve seen many issues with Windows “plug and play”, so much that I usually call it “plug ‘n’ pray”. The other day, I was directing a customer to install a driver, Windows 98, we navigated to the location of the driver, selected the correct .inf file, and it refused to work. Opened the inf file ok, but no drivers present. So, we tried it on automatic install, it was pointing to exactly the same location, and lo and behold it worked. You tell me, what’s up? I’ve seen several instances of Windows saying “unknown hardware” in device manager. The problem? If you have more than one item of hardware that’s not detected, and therefore more than one “unknown hardware”, it’s rather hard to determine which one is which (and obviously install the correct drivers, to the correct piece of hardware).
Windows is far from perfect. Linux supports more hardware out of the box, has for some while now. Windows hardware support is only so good, because it has 3rd party hardware manufacturers writing drivers to support it’s operating system. Nothing more, and nothing less. And, I’m pretty sure that Microsoft has nice dirty deals precluding manufacturers from supporting other operating systems. Of course, no one would dare admit it. Illegal collusion in the software industry, nah!
As dylanmrjones pointed out in his counter post, if you’d stopped being so insulting, you might have actually had a few people off help/suggestions.
As to the hardware qualification program that Microsoft offers, it’s VERY expensive, and in reality is just for a Microsoft tick. It’s simply a way of Microsoft making more money, and means very little to the customer. But, realise that by a hardware manufacturer paying X amount of money to Microsoft for certified drivers logo, it raises the cost of the actual hardware. Think about it.
Dave
Windows 98, how long ago was this? There are people who still use Windows 98, but I would hardly use that as a comparison point.
And you have a point with wireless, which is part of the reason why my laptop doesn’t have Solaris x86 installed on it, no support for the two wireless cards (Broadcom and Proxim) I have in it.
With Windows Certified hardware, yes it costs more but if it works “out of the box” isn’t it worth more money? This is what people pay for when they buy a Mac, everything works with no problems.
I have read posts like this for years about people bitching about driver support, the answer is simple. Vote with your wallet! I only buy hardware that gives me quality drivers and diagnostic tools so I spend more time using my computers and less time fighting with them. If enough people avoided hardware vendors who feel it is not necessary to provide quality drivers and tools, they would not be in business for long. And if that means I have to spend more money for it fine. I guess that is why I haven’t had a significant problem with a computer in 11 years. And this includes using Windows, Linux, OS/2, NetWare, and Solaris x86. In fact the only OS I had problems with hardware wise was SCO UNIX (before the Canopy Group).
Quote: “Windows 98, how long ago was this? There are people who still use Windows 98, but I would hardly use that as a comparison point.”
Au contrare! A lot of people still use Windows 98se, some even plain Windows 98! You’d be very surprised. Penetration of Windows 2000 wasn’t really that great, Windows XP was better, but there’s still a lot of people out there running Windows 98 in all honesty. They don’t have (or don’t want to pay) the money to upgrade both the operating system AND hardware.
Quote: “And you have a point with wireless”
Yup. It’s pretty sad in reality. Hardware manufacturers are really shooting themselves in the foot as far as I’m concerned, because I simply won’t buy or recommend their products, and it’s something that I try and encourage others to do.
Quote: “With Windows Certified hardware, yes it costs more but if it works “out of the box” isn’t it worth more money?”
It’s a waste of money in my honest opinion. Why should Microsoft charge you money for it’s inability to make a stable operating system that can interact with 3rd party software, and hardware drivers?
Quote: “I have read posts like this for years about people bitching about driver support, the answer is simple. Vote with your wallet!”
Totally agreed!
I haven’t played with SCO Unix or OS/2 or Netware, so can’t say too much about their hardware support, but I would imagine it’s worse than Linux, since there’s a much smaller footprint of systems out there!
Dave
Some wireless cards work out of the box, some don’t. As usual, look up HCLs before buying your card.
That said, there are solutions if you card isn’t supported, such as ndiswrapper (free) or Linuxant’s driverloader (not free, but works better than ndiswrapper for some cards, such as Broadcom’s).
Oh, and the reason you were modded down was because of your blatant attempt at flamebaiting, i.e. using the words “Linux fanboys”, and NOT for criticizing Linux. Once you grow up a little you may find that people will be more inclined to listen to you.
My debian system sees my ORiNOCO Silver card. I didn’t have to do anything extra.. Course EVERY OS ever created has drivers for the ORiNOCO chipset….
How typical: modded down for pointing out Linux’ obvious flaws in regard to supporting wireless cards. It appears some people can’t handle the truth.
You shouldn’t be complaining about “fanboys” then.
Your post was insulting and I believe that was the reason for the modding down.
However, if you had removed the “fanboy” part and added some links to cases on the issue, it would probably have been modded up by these “fanboys”.
You can find many flaws in Windows as well, as in Mac OS X and in the Commodore 64.
What’s interesting are the reasons for these flaws
The official ATi Radeon drivers for linux only work with the 8500 and up. I had a Radeon 7200 and had to rely on the OSS drivers which had poor OpenGL support. However, even with those, the Radeon was one of the best cards I’ve used in linux.
This is going to be moot with the new Windows Driver system for Windows Vista.
How is linux going to be able to combat that? I am sure some linux fanboi will say some crap.
Sorry, it won’t be moot.
Those of us who can’t afford to upgrade to a 3ghz computer with 2gb of ram will be more than happy to use our current linux boxes.
Why should that much horsepower be needed for just the OS itself? What about other apps?
heh. I doubt vista will require all that to actually run. I’m sure itll require a hefty system if you want all the glitz and glam… but who wants that? I plan on using is just for some of the new features. not the UI.
>Windows is far from perfect. Linux supports more
>hardware out of the box, has for some while now.
Raise your hand if you think Morgoth is a linux fanboi?
Raises hand pretty high.
Today, a lot of people will make up their own crap just to prove a point. Morgoth is a perfect example of this.
I suspect you need to actually learn a bit. Out of the box, Linux does indeed support more hardware. I didn’t say that the level of support was better than the 3rd party hardware manufacture device drivers that you can get on Windows with most hardware.
Anyways, you stick with Windows, I’ll stick with GNU/Linux.
Dave
I can say from first-hand experience that Linux does indeed offer more complete support for certain devices. My motherboard (ASUS A7N8X-E) has an integrated NForce 5.1 channel audio chipset. In Windows, there is currently no way to utilize the yellow Center/LFE connector on the I/O panel. Instead, I am forced to give up my Mic connector in order to use the Center/LFE plug on my speakers. In Linux, this is not the case. I can choose to either plug it in to the Center/LFE connector or the Mic connector, and I simply select which one to use in the software mixer. It’s nice to be able to record audio and watch a DVD back-to-back without having to pull out my tower and swap cables.
I will even go as far as to say that the only area where Linux lags behind Windows in driver support is printers and scanners, at least in my experience. My Lexmark X6150 all-in-one is a paperweight as far as Ubuntu 5.10 is concerned; there simply isn’t a decent printer driver nor is there a SANE scanning backend at all for that device.
Don’t think so. while linux has more BASIC support OOTB, windows and it’s 3rd party drivers have more advanced drivers.
I can’t vote with my wallet, cos cheapo stuff is usually all i can afford. All work with windows.
Thanks frys.
Did you actually read what I wrote? I wasn’t talking about basic out-of-the-box support. I cited a specific example of Linux drivers for one device actually exceeding the capabilities of and more correctly utilizing the hardware than the device manufacturer’s Windows driver. In other words, the device in question worked better in Linux than in Windows. While that may not always be the case, it is very nice when it does happen.
Concerning printer, turboprint might help you because the driver support most of them. http://www.turboprint.de/english.html
Scanner can be tricky as you will need to manually enable it for users. It has problably something to do with SANE so you contact the developer about the issue.
Thanks for the tip on the printer!
As for the scanner, Lexmark does indeed offer a Linux driver development kit. However, I am not a programmer, and from what I’ve read the DDK is far from complete. I’ll just continue to boot back into Windows for that device until I can find a better arrangement.
That statement is so far from the truth that you couldn’t be harder from the truth …
Take for instance my Audigy 2 ZS. Can I use its full feature set in Linux? No. What about my X800 XT PE? Nope. What about my nForce 3 chipset? Nope. Logitech Cordless MX Duo? Nope.
Sure, Linux “supports” those things so much that I can use them, but in each and every case support is either limited or unstable. I can’t use the full feature sets of most of my hardware in Linux, whereas I can in Windows. Sure, you say “blame the manufacturers”, and I agree to some extent … but let’s not go around making blanket statements like “Windows supports more hardware than Linux, just not as well” — because that is pure and unadulterated bullshit.
yeah and my knoopix CD froze my laptop and I couldn’t even get it to boot with any copy I tried and it’s a laptop that is only four years old. What is your point?
// In Windows? I was shit-out-of-luck.//
How about:
1. Right-clicking on “wireless connection” icon in taskbar,.select “Properties”
2. Click “view available wireless connections”
3. Click “Change order of preferred networks”
4. Find the “unreliable” network in the list. Click “properties” … click “connection”
5. *UNCHECK* “connect when this network is in range”
Boy, that was hard. Stoopid Windows.
You don’t think I tried that? Windows ignored my setting entirely. It will always prefer the SSID-broadcasting network to the non-SSID broadcasting one.