Skip to comments.A Call To "Kill All Proprietary Drivers For Good"
Posted on 04/05/2012 5:44:52 AM PDT by Halfmanhalfamazing
UPDATE: I have been contacted by Qualcomm PR regarding this expected presentation next week. The content of the Qualcomm Atheros developers was not approved by Qualcomm's legal department and the views to be expressed will be of their own personal beliefs.
Next week at the 6th annual Linux Foundation Collaboration Summit in San Francisco, two Qualcomm Atheros engineers will be speaking about their Linux device driver development experiences and will go as far as calling for all proprietary drivers to be killed for good. They talk not just about killing proprietary drivers for Linux, but for all operating systems. Can the plans they lay out to kill all proprietary drivers work or is this just a big pipe-dream?
(Excerpt) Read more at phoronix.com ...
There's always someone willing to write a driver for devices. There's a whole project built around it. http://www.linuxdriverproject.org
This is one way of combating the concept of built-in-obsolence. If my device still works, why am I forced to buy another one?
ping for later
Won’t happen. There are way too many devices where the hardware manufacturer is the sole supplier of the driver. Even serious attempts to standardize (e.g. Postscript) have no good way to anticipate new features to make full use of the hardware in question.
A lot of the stuff is pretty standard already... SATA, monitors, USB 1.1 through 3, Firewire.
But it ain’t gonna happen with that 256bit fancy pants video card ... or with that specialty two sided sheetfed scanner. If the manufacturers aren’t willing to make a version friendly with a Mac Pro (in the case with high end video cards) they aren’t gonna bother with Linux, which is mostly in the server and hobbbyist and low-end consumer arena.
When I read the headline “Kill all proprietary drivers” I thought it was yet another anti-global-warming executive order coming down from the EPA.
This isn’t quite fair - but there is a large grain of sanity in your point of view. I’ve actually delved into writing a Linux device driver. It is NOT for the faint of heart. It is a non-trivial task.
It would be NICE if companies didn’t use proprietary drivers. There are MANY applications where there isn’t a good reason to make the driver proprietary. If the specs of your hardware are already public - which is many times the case, then making the device driver proprietary doesn’t make sense. If you say hardware is never exposed this way - not so. Even the original PC had the entire register set (and roms) published.
However - there can be both competitive and legal reasons why you keep something proprietary. The easiest LEGAL reason to explain is one for a WiFi interface. These have radios in them. Radios that are programmable are a scary thing to groups like FCC. There are legal restrictions on how much power the radio can put out, and this varies with jurisdiction. So a proprietary driver is mandatory from the companies point of view (and the government!)
Certainly - if you have a newfangled piece of hardware that you want to keep secret, yet you want to sell it, then publishing a proprietary driver only makes sense.
Now it comes down to how many different drivers are you going to build, i.e. what OS’s are you going to support? This is the REAL problem for the company.
Ignore the Open Source Zealots - they don’t live in the real world on this topic (I’m an nomrally one of them..one of them...)
Yesterday I installed Windows 7 on this computer. The pre-check I did told me that my printer was compatible. However, after Win 7 installation, my printer didn’t work. Why not? Because Hewlitt Packard decided not to support numerous of it printers with drivers for Windows 7. So, my HP Laserjet 1012 didn’t work.
I went on-line and found a work-around that isn’t perfect, but will allow me to print. That’s the last HP will see of my money.
To heck with killing prop drivers, I would be happy if they killed all those stuipid unneeded “helper” apps that install in my system tray that use .NET framework or some other Library type DLLs to install that then just sit in my app tray and consume massive amounts of memory and processing time.
Like I really need some HP shovel ware sitting on my desktop to tell me that the drivers has been updated or that the toner on the printer I rarely use is still at 89% capacity....
Also banning other crap that “driver suites” tend to install like ANY sort of damned browser toolbar or anythign that runs ar startup that isn’t directly driver related. I don’t need another volume mixer for my sound card when the windows one works just fine for 99.9999% of the users out there, at least give me the option when installing the driver to leave the blankety blank shovel ware off.
[ Yesterday I installed Windows 7 on this computer. The pre-check I did told me that my printer was compatible. However, after Win 7 installation, my printer didnt work. Why not? Because Hewlitt Packard decided not to support numerous of it printers with drivers for Windows 7. So, my HP Laserjet 1012 didnt work.
I went on-line and found a work-around that isnt perfect, but will allow me to print. Thats the last HP will see of my money. ]
Did you try installing the windows driver or using the “Get more Drivers” button, I have installed the driver in windows 7 for an HP Laser Jet 4 thia way and that printer is “Ancient”. I was using the local network and a Jet Direct card to connect to it, by the way.
You wouldn’t believe the systems I have sped up by disabling the startup of “Driver Helper” apps that get loaded on startup. Most of the crap that gets installed During driver isntall besides the driver just sits there and consumes system resources.
All those “Driver Helper” apps do nothing to add any needed functionality and a lot of them are sloppily coded and do nothing but slow down systems to the point that clueless consumers end up buying new computers when their old ones “start running slow”
msconfig is your friend!
Removing all the extra driverware crap can really help a system run a lot faster.
Yes, I did.
The forum that I found recommended a driver for a different HP, told me to change to a different port, and said that when the printer started printing garbage it needed to be turned off and back on to delete buffer contents. Worked like a charm.
I print very little. That’s why I’m cranky about buying anything new when this one has received such light use.
Uhm - Microsoft doesn’t care whether the drivers are proprietary or not. This is really about Linux being a second class citizen AND being zealous about proprietary code being BAD!(TM - GNU.org)
I won’t say “never happen” but put me squarely in the “doubtful” camp.
I have a very good friend who is one of the main driver developers for a very large graphics chip vendor (who could that be?)
He has told me on not one, not two, but on dozens of occasions (and always not in response to a question but on his own) that his company would NEVER do this as there is too much secret sauce rolled up into their proprietary drivers. Basically if you understood their proprietary drivers you would understand way more about how their devices work than the company would want you to understand.
Tell me about it! I had the same problem with an Epson hanging on the network at a customer site. I downloaded the 75MB "driver" and installed it. Course the crapware comes along for the ride without even a box to indicate that you're sane and want it about like you want the plague. So I uninstall the crapware and it deletes the driver!! I didn't realize anything actually deleted drivers. The only workaround I found was to install it on another machine and share it, then configure a printer from Windows on the machine where you really want to have it (selecting the option where Windows goes and gets the driver from the share machine and installs it locally), then delete it from the other machine! What a load of crap all in an effort to try to FORCE their customers to run their spy/mal/crap/shovelware. I have noticed in the last couple years, though, more and more printer mfgrs offering a minimal "driver only" install for download. I still trust the other method more, though, to get only the driver and avoid any of the other stuff. (They may still include what they consider the absolute minimum level of crap they're willing to let you get away with otherwise.)
[ Tell me about it! I had the same problem with an Epson hanging on the network at a customer site. I downloaded the 75MB “driver” and installed it. Course the crapware comes along for the ride without even a box to indicate that you’re sane and want it about like you want the plague. So I uninstall the crapware and it deletes the driver!! I didn’t realize anything actually deleted drivers. The only workaround I found was to install it on another machine and share it, then configure a printer from Windows on the machine where you really want to have it (selecting the option where Windows goes and gets the driver from the share machine and installs it locally), then delete it from the other machine! What a load of crap all in an effort to try to FORCE their customers to run their spy/mal/crap/shovelware. I have noticed in the last couple years, though, more and more printer mfgrs offering a minimal “driver only” install for download. I still trust the other method more, though, to get only the driver and avoid any of the other stuff. (They may still include what they consider the absolute minimum level of crap they’re willing to let you get away with otherwise.) ]
If I can get the installer to extract the files then point windows to where the *.inf files are stored and use the windows driver installer to look at those files it can cut down on a LOT of the crappy shovelware.
It seems that there’s been a bit of misunderstanding of what’s been posted.
—————Wont happen. There are way too many devices where the hardware manufacturer is the sole supplier of the driver.-—————
The point of open source drivers is full circumvention of the hardware manufacturer.
-——————But it aint gonna happen with that 256bit fancy pants video card————————
Both Intel, and more important AMD(Think ATI Radeon cards such as the $500 plus top of the line 6990) cards do very much indeed offer open source drivers for their chips. Just recently, they released the open source driver for the very newest line, the 7xxx.
—————If the manufacturers arent willing to make a version friendly with a Mac Pro (in the case with high end video cards) they arent gonna bother with Linux—————
Again, you’re missing the point. The manufacturer doesn’t have to lift a finger in many cases. They can. Many do. But all they really have to do is release specifications, and somebody else will do the rest of the work.
AMD’s efforts are a great example of this. Who is building AMD’s open source driver for it’s cards? Well AMD is, obviously, for one. But who else? Redhat. Novell. And any other company or random developer who wishes to contribute.
Both Intel as well as AMD(yes including for their high end cards) offer fully open source graphics drivers.
Your friend works for nVidia.
But the issue is not so much opening up a current driver as it is even allowing an open source one to be developed.
Like AMD. They’ll never open Catalyst. That hasn’t stopped them from creating and helping the creation of the open source Gallium3d drivers.
AMD has in a sense outsourced their driver development. In the long run, it could save them millions. They have several companies and individuals doing the hard work for them and to my knowledge, not getting paid for it.
They could theoretically, even realistically some point way down the road only support their newest line of cards for a very short period of time, then let the community support itself after that because the driver is open.
I think it’s the old planned obsolescence thing. That’s what proprietary software vendors have the ability to do.
And it’s why I avoid them whenever it’s practical.
-—————My belief is that the hardware vendors aren’t willing to do their part (release full specs, much less consistently conform to emerging standards) and that the well-meaning open source community won’t be able to generate good drivers in a timely manner.——————
Makes sense. I do think this call will largely fall on deaf ears for now. But the call will be made again. And again. Coming from a set of Qualcomm people is a lot different than seeing it come from some whack nut like Stallman.
With many companies already doing it(you mentioned printers, you should buy HP, they support open source too http://hplipopensource.com/hplip-web/index.html ) at some point the call gets loud enough while at the same time those still coveting their proprietary drivers becomes a smaller and smaller crowd, and eventually the egg breaks.
As to your high end printer and scanner, if you still have your XP install disks, install it in a virtual box.
Yes, virtual box is open source too. :-) When necessary, bring your files into your XP environment and hit print. Voila.
Now that I’ve grown accustomed to virtualization on linux, I don’t know how I ever lived without it.
I’m pretty sure that there exists open source drivers for Nvidia - but it’s an apple to the proprietary orange - again if I have my facts straight. If I’m right the open source one will allow for some functionality of the hardware but to get the cutting edge performance needed for cutting edge applications (read games) - open source just will not cut it.
I don’t see this model changing going forward. If you’re in graphics chip development the whole raison d’etre for your company is to stay on the bleeding edge. And this involves proprietary hardware design. To hint at how the hardware is built by opening up the driver is - well it’s just completely contrary to the whole notion of staying one step ahead of any competition.
People tend to think that there’s a bright line between hardware (chips) and software (driver). There’s not. Whatever you can do in one you can do in the other depending on a number of factors. Just as a graphics chip maker won’t put their (detailed) hardware architecture/implementation in the public domain they won’t put their driver source in the public domain either, for one thing because it provides insight into what the hardware itself is doing and for another thing there may be tricks they are actually doing in software that are equally proprietary.
They have to. Such is the nature of video cards. Different markets different needs.