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
Unless your willing to take that risk, I’d face the facts: the drivers gone…
Mirrors:
http://projects.raphnet.net/#pwc
http://www.xs4all.nl/~jjvrieze/pwc.html
(Posting a link to something illegal is not illegal in the country i reside in <:)
Wow, that’s very mature of you. I wouldn’t expect anything less than from the Linux community though. Do you have an ISOs for SUSE, too? Or maybe CrossOver Office?
I don·t blame IBM for this, in fact, I don·t blame anyone. In fact, I do not even use Linux, I prefer Windows and NetBSD, and waiting for NetBSD 2.0 I have installed Solaris, which also works fine for me.
Yet I still think it is a bad move. First of all, this article isn·t exactly good publicity for Linux, I think. Moreover, if people at home suddenly get problems with previously fully-supported hardware, that will also lead to less trust in Linux. And I think that, somehow, that will influence Linux on the server too.
—
Those mirrors are nice, but they don·t really solve the problem. The first problem is that you need to (re)compile things to make the driver work, which was exactly why the original developer pulled the plug. Moreover, that pwcx-driver is only in binary form so it can·t be developed any further.
Wow, that’s very mature of you. I wouldn’t expect anything less than from the Linux community though. Do you have an ISOs for SUSE, too? Or maybe CrossOver Office?
Ummm….I have several ISOs of various linux distributions, including Suse. And it’s not illegal for many reasons besides the GPL, including the fact that they were downloaded from the distributors’ anonymous ftp servers or authorized mirrors. What’s your flipping point??
Arrgh…I need to learn to stop feeding the trolls…but it’s so addictive!
I don·t blame IBM for this, in fact, I don·t blame anyone. In fact, I do not even use Linux, I prefer Windows and NetBSD, and waiting for NetBSD 2.0 I have installed Solaris, which also works fine for me.
Yet I still think it is a bad move.
So what did you mean by this statement:
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.
????
First of all, this article isn·t exactly good publicity for Linux, I think. Moreover, if people at home suddenly get problems with previously fully-supported hardware, that will also lead to less trust in Linux. And I think that, somehow, that will influence Linux on the server too.
How? How will this affect linux on the server? It won’t and it hasn’t in the years that these hooks have been in the kernel. The fact is that the author doesn’t want it in the kernel anymore. The driver doesn’t follow official policy and the author doesn’t want to fix it. Don’t blame IBM or Linux in general. Blame the author. Complain to the distibution maintainers if you bought a copy of linux. Everyone is using this as a chance to jump on the linux-will-never-make-it bandwagon. The way this issue was framed in the original post makes OSNEWS look more like FOXNEWS. In fact OSNEWS has had equivalant reporting “reporting” styles as FOX for too long now. It’s all about controversy where there is none, for the sake of slandering one side of the argument.
Wow, that’s very mature of you.
Actually i’m not so sure on the (il)legality. Several people asked elsewhere for a link to the binaries, source, so i decided it would be constructive to spread these.
I wouldn’t expect anything less than from the Linux community though.
Nice flame and generalisation.
Do you have an ISOs for SUSE, too?
Hmm, ain’t that legal because SUSE is fully GPL?
Or maybe CrossOver Office?
If you’re interested there’s a demo on their homepage.
So what did you mean by this statement:
Whether the action of removing this hook is a good or a bad idea, it has caused bad publicity, and the most logical person for most people to blame is the USB developer, don’t you think?
Furthermore, I simply cannot believe that publicity has no influence on sales. And I also don’t believe that different markets (desktop/server) don’t have any influence on eachother. Likewise, I am not the only one who thinks that quite some people initially didn’t trust Windows CE, just because of the blue screens of Windows 98.
>By dpi (IP: —.ipv4.freeshell.bofx.net)
“No, if you download some GPL code you do not own that IP. Hence, no ownership.”
Well, Mr. Smart Guy, enlighten us: who owns some GPL code you downloaded?
Can that so-called owner stop you from sharing the code? Can he stop you from maintaining the code? Can he stop you from improving the code and giving improvements back to community?
Finally, can he out of the sudden force you to pay for the use of his GPL code, and IF you refuse to pay- make you stop using GPL code?
May be I am not good in copyright, but I understand property ownership rights.
That is why I said about illusion of ownership of GPL code. You want to prove that GPL code has the rightful owner who can dictate what people can and can not do?
That GPL code has a vendor/developer/owner lock-in?
You must be new to OpenSource. You should re-read not only GPL, but Manifesto too. Like I did, many times- not just read but understand what it is about.
Then, come back and tell us if you still have illusions of ownership in respect to GPL code.
>>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.
My point exactly! He licenced his code under GPL, GPL liberated him and us. He is free to use/work/improve/share his code- so do we. You and me.
It does not mean he is the slave of GPL. He can work on other project that is not GPLed, who cares!
Actually, I believe he just decided to do just that! Good for him. Irrelevant for us. We still have that source code.
The point is- when you go GPL, you are liberated. A vendor/developer/corporate ego or greed can not interfere with community needs. With your and mine needs, mind you.
Embarasing how Microsoft still doesn’t understand this…
I fail to follow you logic and understand what Microsoft has to do with it.
Actually, I fail to follow your logic at all: the logic of not discussing the subject- but trying silly insults at the person who brought the subject. Yes, it is very mature of you.
*********************
@GregV,
“Unless your willing to take that risk, I’d face the facts: the drivers gone…”
I think you are wrong, and GPL will prove you wrong soon.
Here is what I believe will happen, and this is because of GPL:
– A new GPLed driver maintainer will appear, and take the task of maintaining GPLed driver. This can happen any time soon, because it is GPL way. Non-GPL way is: developer is gone- drivers are gone.
– A developer from the country like Brasil, Vietnam, China, Argentina will reverse engineer binary part of driver. You know, there are countries outside USA:), in some of them reverse engineering is still legal.
– A developer from abovementioned country will offer OpenSources GPLed driver code for what was before a proprietary binary. That code can either be supplied to the kernel folks, or if they are too shy- freely downloaded from the Web.
After all, even if in USA it is illegal to reverse engineer binary- any court will find it hard to prosecute someone just for using reverse engineered driver.
Because of all of that, I would actually recommend to buy more Philips cameras instead of dumping them.
The new fully functional driver is near, if you truly believe in GPL, and having many Linux users of a Philips hardware will help to make it happen even sooner.
It is the old, “proprietary” way just to dump hardware when drivers are no longer supported, or stay with older OS where drivers are supported.
Don’t do that. GPL gave you freedom of choice. Use it.
Well, Mr. Smart Guy, enlighten us: who owns some GPL code you downloaded?
As i and others have already explained to you: the copyright is owned by the author as with any copyrighted work. The GPL and BSD license just grant you additional rights beyond basic copyright rules in different ways. It makes use of the copyright laws.
If you don’t understand this, no problem, but is means you do not comprehend that aspect of copyright nor the GPL. It also lowers your credibility when you make related statements to that.
All the other things you say in regard to GPL don’t matter IMO, because that is either stating the obvious or just part of your standard anti-GPL / anti-Linux rhetoric, rants which are of no particular interest unless you start innovating on that aspect or start to actually comprehend how the licenses work.
Have a nice day.
Whether the action of removing this hook is a good or a bad idea, it has caused bad publicity, and the most logical person for most people to blame is the USB developer, don’t you think?
Furthermore, I simply cannot believe that publicity has no influence on sales. And I also don’t believe that different markets (desktop/server) don’t have any influence on eachother. Likewise, I am not the only one who thinks that quite some people initially didn’t trust Windows CE, just because of the blue screens of Windows 98.
Sorry to break it to you but the only place that Linux is receiving bad publicity over this issue is here at OSNEWS. I can’t imagine how a closed source webcam driver losing its place in the kernel can possibely affect server sales for IBM. If anything this whole issue only proves that only drivers with free licenses should be allowed in the kernel.
It’s sad that politics always affects the end-user. How about someone makes the developer’s life hard for a change? How about if they keep changing the API and making silly rules about licensing of kernel modules, that people don’t write drivers for or release specs for them? Oh! That’s exactly what happened. I guess you reap what you sow.
Why hasn’t this discussion hasn’t worked its way around to real issue, namely the foot-dragging by an awful of hardware manufacturers to port their drivers to Linux. Afteral GPL vs. BSDL is already going on in another thread.
But despite Nvidia proving it can be done with destroying the company and despite the fact that Linux’s marketshare is relative to the Mac’s, still Linux users (and others) have to rely on third party hacks. The hardware people don’t seem to have a problem porting drivers to a platform with a whopping 3% marketshare. Maybe Steve Jobs is the difference. Whatever the case is, how long are these companies going to wait before they put out drivers? How much marketshare and/or mindshare does Linux need before they follow Nvidia’s example?
Don’t you get it? No one wants to port their drivers to such a non-stable interface. Drivers break betwen kernel version 2.6.7 and 2.6.7, which could have been released weeks, or even days apart from each other. Linux is a nightmare to support, and certainly minority market.
Also, since it’s not painfully obvious to you, the Mac does have a stable API. Like Windows, it (relatively) recently broke its API, OS9 -> OS X, but it does not do so frequently. How much market share? Probably 10-15% before it’s even close to being worth the intensive support required.
And anyway, if you read these forums, you see all the time that Linux supposedly has better support for hardware than Windows! What a joke.
It is not obvious, that Mac has stable API. In fact, it doesn’t. Jaguar to Panther transition broke not only drivers (10.2 – gcc 2.9x, 10.3 – gcc 3.x, Mac drivers are written in “embedded C++”), but userland applications too.
And remember, how binary driver API helped BeOS? (For those that don’t remember – not at all. Most drivers were ported _from linux_).
It is simply much easier to direct ones anger at Greg-that oh so evil Kernel-USB maintainer or at Nemesoft(can´t recall his name)-that oh so evil guy who threw a tissy fit. These are real people which one can easily point thier fingers at and scream “you’re guilty”. Pointing your finger at a massive corporation like Phillips is a lot more difficult and needless to say far less rewarding.
I actually find it funny seeing how big corporations are so scared of someone seeing their 50k decompression routine. I mean that company can manufacture what it wants-from the IC´s all the way to the finished product-something which no individual ever could-yet the 50k code apparently is the entirety of their competive advantage over all other anufacturers-give me a break. The more “programmable” hardware becomes the more propietary the code becomes-in essence the code is surrogate hardware. I understand the economics of ‘firmware’ and programmable IC´s but it is a joke that hardware manufacturers are to friggin lazy to provide a software api sufficiently abstracted from the hardware to enable drivers to be written which don´t reveal the propietary secrets of the hardware involved.
If the manufacturers would simply agree upon certain standards, publicly open api’s could be written which would allow binary drivers for hardware, which conform to the standard, to be supported on virtually all platform which implement the standard. CAPI is one example of this, OpenGL is another, albeit on another level. Instead we have 3000 different products which are only differentiated by propietary firmware which effectively means no chance of support under Linux or any any other alternative OS-including most of the time Mac OSX. Sure this method is more economical(from a deranged point of view)-its less work to do things this way-but it creates so many problems, not the least of which is the inability to support it on different OS´s. Unfortunately most low end consumer devices are so constructed and this doesnt seem likely to change anytime soon.
As insane as this sounds I actually think it is a good idea to try to implement those parts of the MS Windows driver subsystem under wine to allow use of these things-however a manufacturer writes these firmware things they still have to communicate back and forth over the interfaces which MS provides for manufacturers. Of course it would be far better if open standards were created -but you and I both know Microsoft wouldn’t implement them and if they did they would extend and subversively destroy it. But the manufacturers apparently like being pimped by Microsoft-at least more so than cooperating together to forge new open standards.
I myself was also upset to here about the loss of the Philips drivers. No if, and’s, or but’s about it-the loss of these drivers does hurt-several thousand web cams which were in use are no longer usable, at least for the time being, under Linux. But if the kernel maintainers allow for propietary hooks to be embedded in the kernel we(the Linux/FOSS community) will wake up one day not being able to define our own kernel-instead it will be dictated by propietary structures which are not accounted for.
I would love to see Phillips wake up and smell the roses and actually provide good drivers for their products under Linux. But I would rather see platform independent open standards being developed-which would simplify the work of all those involved-manufacturers, driver writers and kernel developers/maintainers-and for us “users” things would “just work”.
Everyone screams about the big bad monopoly which Microsoft is-but they fail to recognize that literally the entire consumer computer electronics market are “enablers” and parasites in the market ecosphere which Microsoft represents. Linux is slowly beginning to loosen the association of this market with Microsoft- slowly some are beginning to fathom that other OS’s also have a role to play in this market. But Linux can only do so much. As long as the hardware manufacturers insist on creating new devices which are exclusively tied to Microsoft-through failure to provide driver implementations designed for open public api’s- all other OS’s are artificially structurally crippled-and this my friends is the real power of Microsofts monopoly…..
I think we need to form a new group- Microsoft Anonymous, a place for recovering programmers, developers, and designers to meet and come to terms with the abusive misuse entailed in Microsoft dependency. And only those who manage to make it through the 10-step program would be eligible for hire in the market. The first step in the program is recognize that their is a higher power. The second step in the program is to recognize that neither Microsoft, nor that which it represents, is this higher power. The third step is to recognize that this higher power is not propietary. and so on and so forth…..;)
” I think we need to form a new group- Microsoft Anonymous, a place for recovering programmers, developers, and designers to meet and come to terms with the abusive misuse entailed in Microsoft dependency.”
Make sure you have a place in that self group for recovering Windows fanatics who spent years screaming that Linux wasn’t ready for the desktop because they lived in mortal terror of life without Bill.
Can someone knowledgeable please enlighten me, how providing info on how to write a driver (i.e. to get a picture from the cam, send these bytes and expect this and that number of bytes to come back; or to draw a polygon, write the coordinates to these ports and the texture id/address to this port..) violates the HW producer’s trade secret?
I expected, until now, that the trade secrets are in making a chip that draws polygons blazingly fast, not in the way of how to ask the chip to do that. It’s definitely clear that you somehow have to ask the chip to do it. No secret in that.
And as for the camera and the codec… I expect they could provide a decoder part without providing the encoder (which usually is the more complicated part, similar to asymetric crypto). If I am wrong, someone who KNOWS more about this, please explain.
Once upon a time, I had a motherboard with integrated sound. The soundcard had the ad1848 chip. It worked fine with Linux 2.2.x, but broke with 2.4.x. After setting on mixer channel, the card remained silent until the next cold boot. After figuring this out, I tried to eliminate this channel from the mixer part of the driver. As it did ont really help, I ended up hacking aumix and then lived happily ever after for some more time, until I bought myself a PCI sound card.
Try this with binary drivers and binary applications… From this point of view, Linux really has better HW support. If it does not work, I can at least try to fix it (though I’m no kernel hacker). No such chance with closed source drivers.
Yeah, yadda yadda, you had no such problems with sound in Windows. Neither did I, but I had problems with my Bluetooth USB dongle (had to find some older but working version of the driver, newest one were no good). That dongle worked on the first try in Linux 2.4.x and 2.6.x.
I think it would be much more appropriate to direct that particular criticism at the hardware company who isn’t willing to make their product work with anything but MS’s OS.
Why is this sort of criticism always aimed at the Linux/OpenSource/Free-Software targets — are they the ones who are refusing to “play nice” with customers or corporations?
Yes. Linux has become (in some senses) a “product” — but this is in very large part precisely because it has eschewed the proprietary/exploitative model of software developement. Unaware late-comers may now winge about the occasional hic-up that inconveniences them in the use of (GNU)Linux software, but this sort of problem is not unheard of in the proprietary software world either (how much of your hardware had to be put aside in the transition to Win2K, for example).
Crying that these teacup tempests are plain evidence that FLOSS software developers need to abandon the very principles that created and carried forward this amazing succes (despite the derisive, disbeliving sneers and predictions of failure from some more conventional quarters) is a sign that the critic is ignorant either of the history of the (GNU)Linux, the (GNU)BSDs, and the propriatary OSes or has failed to grasp the lesson implicit in this history.
In this spat, AFAICT, a developer chose (for what-ever reason) to disregard a basic — and explicit — design principle. This was both avoidable and fixable. No-one else in similar circumstances seemed to need to contravene this standard. It also led to problems that he wasn’t dealing with, and no one else could, for at least 3 years. Eventually the problems could no longer be ignored, and the developer concerned said “I’ve had enough”, quit, and left us poor non-techie “lusers” unable to continue using a piece of hardware, precisely because that hardware relies exclusively on proprietary code.
To me, the lesson here is that Linux developers should take care to adhere to those principles all the more now, when superficial pressures might tempt them to chase after some particular marketing/business “old paradigm” definition of success.
————————————————————
And about our web-cams:
We’ve dealt with this problem before. But now-a-days, when I inqire about “a real modem — not a win-modem”, it’s a convenient test of whether I’m talking to someone who has a clue, or whether I should deal with someone else. My printer, video card, etc not only work — I can download the linux drivers from the corporate website. I don’t know how long it’ll take, but I’m quite sure that this hassle is in fact temporary. The developer may re-consider — and do a proper module, another may reverse engineer the SW, Philips will get wise, or a competitor will see an opportunity to gain a following among the technically savy elite. At least I know that the kernel is being kept unencumbered. That might be a hassel now — but it makes me more confident about the future.
Ha, it got put back in after all.
That it was nothing but a storm in the glass of water. Driver will be back – as it is known that someone at the linux-usb-devel will pick up pieces from original author officialy, t.i. will take everything with instructions and good blessing – and everything in it’s right place.
About whole disscussion – I guess many people who cried fool over Linux developers didn’t get their situation – users don’t care about legality today, true, but Linux/Libre Software devels almost everyone do – because it can cost them lot of money (get sued) and could bring an end to their programming.
As long as there will be sanity in the legal and ethical checkings about source code and all that stuff in kernel project (and not only there), I guess Linux and Libre Software will succeed.
p.s. it is not that i don’t care about users – I’m support in the local company which is serious about linux desktop (w e sell packaged and supported Debian based distros) and of coarse it won’t help me if this driver would disappear from the Internet completely. Lucky, I wasn’t alone and so we will countinue to enjoy our bought webcams.
About Orange webcams – they are interesting too, I will check it out.
Your argumentation does not convince me, because of the following:
1. There was a technical reason for Microsoft to break driver compatibility between Windows 9x and NT. There was no technical reason to break the pwcx driver.
2. Breaking the pwcx driver is not a strong motivation to Philips, for two reasons.
2a. Firstly, Philips probably has a reason for making a binary driver. If they would make it themselves, it would likely be rejected too.
2b. Webcams are targeted at the desktop market, where Linux market share is very small. Why would Philips invest in a driver?
3. If you want to increase that desktop market share, hardware should work for desktop users. An average home user who runs Linspire or something like that, will probably blame Linux if his newly-bought webcam only works in Windows, and not the webcam itself.
Well well well. The problem is solved:
http://kerneltrap.org/node/view/3747
Multiple maintainers stand up. The hooks will be gone. Nemosoft seems to cooperate Both 3 sides (nemo, kernel devvers, customers) win. The solution was already in development behind the scenes the past days and it took them in total 5 days (24-29th) to solve the conflict.
Eugenia, when is Eureka coming? I’m waiting for an update or new post. A nuanced rectification would also be applicable IMO.