Linux kernel maintainers removed hooks from a semi-binary driver (half open, half closed) for Philips webcams. This has angered the developer of the driver who have worked on the driver for 5 years, resulting on removing the driver completely from the kernel and his site. The developer has an NDA with Philips and he can’t fully open source the driver, even if he wanted to. The losers from the whole story, are the users.My Take: Here’s an example of some linux kernel maintainers that put who-cares politicalities over practicality & user’s interests. Being Free is a good thing, but if I can’t be 100% Free, I’d take that 50% any day. Common users (the kind of mammal species that the Linux platform is trying to captivate away from Windows/Mac) want cheap and easy solutions, and some Linux maintainers fall short in materializing this need because of this or the other (irrelevant to most users) reason.
Strict Open Source Kernel Policies Give an End to a Driver
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
> By Gnu/Linux Debain user (IP: 62.90.243.—) – Posted on >2004-08-27 05:44:24
>The losers from the whole story, are Philips. No linux users >will buy thier webcams. And I prefer useless webcam then >non-free driver!
And do you think they really care? After all, I’m sure they’ll be fine with all the windows/mac users who have webcams with a phillips chipset in them.
The Linux Incompatibility List
http://leenooks.com/1
Well there are lots of people who care that do have a vested interest in Linux drivers. Also do not forget that by annoying just one person can stop an entire organisation from purchasing your hardware, people such as the IT people in an organisation.
When you are selling a product you should never tread on your buyers toes. It will not make you money but instead will casue you to lose it. A simple buisness principle that many IT buisnesses today neglect. Only those with market monopolies can afford to make such mistakes. “Only the paranoid survive”, Andy Grove, Intel.
I’m not implying however that Phillips should do something for Linux about these drivers, just that if they did they would gain more market share, plus they would gain what might be called “geek support” which could be important further down the road. How can you move forward on the path when you are continually throwing stones in front of yourself.
But on the consumer desktop, linux has a *tiny* share of the market. So why would a company such as phillips spend $$$ on developing drivers, which they’ll have to release new versions everytime there is a kernel update just to satisfy some linux geeks… come on..
I think it’s childish of a kernel developer to just like that go and remove a driver that’s been in there for years and which has a audience that uses it. It’s entirely different if the same developer doesn’t want to accept *new* drivers that are more or less binary, but the old ones should IMHO stay in. Why not instead try to help the webcam developer reverse engineer the webcam so that a fully Open Source solution can be created?
Because those geeks just might be the purchasing officer for some megacorp. How can I make this point any clearer.
Imagine that you wrote a software application for Windows. It has a few major bugs but nothing that seems that bad. Your organisation is about to be purchased by Microsoft for some 48 Million US dollars, in order to absorb this application. Then Bill Gates trys out your app on his PC and it wipes all his files, due to one of the undiscovered bugs.
Question where will your 48 Million go?
One person can make a big difference. Got it?
It always surprises how many people still don’t understand that Linux is first of all a FREE operating system. Free as in speech, of course. This is the main reason for linux, gnome, kde, etc, to exist.
So the politcs ARE IMPORTANT, very important. The linux community is creating the best operating system. Sure there is a lot to to, but this must be done without forgetting the mission.
In my opinion the winners of such a decision are the USERS, the users who remeber why Linux was born and how it has become such a big movement.
I’m sorry if the point has already been brought up, but its late and I don’t have time to read all 107 current comments.
There’s no denying the appeal of a stable interface for binary drivers. I’m sure there’re several companies that would release drivers for Linux if such an option were viable.
However, I don’t see this coming from the kernel guys and gals. Why don’t these companies do it themselves. Noone’s stopping them and they don’t have to wait on the Linux kernel team to do it for them.
They’ll have to manage politics and legalities, but if they want a stable interface for binary drivers, they’ll probably have to do it themselves. nVidia does something like this already, and I suppose there is nothing stopping some industrious companies from forming a common framework of similar style. I’m not inclined to believe that nVidia would share their work (IIRC, their corporate culture disallows such sharing).
Really, and truly, if businesses see value in having a stable kernel interface for binary drivers, they’ll do it themselves. That’s the power and responsibility of FOSS: the finger points both ways.
Leaving apart the fact that Linux is about Openess and that letting the close guys run their products in a close way into the kernel would just make everybody stay closed, the fact is that Linux is so consistent in front of Microsoft because it is Open. Most of the deep blue screens of windows are result of two drivers not knowing how to talk to each other or with the kernel or the kernel not knowing how to exactly talk with the drivers.
If Philips is not going to participate into the party, it is ok. I’m not going to buy an nVidia card anymore, for example, while they don’t release their drivers into the source code because it is very anoying everytime I want to upgrade the kernel It results in no X Windows until I reinstall the NVidia drivers.
But, came on, what is soo important about the Philips binnary drivers? A new codec? There is a thousand of them. What makes they think theirs is so incredible and extraorinary? I think nothing. I think they are simply used to run their bussiness closed.
If you just want all your devices to work and you don’t mind closed drivers use Windows, it’s as simple as that. Linux is a free kernel, I want it to stay as it is. I am careful buying only hardware I know it runs fine in Linux, so I get good practical results and can reward the manufacturers who make it easy for free software developers.
The people who bought a philips webcam (because it was supported) are sorry, the people who didn’t are zealots.
This is kindof a hard decision for all – but as some of you said, Linux got only so far, because they were strict, because they don’t want this kind of thing. The problem was _not_ that there was a binary module, as it seems, but that there was a hook, that exported a function into userspace. This is _not_ about open/closed, but about a design issue.
I kindof understand the disappointment of the developer, but why doesn’t he continue? There is no problem in releasing a kernel patch with the driver that makes it work for those who want it to. The only point is, that it’s not wanted in the kernel-tree. He could provide binary modules against all 2.6.x versions, which is not a lot of hard work btw, and people would still be happy…
Don’t get me wrong, but I think something terrible happened (Discontinuation of the development effort). I think the developer could have managed the situation in a much better way imho….
And lets not forget…. there are still numerous kernel versions that still _do_ work. So it’s not that you cannot use them anymore, but that they wont work with future versions of the kernel…
So how do you explain Solaris or AIX or Mac OSX’s stability?
The Linux kernel has no driver API so how can you say it’s
consistent?. It’s stable but sure as hell not consistent.
As for you not buying an nVidia card because their drivers are closed, it’s your loss – I’m right now playing UT2004 with proper OpenGL.
Best regards
Dev Mazumdar
I see a lot of good things coming, as a result of this situation. I have a certain detachment from the whole mess, because I now use Linux only about 10% of my computer time, and I do not own or use any webcams.
How about Mac OSX? Why does Windows == Closed Source?
Here’s a direct quote from Linus Torvalds:
From: Linus Torvalds <[email protected]>
To: Daniel Phillips <[email protected]>
Subject: Re: [PATCH] Remove Bitkeeper documentation from Linux tree
Date: Sat, 20 Apr 2002 09:56:21 -0700 (PDT)
Quite frankly, I don’t _want_ people using Linux for ideological reasons. I think ideology sucks. This world would be a much better place if people had less ideology, and a whole lot more “I do this because it’s FUN and because others might find it useful, not because I got religion”.
So there you have it!. Your guru Linus Torvalds tells you that you use Linux because its fun and not for some ideological oppostion to Windows or closed source.
Best regards
Dev Mazumdar
> The losers from the whole story, are the users.
No. The losers from this is Philips, and they had it coming.
If the code developed really deserves to be in a GPL product it will still be available legally after the original developer has stopped. If it’s not, then it only proves removing it was the right decision.
Move to Windows. You can install binary drivers or compile your own, without bothering with open source politics or worrying about the kernel suddenly changing it’s API to the hardware. With it’s superior device driver scheme, Windows can even support hybrid compiled/binary drivers!
(And, for that moderator who thinks he’s about to mod me down, this is a transposition of exactly the same argument that survives moderation every time an article is posted containing the words “Windows” and “Security”)
Quote: “So how do you explain Solaris or AIX or Mac OSX’s stability?
The Linux kernel has no driver API so how can you say it’s
consistent?. It’s stable but sure as hell not consistent.”
Solaris supports bugger all drivers, so does AIX. Try getting half the hardware that runs on Linux to run on either of them and guess what – won’t work.
Mac OS X – has very limited hardware support in reality – buy something that’s not on the Apple recommended list and guess what? Yup, won’t work. Apple hardware works well because Apple are a closed source hardware manufacturer that doesn’t accept every tom, dick and harry hardware manufacturers hardware on their platform. They pick and choose and therefore limit any potential problems. Results in a very small hardware base.
Quote: ” With it’s superior device driver scheme, Windows can even support hybrid compiled/binary drivers! ”
What a load of rubbish. It’s certainly not superior, half of the problems with windows are results of crappy API conflicts between drivers and applications. But since it’s pretty closed source, and none of the hardware manufacturers like to talk to each other about interaction between their hardware cos they’re so afraid of losing IP, it just causes issues. The customers in the long run pay for it.
If i’m buying hardware from a manufacturer, then they must list what o/s are supported. If it’s windows only I don’t buy it. I don’t recommend it. Simple. Period. Their loss of income and if enough Linux users follow that strategy things will slowly change. I’m sick and tired of poor support from hardware vendors due to greed and laziness. Most vendor device drivers are poorly written anyways. They don’t care about the software/drivers, they care about hardware sales. Nothing more, nothing less.
The best thing would be to limit copyrights and patents to a maximum of FIVE years. At the rate that the IT/Computer industry moves, 5 years is a long time, and most 5 year old hardware is obsolete (so why keep the drivers and system calls closed?). See what I mean?
Dave
There’s this little law call the DMCA. I don’t know if you have heard of it or not but it can be used to prevent backwards engineering attemps by programmers.
> With it’s superior device driver scheme, Windows can
> even support hybrid compiled/binary drivers!
Are you saying that binary drivers arent compiled?
I guess they’re handcrafted from scratch in a hexeditor…
Man, there goes my webcam support under Linux. Most of Logitech webcams worked using that driver 🙁
A sad day for me.
Is that you are 100% free to add that hook back in to the kernel yourself. It’s not *gone*, it’s just not in vanilla!
Tell your distro that you want to use that driver, and the might include the hooks.
What’s the problem?
The Linux kernel is open source. Just rip the code, write a patch and release a new kernel. Call it Flea-Nux or something. Then let the people choose!
Oh, the joy of open source! You can do practically anything you want as long as you open source your changes and don’t use trademarks (eg Linux).
I find all this quite hilarious, I must say. So much fuss about a few hooks…
I thought open source was all about freedom, about choice? This way, Greg Kroah limits the choice of the Linux user because of his personal beliefs. A lot of companies aren’t gonna like this. Also, I find it rather childish. When is that guy going to realize that what he is doing to the kernel doesn’t affect only him? There are millions of users out there, and he has respnsabilities towards them. Yet he doesn’t seem to care.
A shame.
why not look into possible solutions?
After all, Linux webcam users are being hurt by this.
Solution 1: Phillips opens up the binary code; this may or may not be possible, because even Phillips may not be the author of the binary module. So we just don’t know if this is a solution or not.
Solution 2: the USB kernel maintainer allows this exceptional hook to stay in the kernel. This may or may not be in the interest of the entire Linux community. Also this is a can of worms, because if we make an exception for Phillips, why not for any other manufacturer?
Solution 3: somebody reverse-engineers the binary module written by Phillips, and puts the code in the kernel under a standard GPL license.
Somehow I feel solution 3 is going to happen sooner or later.
Dev Mazumdar: “Redhat was able to got IP”
By coding software (under e.g. the GPL). GPLed code IS IP. It is just a liberal version of it, but it ain’t public domain. They still have all the control over what they’ve written. FYI: BSD licensed code IS IP as well.
Dev Mazumdar: “Why is it ok to have an application that can be closed source but not the kernel?. Where does your set of principles get compromised?.”
Linus himself ain’t some license purist (e.g. see above quoted psot from yourself, and he uses BitKeeper, and …) but in the case of the Linux kernel, the license can be a problem. It ain’t a matter of license principles on which your post seems to be based upon, it is (for some) also a matter of legal, stability and comfort which you failed to address in this anti-Linux rant of yours.
—-
@ Leo:
By the way, binary drivers lower innovative changes?
Uhm, no, you look to it from a point of view i didn’t meant with my point. What i meant by saying that, is that the Linux kernel developers have to keep constanly in mind those binary drivers and users complainig when a change was made. This costs either time via support/noise, or changes aren’t made because of this, or people frown upon them when changes were made.
http://marc.theaimsgroup.com/?l=linux-kernel&m=108924181725688&w=2
Is related to it and is noise.
http://marc.theaimsgroup.com/?l=linux-kernel&m=108782184709762&w=2
Noise and no support was given hence the user went to LKML.
http://marc.theaimsgroup.com/?l=linux-kernel&m=108427539920124&w=2
Noise as a proprietary module was loaded hence no support is given. Also consider it less easy to reproduce.
http://marc.theaimsgroup.com/?l=linux-kernel&m=107036294804223&w=2
Another noise, see the last line.
Also, consider adding the other taints as well given those for ther very same reason also lower the signal-to-noise ratio.
Finally, i’m missing some aspects. The author says he had several discussions with the kernel developers regarding his module. I’ve not been able to find those (yet) while they may give other insights and details. Anyone who wishes to draw conclusions should read these.
However, note that after those URLs i stopped searching because i’ve found out something several more interesting aspects.
—-
Also interesting:
Having a hook in the kernel (in GPLed code) for the explicit purpose of allowing a binary module is not allowed. Go read Linus’s statements about this in the archives. — Greg KH.
http://marc.theaimsgroup.com/?l=linux-kernel&m=109356538012562&w=2
That seems to be the reason. A design reason. This is easy to circumvent.
As said by Andrew Morton:
Well yeah. Everybody hates seeing functionality which is useful to people go away. But the license has served us well thus far and if we’re to go bending further than we have for practical reasons, it should be an up-front decision rather than a backdoor sneak-in. — Andrew Morton. Note how this refers to GPL purity and again not “i want everything to be GPL” kind of zealotry. Some people consider it illegal to have a binary only license as module whereas others don’t. This is what Andrew refers to.
Also, he said:
In this case, special-purpose hooks for a binary-only module are a no-no.
Leaning on the binary driver vendor is the best approach although speaking realistically, someone needs to get down and implement a driver which uses the normal kernel APIs and which wraps the binary module (cf: nVidia driver). That doesn’t sound too hard.
Source: http://marc.theaimsgroup.com/?l=linux-usb-devel&m=109357481719361&w…
This means:
1) This binary-only driver is different than the ATI and NVidia ones.
2) There is no discrimination; actually its consistency
3) If the author wants to he’s able to do this but instead he opted for stopping any kind of development or support and put up a rant on his homepage without referring to the discussion and without referring to the other side who argues otherwise (argues otherwise with some valid points worth to be considered, if i may add).
4) As a side note to #3, he still has the option to talk to Philips, or consider the NDA expired thus legally able to GPL the code. Which hasn’t been considered nor addressed by the author.
—-
Finally, those users who do not agree with this have the ability to start a petition and give that to the Linux kernel maintainers and to the author of the proprietary driver, Nemosoft. Doing nothing and complaining here is not constructive, but showing you care does. Who knows what will follow after that, maybe the hooks are placed back, maybe the author will take the NVidia-approach, or maybe it’ll be an argument why the current development process lacks respect to the users (but claiming the latter right now is rather pointless cause we miss pieces of the puzzle.)
—-
Quite frankly, i feel pity for the clowns asserting all kind of crap (various conclusions and global statements) while the discussion is still going onwards, while facts are still not clear including the very clown who posted this on her circus-playground. Shame on you, always with those conclusions and rants ready :
… is not (cheerfuly and fortunately) and should be not our own.
“… but if I can’t be 100% Free, I’d take that 50% any day. Common users (the kind of mammal species that the Linux platform is trying to captivate away from Windows/Mac) want cheap and easy solutions, and some Linux maintainers fall short in materializing this need because of this or the other (irrelevant to most users) reason”
I for one am not happy when someone has a really stupid idea and then shouts it out loud thinking (s)he has a point. Thinking it over a bit may help you know. Some broader view and understanding of OSS and Linux (which from your writings I thought you had up to now) would be useful when trying to sell bullsh*t for the rest of the people. Anyone can tell rubbish, it isn’t rocket science.
And you know what ? “irrelevant to most users” would be a nice idea if and ever you knew what we (you know, other people out there in the wild) think and want. If you ever took a bit to think about why Linux is where it is, why the kernel is how it is, why the kernel dev process goes as it is, or how third party fully or partially binary drivers are used in Linux.
Irrelevant ? Yeah, guess who I think that is.
Razor’s just popped into my mind: never attribute to malice that which can be adequately explained by stupidity. Take this by word.
See my above post and consider my 2nd quote of Andrew Morton:
Leaning on the binary driver vendor is the best approach although speaking realistically, someone needs to get down and implement a driver which uses the normal kernel APIs and which wraps the binary module (cf: nVidia driver). That doesn’t sound too hard.
Source: http://marc.theaimsgroup.com/?l=linux-usb-devel&m=109357481719361&w…
Here’s the patch: http://marc.theaimsgroup.com/?l=linux-usb-devel&m=109078730609347&w…
I don’t know where the binary-only driver is. Hopefully someone else does have a link to that.
Linux-usb-devel has a whole lot more information on the subject, too. I’m checking it out now.
PS: Leo, i’ve found some more noise on the list, but i’m conceding on that point because i’ve found out it ain’t the reason of this happening.
Here we are:
ChangeSet 1.1843.4.21, 2004/08/24 15:58:54-07:00, [email protected]
USB: rip the pwc decompressor hooks out of the kernel, as they are a GPL violation.
[…]
http://marc.theaimsgroup.com/?l=linux-usb-devel&m=109356614909456&w…
THE proof it IS seen as a license issue. Now, that makes quite a lot of comments useless, doesn’t it? I’ve had this feeling from the beginning, but what the heck. Eugenia’s rant is totally out of line because of this. Had she read ONLY this mail, she’d have at least put a link on it and i hope some other point of view (or rather, not posting her rant). As i said, she failed to research the other side of the story. I guess she hasn’t even read the mailing list about this. Is this quality journalism? :\
I thought open source was all about freedom, about choice? This way, Greg Kroah limits the choice of the Linux user because of his personal beliefs. A lot of companies aren’t gonna like this. Also, I find it rather childish. When is that guy going to realize that what he is doing to the kernel doesn’t affect only him? There are millions of users out there, and he has respnsabilities towards them. Yet he doesn’t seem to care.
Millions of users of this driver? Doubtable.
If you only read the mailing lists or some of the more valuable comments here you’d have found out it there ain’t political reasons and that there are solutions to the problem which currently aren’t considered. The discussion also ain’t over yet.
But you’re too entitled to your opinions about open-source already, evading the possibility to grok the subject. I find your anti-open-source opinions also quite useless, but you problably know that already. Btw, when are you gonna post your scientific research regarding X in the Haiku thread? Perhaps you’re able contribute something unique to the tech-world by doing so. I’m looking forward to them.
I think Philips don’t realize the importance that “Linux compatible” has for them. I think they see the big numbers and they forget about the trends. Linux is somewere between 1% to 5% right now all over the world. But what percentage of geeks that buy Webcams use Linux? I guess that can be easily now a 7 to 15%. And in a year? How many of them will use linux? probably 20%. Surely not, but this could happen. And how long will it take for them to have a good enough open source driver to make into the kernel tree? Bad documented or lousy writen code don’t get into the kernel.
And even when they manage to have their code into the kernel, how much time wiill it take for the [let’s say] 2.6.17 version of the kernel to get into the mainstream distributions?
I’m not telling that will happen, but it could happen easily that, in a year, they discover some small company is selling a lot of webcams because they do have a Linux 2.6.8 and above Compatible” logo at their product’s cover.
When I started in the linux world, installing a network card whas quite a nightmare, because lots of network card companies acted like Philips. But right now, you simply forget about any possible problem with the network cards (eth). They simply work out of the box.
If someday we accept closed drivers into Linux, as newcomers would like to, it will happen the same that happens with Windows. To reinstall Windows you need the Windows CD, the mouse CD, the keyboard CD, the Webcam CD, the DVD CD, The CD-Writer CD, the printer CD, the scanner CD, the nVidia CD and to restart your system a dozen times in order to install it all.
Please, newcomers, you are very welcomed to our comunity, but let us continue acting “fundamentalist” because this is the way we’ll manage to have a better system for everybody. We know it is very disturbing when some device is not recognized. Be patient. Some day it will be.
you seem to asume that binary drivers are stable. and give windows binary drivers as an exampel but that dont add cus microsoft inspects third party drivers and sign them linux kernel devs have no possible way to do this with binary drivers for linux.
the nvidia drivers are mostly stabel but som versions of it have ben really crappy ati binary drivers arent that good.
Millions of users of this driver? Doubtable.
I was talking about Linux users (to your surprise: that includes myself!) in general. If he does this to a webcam driver, what prevents him from doing it do other drivers too? Removing hooks as he pleases because his beliefs are in the way? Of course he has the right to respect his own beliefs (excuse the weird sentence
) but then he should’ve just removed it while compiling his own kernel, not the kernel millions out there use (and enjoy).
If you only read the mailing lists or some of the more valuable comments here you’d have found out it there ain’t political reasons and that there are solutions to the problem which currently aren’t considered. The discussion also ain’t over yet.
Of course not, I never said there aren’t any solutions. It’s just that at the moment, there ain’t one available yet. At least, not an easy one.
But you’re too entitled to your opinions about open-source already, evading the possibility to grok the subject. I find your anti-open-source opinions also quite useless, but you problably know that already.
I’m not anti-open source (something I tried to emphazise around here very often). I just don’t follow it blindly without questioning it all. If you can’t stand critique, then that’s your problem. Please don’t bother me with it.
I am a Linux users and have been at times been over zelous in my support and promotion of Linux. I have read the many comments and my two cents are: I think that this is not good for Linux. In this case purity of the kernel be dammed. The kernnel maintainers should be hung out to dry. If this was such an issue then they should have contacted Philliphs and tryed to get them to produce a binary only or open source solution. They could have worked with Philliphs and the current developer. But no they acted like misbehaving brates and pissed on us all. My two cents. This stinks.
you realize that to grants someone’s new freedoms you often have to take away someone else’s rights/privileges.
All are right. There is no one wrong here. The Linux guys want a pure OS that complies with the GPL. The users want their stuff to work with Linux. The coder wants to fill a need. And the company want to protect what they licensed/made.
Nemosoft guy should switch to BSD. He shouldn’t stop coding his drivers. Making stuff is fun
But no they acted like misbehaving brates and pissed on us all. My two cents. This stinks.
—
do you know why greg removed the hooks?. did you even try to understand. read lkml before you jump out lashing the kernel developers.
There are a few implementations of Userland drivers out there, and I think that it’s time we allow them as an option in the mainline kernel.
It’s not like the PWC driver requires terribly high bandwidth. It has been found that Userland drivers do not perform much worse (sometimes better) than kernel drivers. Even IDE drivers can be in userland without much of a performance hit. By LGPL’ing the driver library, we allow third party drivers to be written that work across all distributions without a recompile.
To summarise some of the benefits of Userland drivers:
– Avoid license issues as indicated in the article
– Can compile and distribute drivers across all linux distributions supporting userland drivers. No more requiring a user to compile a driver for their specific kernel!
– Performance comparable to in-kernel drivers. Slightly worse bandwidth/latency but nothing that would be noticeable to the user.
It should be noted that the ndiswrapper is a userland kernel driver. The windows driver is running in userspace. Imagine that! Installing a windows driver in Linux is _easier_ than installing a native Linux driver, because no recompile is necessary.
I am a staunch Linux supporter, and I use it as my primary OS. But the whole “thou shalt not allow binary modules” really annoys me. I’m an engineer, and I can wholeheartedly understand why some companies can’t give out the source. Sometimes it’s not possible, we have to face that reality. Sure, put pressure on so that 95% of drivers are Open Source. But really.. be practical!
” If he does this to a webcam driver, what prevents him from doing it do other drivers too? Removing hooks as he pleases because his beliefs are in the way?”
If you had read the lkml you would know that this was the _only_ instance were hooks like these in the kernel were required and that this was exactly the problem.
You would also have noticed that he didn’t remove them because his beliefs are in the way but because of legal problems.
Let’s say you live in an area that’s 10% forest, but it’s getting more populated and folks want more houses, stores and schools. So you “compromise” and take just a bit of that (2% of the area, let’s say) for development. And 10 years later, you need to do the same again, except this time you need to grow by 3%. Keep this going by little bits and pretty soon, no forest – just another urban sprawl same as all the rest, so that all the folks who moved to the area years ago to enjoy the woods now wonder why they came.
Let’s say you’ve started using Linux because of all the choice, the freedom to tinker with or change nearly anything to suit yourself, the openness about security matters – all the little and big ways that Linux and Linux distros differ from proprietary kernels and OSs. Of course, Linux is only, let’s say, 5% or so of the software landscape. Now you run into this really cool hardware you’d like to use, but it requires putting something proprietary in the kernel. That’s OK, it’s just a bit of your system. And (because of the pace of tech development) in 3-6 months or less, there will be another amazing game or cool piece of hardware, then another, and another…. Then one day you’ll look at your system and notice it’s full of proprietary stuff you can’t change, the same “urban sprawl” that exists everywhere else, and you’ll wonder why you ever went to the trouble of using Linux.
From Logitech website(Quickcam Pro 4000)”All QuickCam cameras, including those produced under the Connectix and Logitech names, are not supported in non-Windows operating systems.”
By buying such a product we are perpetuating and endorsing this arrogant policy with our hard-earned dollars, which seems to be the only thing that matters to such big corporations these days. (and I only use Logitech as an example since I was stupid enough to buy a Pro 4000). Instead, let’s buy from companies who appreciate our money enough that they won’t try to dictate what O/S we must use. Then send a letter to their winhardware competitor(s) to let them know they lost a sale because of this policy.
“In case the answer is “No”, then I will:
– demand that the PWC driver is removed from any further Linux kernel
releases; Open source or not, it’s still _my_ work.”
So this guy doesn’t understand how open source works ….
Of course he has the right to respect his own beliefs (excuse the weird sentence
) but then he should’ve just removed it while compiling his own kernel, not the kernel millions out there use (and enjoy).
His belief is based on a legal issue. If you believe he is wrong, feel free to make a constructive argument as reply to that (but as unconstructive as that X rant of yours, please). Personally, i take a belief of a lawyer or paralegal more into account regarding legal issues than that of a home user of the software, or from a kernel maintainer, but i’m open for a pleasant surprise.
Otherwise, i humbly suggest you don’t comment on the issue. You seem to be unaware of the issues involved, not wanting to learn about it or unable to understand it. I agree with Ralph. To learn about the issues, check out LKML. Several readers here have posted links to some arguments.
Of course not, I never said there aren’t any solutions. It’s just that at the moment, there ain’t one available yet. At least, not an easy one.
Why make conclusions as if the world is falling down? Such conclusions are based on negative assumptions while they also do not take into account Greg’s reasons which is quite unfair. This is only a momentum. In the case where a conflict is to be resolved, one should not become hopeless during that conflict for the final resolution is still open and might be regarded as a positive one. Because of this, and also because of the rant and one-sidedness of Fraulein Eugenia, this article is of bad quality.
I just don’t follow it blindly without questioning it all.
Are you sure about this? What have you read? A lot of people are blindly following what Eugenia and/or the author of PWCX wrote without researching futher. Futher checking out the other side of the story. (Sides, actually, given there are multiple viewpoints.)
Eugenia only linked to the story of the PWCX driver, not from the kernel maintainer’s point of view neither were in depth arguments nor the discussion referred to. This is a one-sided view of the problem, and many comments are already based on this. This is a quite common, problematic behaviour and reflects bad journalism. A good journalist tries to apply an objective point of view, researches the subject, shows various arguments of those involved in the conflict, links to these arguments, and arguments his/her own opinion based on solid ground. Eugenia did none of this. A pity.
(I find Eugenia also a bad journalist and i wish i had better alternatives than Slashdot, Groklaw and OSnews because i find the quality of the editors systematically far from acceptable. Luckily, i found a few and i’m keeping an eye on these.)
If you can’t stand critique, then that’s your problem. Please don’t bother me with it.
I love constructive criticism, but more often than not, that seems to be impossible on this website. Leo, for example, made some constructive comments in this thread with which i mostly didn’t agreed, but at least they were constructive. This is very much appreciated. You are just repeating what many other readers already said, and as it seems, without taking the other side of the story into account. Pretty sad, IMO.
(but NOT as unconstructive as that X rant of yours, please)
Thank you.
they are right, the kernel should stay 100% free, always!
that does _not_ mean you can’t have your philips webcam drivers. the nvidia driver is also not included in the default kernel, however a lot of people are using these cards with linux wo problems. just make packages available as seperate releases.
my take is that not the users are losing, but philips has lost in this case.
> But really.. be practical!
If you ignore the religious aspects, there is a practical side to the open source driver mandate. Here are a few:
* Binary drivers are not portable. The x86 is not the only game in town.
* There is no way to audit a binary driver (for security and privacy issues — do you want root-level spyware in the kernel?)
* When there’s a bug in a binary driver, no distribution can fix it. (This, and the previous, reason are why several distributions don’t ship with binary drivers)
* The while issue of binary GPL modules is rather questionable. Linus’s interpretation of the GPL is that they’re possible, but his GPL licensing of the kernel makes no explicit mention of it. More than one kernel contributor, including the 2.4 maintainer Alan Cox interpretes the GPL under more restrictive terms. When you combine restrictive (pure GPL) and non-restrictive code (GPL+exceptions) together, you get restrictive code.
Userland drivers help, but they they’re still a pain in the neck. Take a look at the Fedora Core 2 and Gentoo mailing lists. Notice how many posts are about difficulty installing or using VMware with these distributions. X.org (a userland app) is one of the key stumbling blocks. VMWare’s support has been excellent and I have high praise for VMWare, but even it can’t keep up with all the changes that go on in the X and kernel world.
If VMWare worked more closely with the kernel and X hackers, it might be able arrive at a general interface so the X and kernel hackers would maintain the VMWare drivers
How am I supposed to be a good little troll if all the trolling happens in the article texts?
Nemosoft asked Greg to remove the code.
Legally of course the license has been given and cannot unilaterally be revoked. But his name is on the driver and he gets the mails about it. So unless somebody takes over full maintenance, he should be allowed to shoot his own dog and Greg has announced that he would take such a patch.
See http://marc.theaimsgroup.com/?l=linux-kernel&m=109352140027810&w=2
Greg just did what he was asked to do: removing the code.
Nemosoft therefore did not consider other solutions such as:
* NVidia-like wrapper solution;
* Contacting Philips wether the code may be GPLed since the NDA is expired;
* Releasing the code under the GPL since the NDA is expired.
While he did say the NDA was based on trust and he doesn’t want to release it under the GPL because the NDA got expired i’m not accurate when i state he didn’t consider that, but it seems the other 2 options are evaded. Why?
I find it hard to blame Greg for doing something based on both legal and a technical logic while it still opens at least 2 solutions to the problem created by this, which the author of the PWCX driver doesn’t want to follow. Why is this?
Separate from the above debate around the removal of the hook, following the request for the driver’s removal by its author it was pointed out that the driver is GPL’d, so the source code should legally be able to stay in the kernel regarldess of its author’s requests. Linux creator Linus Torvalds replied, “yes and no. From a legal standpoint you’re right. However, we should also be polite. If he’s the sole author, and he asks for it, I think it’s reasonable to honor his wishes. Of course if some new maintainer shows up and decides to infer how the device worked by looking at the original open-source code, that’s also clearly fine. I don’t want people to play lawyer. Honoring peoples rights to the code they write is more important than just the law.”
Source: http://kerneltrap.org/node/view/3729
While you’re at it, also check out how Kerneltrap.org reported on the issue (see URL above). Compare it with how OSnews reported on the issue. Do you see the difference? Do you see how they do link to various sources and do try to be objective? At least, that is how i perceive it. Kerneltrap is reporting; OSnews is journaling. Kudos to Kerneltrap for writing such an excellent article, although adding the NVidia-wrapper solutions as well as the NDA being expired wuold have been interesting details as well.
Hardware is often proprietary and so are the drivers. Don’t we yet have a way to go round that fact of life ?
Polite….wow you gotta love that ’em for bein polite even if you hate the politics.
The QC Pro 4000/Sphere and PCVC840k has both great quality when compression is enabled. Are there any alternatives that can compete with them? CCD and compression is a must over usb 1.1.
What is the point of keeping a non functioning driver in the kernel tree ? Why keep the driver if you remove the bit that makes it useful ?
Phillips is well within thier rights to withold whatever IP they want from whomever they want , just as those who are responsible for the Linux Kerenel should be able to say what does and does not belong in the Linux kernel.
Theres actually been many complaints with some aspects of kernel compatibility recently, the main being ABI compatibility..
What really needs to be done is that someone needs to really fork the kernel and focus on ABI compatibility to encourage mainline to do the same..
Its only a matter of time anyway, and the thing is, that such a fork would probably gain more attention off people then the actual kernel itself and force linus to reconsider his priorities.
Doing so would also revolutionise linux installs too, as finally ppl wouldn’t need to wait for a newer distro release to support their hardware sometimes during install.
yes it would encourage some closed source driver development, but in some cases companies can make better drivers since they know their hardware well, and also wont penalise drivers that must be kept closed source due to legal restrictions
I don’t agree with the users being the losers here. Yes, probably quite a few users will end up with either webcams not working or running an “outdated” kernel (Which really isn’t that much of a problem, I’m still running both 2.0 and 2.2 kernels on a couple of servers). If people quit chasing the “latest and greatest” all the time, they could still use their cameras. The driver is probably available from other places than the authors site.
For those wanting the “L&G”, DON’T BUY PHILLIPS.
In the end, it is phillips that will lose this. By opening up their driver, they would basically guarantee the inclusion of that driver into the kernel, and continued maintenance.
I, for one now know that if I need a camera, I’m not going to buy phillips. If phillips does not want to support the hardware I’m considering under my primary operating system, then screw them.
(Anyway, someone will probably be enraged enough to hack a driver together during the next few weeks)
From
http://marc.theaimsgroup.com/?l=linux-usb-devel&m=109357892107844&w…
> On 04.08.26, Greg KH wrote:
> > Greg Kroah-Hartman:
> > o USB: rip out the whole pwc driver as the author wishes to have done
> > o USB: rip the pwc decompressor hooks out of the kernel, as they are a GPL violation
>
> The decompressor hooks may be a Linux kernel policy violation,
> but I challenge your contention that they are a GPL violation.
Doesn’t matter. Whether they are a GPL violation is a gray area. They were
removed because of a policy. The author then complained, and the _driver_
was removed for that reason.
At no point was it a legal argument. In fact, since none of the people
involved were layers, you shouldn’t even try to _make_ it a legal
arguments.
You do legal arguments in front of a judge. Until you reach that point,
you do what’s right. And I think Greg did what is right.
Now, we can whine all we like about the author being childish, or about
the fact that since it was GPL’d, the hooks can clearly legally be removed
regardless of his wishes. That’s not what this is all about. There are
more important things involved.
I repeat: if somebody wants to step up as maintainer, I would certainly be
more than happy to have a pwc driver, and I bet Greg would be too. But you
don’t just take somebody elses code against his wishes – regardless of
whether you have the “legal right” or not.
Let’s put it this way: if you need to ask a lawyer whether what you do is
“right” or not, you are morally corrupt. Let’s not go there. We don’t base
our morality on law.
Linus
@Ximinez
In the end, it is phillips that will lose this.
They’ll probably loose something like less than 5$. Linux desktop users are few and not a priority to support.
What happened is that those few Linux users just lost choice because of lack of flexibility of license and for political’s sake.
And that’s the bottom line.
Why does all the drivers have to be included in the kernel package?
Instead of trying every single driver to see if it works correctly with the kernel, wouldn’t it be simpler to just create an interface that would allow external drivers to work on the kernel level? This probably already exists, so why not make use of it?
I mean, all other OS allows drivers to be loaded during runtime. Why does the drivers have to be included in the kernel package?
“Being Free is a good thing, but if I can’t be 100% Free, I’d take that 50% any day.” Is a ridiculous quote for this situation. It is obvious the code was removed because of design. Like one poster said, nvidia can make a closed module just fine without puting any code into the kernel. I dont want to sound like an oss “purist”, but let phillips sit on their “sekrits”. The code removal was plenty justified.
Oh, and the subject of stealing users from windows by holding their hands is old. Linux wont just disappear without your grandma using it or “that kind of mammal”.
…and Linus decided to respect his wishes:
http://lwn.net/Articles/99617/
“..if an author asks us to remove something, I think we
should strive to follow his wishes.”
Linus
After reading most of the comments here:
1. Keep Linux free. I agree with the Linux maintainers on this one. They are duty-bound to uphold the idealogy whether “we” like it or not.
2. Keep Linux off my computer. I don’t care whose fault it is, but if a piece of hardware doesn’t work, then tough.
And the real losers? The Linux community actually. Unless Philips sees economic viability is writing a driver for Linux, then they won’t do it. Clearly they didn’t because when someone offered to write a driver, they said “sure, but here is the NDA”. They got Linux support, and they didn’t have to do it. So we shouldn’t expect too much from Philips (the optimist in me says maybe this fiasco would motivate them otherwise).
So without Philips, Linux loses support for one important piece of hardware. That’s one extra black dot against Linux. So who lost? The Linux community – there is one less person. Not the users – they have Windows.
crap, the “I dont want to sound like…” sentence should be in the bottom paragraph.
I don’t think the luser loose anything.
And it would be nice to have some non-opinative articles to begin with! press should inform, not manipulate audience. I find that OFFENSIVE.
well. about the driver. I would NEVER feels confortable having a builtin driver for a camera running in my server. What if the closed source owner is in a desesperate state and resolve to undermine the competition by putting some code that will crash when detecting competition hardware on the machine.
You can call me paranoic, because i’m paranoic. So, on my machines, i want the most clean, audited code ever.
Before i buy some hardware (be it a camera) my top priority is compatibility. All my canon cameras uses compact flash cards that can be readed by any device and accessed trhu STANDARD usb-mass storage.
If i ever buy something that’s not STANTARD (like sony cameras) then i would KNOW that i will have to download closed source drivers and install them… isn’t it so with windows? so, what’s wrong? wipe that crap of the kernel and move along. The author removing it from the site was CHILDISH. Now should HP remove every drive from their site that isn’t distributed within windows?
Want to make Linux a micro-kernel OS ;o) ?
Just becouse all the drivers wasn’t included into the kernel, wouldn’t make linux a mirco-kernel.
And to the post you replyed to…:
Why wouldn’t you have as many drivers into the kernel-source as you can? There is still much room for more drivers (modules) like the nvidia’s driver.
His belief is based on a legal issue. If you believe he is wrong, feel free to make a constructive argument as reply to that (but as unconstructive as that X rant of yours, please). Personally, i take a belief of a lawyer or paralegal more into account regarding legal issues than that of a home user of the software, or from a kernel maintainer, but i’m open for a pleasant surprise.
What exactly has changed yesterday that made him decide to remove the hooks all of a sudden? For three years, the hooks were in. Yes, Kroah already made points of it. But yet, he kept them in. But suddenly, he decides they’re out. I think, that if you are really worried about the partial closed-sourceness, then he shouldn’t have let them in AT ALL. It’s just simply unfair and unacceptable to remove them after three years. He should’ve been real, and never have let them in at all.
Otherwise, i humbly suggest you don’t comment on the issue. You seem to be unaware of the issues involved, not wanting to learn about it or unable to understand it.
If it were up to you, so it seems, everyone not agreeing with you shouldn’t comment on this issue. But, then again, I know I have troubles with getting along with open-source fan(atic?)s. No big deal, as long as we keep it civilized, and only on the area of software
.
“I just don’t follow it blindly without questioning it all.”
Are you sure about this? What have you read? A lot of people are blindly following what Eugenia and/or the author of PWCX wrote without researching futher. Futher checking out the other side of the story. (Sides, actually, given there are multiple viewpoints.)
I DO understand WHY he has removed those hooks, my problem doesn’t lie there (well, for a bit it does, but that’s a personal issue*). My problem mostly lies in the fact that for THREE YEARS they were in, even though he obviously didn’t agree with it. I simply find it unacceptable to leave them in for three years while disagreeing, and then suddenly just remove them. That’s not fair towards the users and the developper.
* I’m personally kind of fed up with all the licensing “nonsense”. Of course I fully understand the whole world probably is, and of course I understand licenses are needed and all. But you have to agree with me, they seem to cause troubles and worries all over the place (SCO, Haiku, now this). At least with closed source one doesn’t have to worry about whether or not they follow public licenses
(but there are a million other things to worry about, thus equalling (is that a word? I’m Dutch) the scale 
. )
Quoting totally out of context, using wit to prove a point yadda yadda, but here goes:
“They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.”
— Benjamin Franklin
Yeah, mod me down for quoting a dead person.
“I think more than anything this serves as a reminder not to rely on non-GPL code, you never know when the author/company etc. is going to stop supporting it.”
BINGO! .. you’re right 100%. Thanks to the kernel developer’s for enforcing this. This is a great example of how a closed source program controlled by an individual or company can try to hold you hostage. I thank this individual for his years of work but to put it bluntly – “good riddance, and don’t let the door hit your ass on the way out”. If the code was GPLed than someone else could of taken over maintainership or they could fork the code if a dispute arose and the users could decide which is the better program. An example of this is Xfree86 and X.org. X.org forked XFree86 because of their differences and the USERS have decided which they will use – X.org I hear is doing fairly well. The GPL gives control back to users (and yes all users even if you’re not a programmer). So to the ex-maintainer: Thanks for the memories …
They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.
Well, what else do you expect from freeloaders?
On the one side, they neither wont to pay for a proprietary product, where there are no such problems, due to its proprietary nature, nor do they want to stick to some principles (Free Software) and for example do not buy a product that doesnt come with a Free Software linux driver. But no, instead of both, they chose to take Free Software just because it happens to be gratis/umsonst, and then bitch like bitches when the principles this software is built on causes them even the slightest inconvenience.
Example: XChat recently went the shareware way. Well, not exactly the usual shareware way, but the developer chose not to provide Windows binaries on his site for free, but now wants a small shareware fee, just for providing these binaries, which are still under the GPL, and come with full sources. His user forums have been filled with XChat abandoning threats and the most disgusting bitching. Just take a look for yourself at the user forums at xchat.org.
The kernels that debian packages are patched to remove features that are incompatible (in terms of license, patent etc) with their guidelines anyway, clearly this hasn’t held anyone back too much.
I would say here that the kernel maintainters made a good, if not entirely necessary choice. The issue could probably have been ignored indefinitely, but that doesn’t mean there was anything wrong with the decision to act. The driver maintainer was just childish to clear off entirely though, lots of people work on drivers outside the kernel tree and don’t act indignant about not being included.
And a question for anyone more knowledgeable than me, would this situation be different in a microkernel OS? Clearly, the driver wouldn’t be linked into the GPLd kernel code, but it would still be interfacing with it. I don’t know the GPL well enough to know what exactly is allowed.
His belief is based on a legal issue. If you believe he is wrong, feel free to make a constructive argument as reply to that (but as unconstructive as that X rant of yours, please). Personally, i take a belief of a lawyer or paralegal more into account regarding legal issues than that of a home user of the software, or from a kernel maintainer, but i’m open for a pleasant surprise.
What exactly has changed yesterday that made him decide to remove the hooks all of a sudden? For three years, the hooks were in. Yes, Kroah already made points of it. But yet, he kept them in. But suddenly, he decides they’re out. I think, that if you are really worried about the partial closed-sourceness, then he shouldn’t have let them in AT ALL. It’s just simply unfair and unacceptable to remove them after three years. He should’ve been real, and never have let them in at all.
Thom, you often sounds like you seriously lacks some experience in development and how organizations are conducted. On the development issue, have you read previous posters pointing to LKML threads that states clearly that those “hooks” now were there just for this driver alone and that’s nonsense to keeping them in the kernel any further? Do you know how hard is to develop a really good and clean code allowing exception of any sort all the time?
On the organization issue, I haven’t seen any evidence that this Greg Kroah allowed them (the hooks) in in the first place. Probably, he wasn’t even the USB stack maintainer back then. I’ll concede that it was up to the previous maintainer, but even then, the author had the option to create it (the driver) as a module – just like nVidia – instead of going the easy way.
But I really do understand your frustation. Even Eugenia’s, but her words were a bit misguided, though. It’s truly sad to see a perfectly functional driver go away, no matter what it do to work. The author, alone, probably can’t provide the man power needed for such overhaul of code and the NDA things surely prohibits him of open source the binary parts of his work (even if it is already expired… These things never are clearly enough).
Maybe Phillips will step up and open either the code or disclose the specs needed for someone create a new driver. And maybe not. But, for the long run, it will be better this way for the Linux Kernel. I’m not sure if I want to see another cat fight like that SCO is promoting, even if they are up to nothing, just because of small bits of proprietary code in the kernel.
It’s not like Linux will die without the help of the big companies or something like that. It is where it is because of its merits to the point that IBM, SGI and plenty of others companies are happy to (now) release their IP in order to offer what their customers are asking them to.
Anyway, this article came in a good moment… I was wondering about Linux-compatible webcams to purchase one in a few days… Now, I know that Phillips’, Logitech’s and any that relies in that pcwx.o driver won’t be a good choice. 🙂
Best regards,
DeadFish Man
Hi,
I’ve got one (1) Webcam for sale (Phillips is the brand). Works only on yesterdays platform (tipical Philips).
For those who are saying Linux represents a few percentage for manufacturers like Philips, consider this. Countries like Brazil who recently use Linux as their standard OS will never accept that an individual manufacurer will held the kernel hostage. Philips is only a single company therefore people will another alternative by selling their unsupported Linux webcam.
Should be kept out if they are proprietary.
Lets not forget why we even are having this discussion in the first place: Linux is an Open Source Kernel.
Sorta Oxymornic I think if it becomes a Open Source Kernel with Proprietary Kernel Hooks(tm).
I bet Bill Gates loves the sound of that, or is that Darl McBride?
It even sounds bad….GAD. Why is this fundamentalistic poo? The reason why the author could even write a camera driver for his 15 minutes of fame, as he puts it, ie because the kernel comes with the source code. So he decided to just yank the entire thing.
?????
It sorta begs the question who is being extreme here with that sort of logic.
What he should do, is if he truly is not interested is to simply turn over the source to someone else who is a little less extreme….err…fundamentally less poo than he is.
I know I very much appreciated
Ah, yes, the power of source code will not be denied…as for phillips, I will just get another camera….just that, it won’t be a phillips one.
If that is OK for Phillips, that is OK for me.
-gc
PS: Mental Note: Do not buy phillips products if possible.
I’m sure its already been said, but I think this is great. Its better to be safe than sorry.
When Philips wants to sell their webcams to Linux users they can write their own drivers and support their own product or deal with the lack of good drivers because of their own secrecy and unwillingness to provide documentation.
I have a Toshiba laptop that still crashes while playing video because it uses a Trident chipset that cannot be fully supported by my OS. This is not the fault of Linux, it would have full support if the developers had access to documentation, so I will never give my money to Toshiba or Trident again. They will have to rely on other customers to purchase their products, and they better pray those customers don’t talk to me first. I paid $1200 for a system with the capability to do decent 3D graphics and smooth fullscreen video playback streaming from my network, but I can’t use these features because they are Trident’s proprietary Windows-only features. My next laptop will have full support for my favorite OS. I have many architectures and brand names to choose from.
Fortunately I can still use my favorite OS on this laptop for 90% of the things I want to do with plenty of performance and features to make me happy. I’m just disappointed in commercial organizations for not giving me the full value of my dollar, by keeping secrets that would force me to run their expensive prefered software instead of my personally favorite high-quality software.
I am confident these bugs will be fixed. However, I have lost confidence as a customer. I will be sure to make a more informed decision next time.
If you want non-free software with good hardware support, try Windows. If you want free software and will accept some difficulties, Linux is a good choice.
I like the fact that the core of Linux is as free as they can possibly make it. Even if I may choose to install the odd non-free userspace app on top, at least the distinction is easy to spot.
Good that we have a choice really. If Linux becomes non-free like Windows, then that choice diminishes. And yes, it is a slippery slope. You make an exception here, why not there too?
I hope Linux hardware support overall improves, but only time will tell. The developers have their own reasons for doing the kernel, and if they don’t want non-free software in it, why should they allow it?
This particular driver author seems to have engaged in a big online sulk by pulling every last bit of info offline. Well good luck to him.
As for Philips, it would be good if they improve their hardware support. I always prefer hardware vendors that try to support Linux.
I switched from Linux to FreeBSD three monthes ago, when this open-source-and-gpl-to-the-limit reached its first non-sense point (remember the XFree86 affair?).
Now we have a “second coming” of the open-source-to-the-limit paranoia with Philips webcams, and soon we will have a third edition of the madness with Apache 2.
How long until Linux digs its own grave?
He said:
“Open source or not, it”s still _my_ work.
– Nemosoft”
Wrong! It is not his work in the sense “I have property rights on my work which I can use to tell you what to do.”
He does not own this code. GPL made him free of any property claims on his code, as well as any obligations for that code.
Yes, he designed the code. To create something and to own something is different. Ford engineer who created my car does not own it. No do I, by the way- I lease it.:)
He wrote the code. So what? A worker at assembly line putting together my car- does he own it? No! Can he tell me what to do with it? No!
Yes, he maintained that code for 4-5 years. It does not give him property or controlling rights over that code. Not more than car mechanic maintaining my car controls it, and family doctor helping me to maintain my body owns me. Not at all!
I don’t even know who designed my car, and not interested to know who put it together. I should not be bothered to know who changed oil in my car last time at the service station. I do not care. It could be anyone with the proper qualification.
GPL liberates developers from silly notion of some kind of ownership. You own code you created and maintained as much as any other member of community. You have no rights to tell what people can and can not do with it.
On the other side, GPL liberated you from any obligations for that code. You maintained it? Good for you. You don’t want any more- who cares! Just say magic words: “I can not maintain my code any more (enter politically correct reason here), I leave it to community to find next maintainer.”
You, as a software developer, are not more than a small part in the OSS machine, a screw, perhaps. You are replaceable by another screw- they all fit.
GPL can not let you keep your silly secrets behind binary code.
So, you want to walk away- walk away, no one will sue you for breach of the contract or for misleading users in respect to support.
GPL finally makes software what it should be: a product of industrial world, not some kind of a product of an artist.
There will be, of course, tensions for some time, while some developers, mostly volunteers, will fail to break from their attachment to the old world of proprietary code and imaginary IP ownership on the basis of their unpaid contribution to the source code.
These tesions should be resolved the way this incident is resolved: by swift measures. You think you can tell community leaders what to do? Well, community leaders will tell you! Obey or leave. Embrace GPL and be free. Free from any intellectual property you think is yours, from any attachment to ones and zeroes. They mean nothing and are irrelevant.
As soon as they understand that GPL makes them equal members of a community with the same rights, no less no more, as any other member- they’ll come back to their senses and will work in software development industrial way.
The way that says: you maintain code today,- I maintain it tomorrow. The freedom way.
“Your source code belongs to all of us”- the true GPL way.
XFree86 affair? its that the one where we got a new project
x.org who are making a cool xserver with feauters that we would never get with xfree86.
in that case wath about it its the best thing that could have happend
Let’s be polite and respect one’s _effort_ to make things work that otherwise wouldn’t. There are lessons to be learned here in order to avoid MAYOR frustration.
“He wrote the code. So what? A worker at assembly line putting together my car- does he own it? No! Can he tell me what to do with it? No!”
Well, that assembly worker owes your respect for the work done. And by the way, you _bought_ that car, or do you think that one was GPL-ed…
however….we all do KNOW that political ideology is pure fantasy….there’s no perfect socialism (that’s why they met their downfall), and there’s no perfect democracy either. And there will be no perfect ‘Free-Software World’ w/o big evil brother like Billy Gates…
I gave up the fantasy few years ago….the world is imperfect, and forever will be. That’s how the world is evolving though…something is not perfect so you strive to make it better…but still no satisifaction and you try again…until you are messed up and wasted up and die wishing for better. Amen.
And if every spec and technical stuff is completely transparent and everyone knows how to do it, then you are completely breaking our economy.
Say, every graphic card vendors gave up their company secret and release every technical info to public as GPLed, then what will be happening?
Some linux zealots will proclaim that the true day of salvation has arrived.
But it will be nothing but the death of graphic card market…
The revolutionary (explosive) development of graphics card market was possible because there was serious competition between graphics chipset makers (say ATi and nVidia) And to make that competition possible, then these companies must keep their hardware’s spec and tech detail as a corporate secret. Without that, every innovation one of them makes will be instantly used by their competitor, and vice versa. And there’s no profit at all…and these companies need to make money in order to continue their development…
If USERS MAKE GRAPHIC HARDWARE THEMSELVES, which is ultrahighly unlikely, then public specs and tech details would be beyond bliss. However, the reality tells us it is not.
So, it is time to wake up from our dream and idealism and face the cold, remorseless and terrible truth. Grow up, human!
Eugenia, you are welcome to your beliefs. If you’re right, and the entire Linux kernel development team are wrong, I hereby grant you three “hip-hip-hurray’s”.
Go talk about somewhere you have relevance.
“But Nemosoft certainly knew that his driver was not conforming to the kernel guidelines. He knew that all along. He did not play by the rules and was not loyal to the Linux developers and community. He should have adopted a different technical solution for his driver or he should never have included his driver in the kernel tree.”
http://kerneltrap.org/node/view/3729
Ok guys who work on Kernel stuff…
Heres the deal. No one wants an operating system you can do nothing with. This is way hobby OS’s have a hard time converting everyday normal people. Disallowing a driver because it is not open source is hindering your task as developers.
Now, before you get all like i dunno pissed at me for saying this i’d like to add i dont have a Phillips webcam, however I am a developer for both Windows and SkyOS(yes one of those fabled hobby os’s). I know that users want featurs, basically taking a driver out of your system because it is not fully open source is crazy.
How you can “calm the storm” with out de-gpl-ing your code is to add runtime driver addition ability. THen you have 100% open source code and you have the drivers users want…
Just my 2cents
“I know that users want featurs, basically taking a driver out of your system because it is not fully open source is crazy.”
The driver was in violation of the kernel guidelines, ie, it isn’t just a GPL issue, from what I have read on the mailing list. The hook should hot have been there. So take the same code, write another driver without the hook.
Your argument is like saying people don’t want to drive cars if they can’t speed. Well, sorry, it’s against the rules. Strange as it may sound, but with the millions of lines of code in the Linux kernel, some lines MUST be drawn somewhere.
“Open source or not, its still _my_ work.”
“Blablabla”
By stating what you do, you proof you are not qualified to tell anything in regards of the GPL. You also proof you do not understand basic copyright laws.
The code is his work. Always applies, no matter what the license is. Note how this is different from public domain. Because it is his license, he can relicense his work anytime he wishes to. However, anyone who has obtained a copy of his work under the current license -no matter which one- has a valid copy of it, under that license, which cannot be revoked.
So cut down the FUD. Go code on Longhorn, fool.
What exactly has changed yesterday that made him decide to remove the hooks all of a sudden? For three years, the hooks were in.
Answer found here in the FAQ: http://kerneltrap.org/node/view/3729
That’s not fair towards the users and the developper.
Consistency is fair for the kernel developers. A common set of rules, applied on non-dicriminational basis, means there is a solid ground. Doing X one day and Y that day (SSP: the Sun Schizophrenia Phenomenon) does not create trust, or a solid ground to work on.
It is easy to make Greg a black sheep, and it is easy to make Nemosoft one too. Although Greg seems to be targetted more. I suggest to look into the various happenings, and check out what lead to what. If you do that, understand aspects such as here above or what Deadfish Man wrote about developmet, it’ll be at least much harder to point out one black sheep.
I hope Nemosoft will reconsider his views and i hope he’ll point out to the other side of the story on his homepage.
So this guy pulls his driver. It’s HIS, right? And the kernel maintainers don’t want any potential problems in the future because of this hybrid open/closed project. So what’s wrong with that?
And, if the driver author chooses not to move forward with the project by modularizing it, again so what? Less work for him, maybe even a chance to use the freed-up time to work on an ever better project. Of course, as several people have pointed out, it kinda sucks for the affected webcam owners, but if it’s that “in demand”, someone else will eventually come up with another, possibly even better, driver solution. Or, perhaps Philips or someone else will release completely new hardware. This will come with either a binary-only driver that some kind programmer will use with a module, or even better, they (the manufacturers) will see the light and release a fully open-source driver.
Of course, I say all of this and yet I’m not affected in the first place. Well, at least, not until I get around to buying a webcam…
He does not own this code.
Wrong. He still have all rights to the code. He is the only one who can license the code under another license than GPL.
He can not take back the GPL licensen from those who have the code or prevent them from distributing it, but he still holds copyright to the code.
The kernel developer’s may be right or wrong as may Nemosoft, but so what? The point Nemosoft was making was that there’s been a lot of hard work getting the NDA, maintaining the kernel driver, maintaining the binary part. He accepts that he could go back and re-do the driver other ways to be more compatible.
But it probably felt like one kick-in-the-teath too many, so he’s moving on. It would either be this, or something else further down the line.
The kernel developer’s may be right or wrong as may Nemosoft, but so what?
—
you realise that maintaining a GPL driver for hooks to a proprietary module is legally a grey area and against the kernel rules and also that the guy’s NDA has already expired?
first of, gpling the source dont mean you cant sell the software. it just means that anyone can look at the code and change it. sure it allso means that someone else can pick up the code and sell it on. what they cant do however is claim they wrote it. copyright and creator rights are maybe mixed up in the public view but they are 2 diffrent beasts (or atleast should be).
and all the hardware creators have to do is release info on what singals to send to the chips to get the right effect. they dont have to tell how a chip is designed or made so its not like the hardware industry would collapse if interface/driver specs are released…
Isn’t this simply the Open-Source vs. Free Software thing?
I am more of the Open-Source idea. I mean, where is making software all about? About “scratching an itch”? Not in the case of Linux. Linux, you see, is being used in enterprises, governments, universities and increasingly at home too. As such, it simply isn’t a hobby project anymore, it has become a product.
And as every product, they should care about the king customer. In the case of the kernel you might think it is not about the money, but ultimately it is. Someone already mentioned the USB developer works for IBM, and IBM is depent on Linux sales. Now if the trust in Linux is damaged by this move, this will lead to less sales, and what would IBM do if Linux doesn’t sell anymore? The only logical thing – fire the responsible people (the USB developer) first, and if that doesn’t help, fire the entire Linux division.
And in the case of really non-commercial products, the customer matters too. Less users (customers) means less feedback, less bugreports, less feature requests and so ultimately worse software.
So in the end, the end user is important to any project. And most end users couldn’t care less whether something is open-source or not, they simply want their PC to work correctly, with no effort if possible. Therefore I think it is a bad idea to remove such hooks if they don’t form a technical or legal problem.
“The code is his work.”
The code is his work as much as my car is a work of men who put it together at the assembly line.
So what? They do not own my car, they can not tell me what to do with it: to stop using it if they don’t like me, for example. That was my point.
“anyone who has obtained a copy of his work under the current license- has a valid copy of it”
Exactly! Thanks for supporting my point. I would only add that not only everyone has a valid copy, but they can take that copy and support it, update it, fix bugs, add features.
It is because the original coder, Nemosoft, was freed from code ownership claims by GPL, and because GPL makes everyone free from property claims of any person with illusions of IP ownership on GPLed software he created.
His source code belongs to all of us. To me and to you, too. He is free to leave, of course. It does not matter for community.
So cut down the FUD.
I think it makes you really unhappy that what I said is the truth. You can not argue with the truth, so you don’t, but you can use bad words to call a person laying down the truth in front of you.
It is ugly truth, but it is truth: a single person does not matter in the GPL software world. You do not matter. I do not matter. Community matters.
A community through its leaders or members can take (copy) a work of any person at any time and pass it to another person (to maintain and improve). It happened before and it will happen again.
Never GPL software will be made a victim of a developer ambitions. Never GPL software will experience a developer lock-in.
You want to write and maintain GPL code? Thank you. You want to leave? See ya. Leave your sources behind, they belong to all of us.
A developer and his GPLed code are soon parted. How soon? As soon as necessary for community. As soon as developer tries to wag the community leaders.
Community rules through its leaders, in that case- through kernel maintainers.
Linus was right. Greg was right. Greg sounded like corporate drone, but it is his job to obey the rules. Everyone wins at the end.
After all, worst case, spending $100 for a new web camera is not such a big deal when rules and regulations are at stake.
Imagine what would happen if a developer of non-GPLed device driver would abruptly cut support for it, compare with GPLed device driver and see the difference.
Ok, yes, its his code, and well GPL says that ppl can change it(one reason I will NEVER release any code under that license, this creates harder to patch bugs and other stuff).
But what your missing is HE had the license from Phillips, if any other persons release anything that uses the binary file from Phillips can be sued/jailed…
Unless your willing to take that risk, I’d face the facts: the drivers gone…
It is ugly truth, but it is truth: a single person does not matter in the GPL software world. You do not matter. I do not matter. Community matters.
A community through its leaders or members can take (copy) a work of any person at any time and pass it to another person (to maintain and improve). It happened before and it will happen again.
So according to what you’ve just said the views held by the glorious leaders of the revolution, oops I mean “community”, matter more than an ordinary individual?
Have you read Animal Farm?
Yes, I’m being picky, but there’s something about this that doesn’t feel totally right. I’m curious as to which side of the moral fence the developers will fall when the first conflict of interest arises. Eg What would have happened if it had been IBM, not Philips?
***
My understanding of copyright law is hazy, but isn’t the original author allowed to change licensing as they see fit for a certain number of years after initial release? Could be remembering something wrong, but I’m sure that’s how it works.
And as every product, they should care about the king customer. In the case of the kernel you might think it is not about the money, but ultimately it is. Someone already mentioned the USB developer works for IBM, and IBM is depent on Linux sales. Now if the trust in Linux is damaged by this move, this will lead to less sales, and what would IBM do if Linux doesn’t sell anymore? The only logical thing – fire the responsible people (the USB developer) first, and if that doesn’t help, fire the entire Linux division.
There is a problem in your line of reasoning. IBM sells servers with Linux. They don’t sell desktop linux. The only companies that you can blame for not having a full featured desktop is a desktop linux company. If you downloaded a linux distibution for free you can’t really demand anything. If you buy a desktop linux distribution then and only then do you have a right to complain, or at least a right to be heard.
I find it odd that people blame IBM for desktop linux when they don’t really sell desktop linux. I hear it a lot in the osnews comments and it really doesn’t make any sense if you think about it. Go ahead and blame Suse and Lycoris and Linspire if you want to but not IBM.
It is because the original coder, Nemosoft, was freed from code ownership claims by GPL, and because GPL makes everyone free from property claims of any person with illusions of IP ownership on GPLed software he created.
No, if you download some GPL code you do not own that IP. Hence, no ownership. Again, you fail to understand the basics of copyright law. Wether the code is released under GPL or the abacadabra license does not matter for the whole discussion.
His source code belongs to all of us. To me and to you, too.
In the sense of GPL, yes. Because he decided to license his code under the GPL. He is still the author though and he all the rights based on the copyright law.
Also if you had read what Linus said, license rather bases himself on morals (“if he wants the code removed, we do that”) instead of the law.
Embarasing how Microsoft still doesn’t understand this…