Slashdot Mirror
Kevin Carmony Responds to Criticism
sharkscott writes to tell us that LXer's Don Parris took a few minutes to get Kevin Carmony's response to the large amount of criticism he has been taking over offering non-free software in Linspire. From the article: "Essentially, Carmony's position is that, in ten years of holding out, the FOSS community has made relatively few gains, in terms of convincing vendors to release libre codecs and drivers. In other words, the strategy doesn't seem to be working. Additionally, while some will be patient, most users would prefer to have something - anything - that works in the meanwhile."
Note that Mr. Paris pointed out to me that Robertson stepped down as CEO. Carmony is running the show now. (Just in case you pay as little attention to Linspire as I do.)
My point still holds, though. There's nothing "wrong" with what Linspire is doing with the Freespire project. They're giving away free binaries (which they don't have to give you) along with all the source code they owe you. In exchange, you may or may not become a Click and Run customer. I don't see an issue here. And no, I don't think that Linspire is really expecting a huge outpouring of volunteer programmers, either.
On another topic (since I can't make fun of poor Mr. Robertson's Linspire work anymore), has anyone noticed the latest from AJAX Launch? It seems that they have added an Excel "Demo" (a pretty bit of XUL that looks like a real spreadsheet), a media player that seems no more sophisticated than the one in sharkscott's link in the summary (if I wanted your website to make noise... grrr...), and a RealPlayer video of the "AJAX Desktop" of the Future.
Are you amazed yet? Ecstatic? Hopping up and down in excitement? Holding your breath in bated anticipation?
No, neither am I.
Javascript + Nintendo DSi = DSiCade
If Linux proponents expect to see any sort of growth in desktop Linux usage, they are going to have to back down on this issue. Users want their MP3s to play. They want their videos to play. They don't want to deal with some complicated installation procedure just to get basic functionality that they can get easily, out-of-the-box in an install of another operating system.
Linspire realizes this, so they're doing all they can to make it easy as they can for new Linux users to use Linux and do what they want. People shouldn't be giving them flack for this.
Arrrrrrr
...that the _real world_ does not share their view that politics is the most important thing in software... Functionality is...
Idiot.
This is the first I've heard of this situation involving Linspire. All I can say is, "Sounds good!"
I can't believe how many times I have been stymied when configuring Linux because it didn't support my major-vendor video card. The "Open Source" version of certain drivers don't work. I tried an OSS implementation of some Nvidia drivers and it could barely spit out any video at all, much less allow me to use the advanced options on the card. I know the OSS developers tried hard, and I appreciate that. However, it just didn't work.
At times like these, I don't really care about politics or philosophy. I'm just trying to get the computer working, and if I get stuck because of OSS, I'll just abandon the project.
I suppose this is the reason why I haven't been a serious user of any Linux Desktop software for years. I use Linux as a server all the time, on dozens of different machines. It works great as a Server.
94% of Repubs and 21% of Dems voted to renew the Patriot Act
but it seems to be making him money, so who am I to argue?
IMHO, there has been little to no success getting vendors to release documentation to write drivers for three major reasons:
1) To date, the market share represented by open-source operating systems is very small.
2) The users and distributors of opensource operating systems have not presented a united front when it comes to the inclusion of propriety drivers and code. In fact, it seems the vast majority of distributors and users are more than willing to settle for closed, propriety drivers (even when they are crap!)
3) American corporate culture reflexively resists voluntarily releasing information of any kind. It is always easier to say no. Some Taiwanese vendors, for example, have been found by some opensource projects to be rather cooperative when it comes to releasing information. Major American corporations by constrast are a guaranteed stonewall.
We believe that the only way the world can successfully advance in the field of computer software is by eventually replacing all closed source systems with open source ones.
Take an example of Apple's recent success with Mac OS X. This software, although it contains tons of closed source code, is based on open source code and contains literally hundreds and hundreds of free software packages. Apple would never have succeeded in creating such a feature-rich operating system in the time it took to make it without the availability and use of such open source code.
This is why this Linspire debacle is happening. People know that although the expedient thing to do is to continue using closed source proprietary stuff, the correct thing to do is to get ourselves off that addiction and on to some better software.
Last week's Slashdot article on Theo de Raadt was about how he's not using binary drivers.
Ever try writing a 1 to the /proc/self/seccomp file? That blocks everything except read/write on already-open file descriptors, exit, and some stuff for returning from signal handlers. On x86, the cycle counter is disabled too.
The alternative is an extremely strict SE Linux policy, but seccomp is probably better for this job. One could use both at the same time I suppose.
I don't want some spyware crap telling Sony/Microsoft/Real/Sorensen about everything I do and probably acting as a backdoor.
Gundam Wing expresses it nicely:
From a historical point of view, warriors who have lost what they were protecting and were further betrayed by those they were protecting are losers. But they do not recognize themselves as such. Not only that, but they retain a strong will to continue fighting. The emotions of those who were thought to be beautiful are always full of sorrow, and honored tradition disappears in the cry of the weak. Winners of a battle will eventually decline in power and become losers, and then those 'losers' will cultivate a new leader.
Keep in mind that you can't rely on the nVidia drivers to work, because they are closed source. Opting for a closed source driver is accepting that Bad Things may happen to your system, and you may not get any help if they do.
Sir, I'm not sure, but I think you might be dumb.
/\SS!
Configuring OSS is not like hazing. Hazing is what "they" do to you so you can be accepted by "them".
You configure OSS so that it will do what you want it to do. And you learn something about how it works in the process.
We used to spend time compiling apps. Then package managers got better and adding apps is now easier and automated.
We used to spend time configuring X, now most distros can handle that for you.
The progression we are seeing is not one where folks are becoming less interested in OSS as it becomes easier to install/configure/use. In fact, it's the opposite.
The functionality bar keeps getting raised by the community, who continue to dream up new abilities and uses for OSS. They only have so much time, and when it's freed up from not having to spend a lot of it configuring a basic system, then they have more time to dedicate to developing that application they really are interested in.
And, as for the folks you refer to as 'losers', they will continue to configure linux to do what they want it do- only their expectations of what constitutes basic functionality will continue to grow.
Well, there are quite a few college students that have joined the cult. But they usually see the light when they hit the real world.
Two weeks ago, I posted a message after Kevin Carmody introduced Freespire at the Linux Desktop Summit. For those who didn't see it, the message seems relevant to this discussion, so I thought that I would ask forgiveness for repeating myself and post it again. Here's what I wrote: "Linspire is, of course, a purely commercial effort, with the goal of selling a shrink-wrapped OS that looks externally as much like Windows as possible. His target audience is not the Slashdot crowd, but rather the people who buy their computers at Wal-Mart. Really! For them, it's all about the out-of-the-box experience, starting up a computer with preinstalled OS and apps and just using it. As someone who has recently installed Mandriva, Fedora 5, and Ubuntu Breezy on various machines, I think that the experience is much better than it once was, but still falls short of the "Wal-Mart" or even the Windows experience. To listen to Kevin Carmody, Freespire is offered in the spirit of recognizing the contributions of the open source community, and giving people the opportunity to stay "pure", i.e., without licensed and proprietary pieces, or hybrid, where the user can choose to download and perhaps pay for the licensed and proprietary pieces. He gave an analogy with food, where the choices were Junk Food (Windows and proprietary software), Healthy Food, and Vegan. Open source vegans, of course, are those who would never want music in the proprietary MP3 format or images in the proprietary JPG format. His belief is that most consumers and business people would like Healthy Food, which is some mix of Linux and those proprietary formats, plus some drivers for graphics cards, etc.. He and his company are actually going out to Fortune 500 companies and talking to them about why they should consider a move to something like Freespire rather than suffering the pain and expense of migrating to Windows Vista (if and when it ever ships). This is a fairly brave, not to say crazy, thing to do, and I think that they deserve some credit and support for their evangelism, even from people who don't care for the whole Linspire business. Getting 3-4% penetration of Linux (any flavor) on corporate desktops would be quite an achievement, and it won't come from Linspire on its own. Carmody also said that they are going to open source Click N Run because they think that it is the best updating program, and are offering it to others for the taking. If I were responsible for Ubuntu or other Debian-based distros, I would be very tempted to take them up on their offer. I've done enough "apt-get"s."
most users would prefer to have something - anything - that works in the meanwhile.
Jesus fucking Christ! Is this something that has slipped past the great minds behind the Linux revolution!?!? Yet you fucktards STILL don't understand why Linux isn't catching on. Man, this kind of stupidity really burns my ass.
Their windows drivers are closed source, but I expect them to work. I also /do/ expect the nvidia linux drivers to work, because... they do!
Sure, nvidia "might not" keep the drivers up to date with all the linux kernel side changes that are going on, but they "might not" with windows as well... but they do, so their customers can use their product.
Okay, this weeks drivers might not work with next weeks kernel, but this is a problem with the linux kernel not having the same backwards compatibility as windows. Can hardly blaim nvidia for that.
It's so not as big a deal as everyone keeps making out.
The revolution will not be televised... but it will have a page on Wikipedia
The distro I use, Gentoo, lets you play mp3s easily. In fact most Linux distributions do. I don't think it's a controversial issue that people want interoperability with their closed format files.
But that's not the issue people have with Linspire.
I'll probably be modded down for this...
The fact that the kernel isn't obligated to accrete cruft continually to preserve backward compatibility is an advantage, not a problem.
Nvidias unwillingness to simply document their product so it can be properly supported is not something you can blame on anyone but them. Old-think dies hard, but it does die.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Friends don't let friends enable ecmascript.
The only difference between open and closed in this case is that in open source, you can't blantly do something the user doesn't want.
I know that I like the concept of open source, the same way I like merge-sort, it's solid in theory. I say this because I know I haven't read the source of any code I use (except for stuff I write). So the benefit of openess is just a different web of trust to me anyway. But realistically, there are trade-offs in either direction.
I have 0 guarentees of help, or correctness using open-source, even in closed source I most often can only find a shitty out-of-date web site that is trying to be idiot-friendly and helpful (the two don't mix all the time, especially in this case) but utterly failing. Yes, there are communities, but they are far from perfect. And in either open or closed source, we've all run into a problem that has been noted and will be fixed at the earliest convience (or interpreted as a feature and you're SOL, and yes, in the open world you CAN change whatever you want, but you need things like time and knowledge which many people do not have in abundance).
You run a computer, it will get FUBAR'd. Try your best to minimize those experiences (or at least learn from them), and stay as productive as possible. If you don't like the heat, get a liquid cooling system.
Although I appreciate a clean, cruft-less kernel, wouldn't lack of backwards compatibility doom all dfriver projects to eternal development?
Great for hobbyists, but lethal for anyone wanting to write drivers. Today's working driver would be tomorrow's kernel panic factory.
What happens to less popular drivers? Do they just die for lack of skilled people to work on them when the kernel changes?
I'm guessing that this response was motivated (at least in part) by the effort of PJ at Groklaw. For those that have visited recently, PJ did a scathing article on Linspire/Freespire. Really harsh stuff which a lot of people found pretty unwarranted. Myself included.
... deleted. Then it descended into farce.
So I posted anonymously as I usually do. The odd thing that happened to me was that I found my post deleted. So I posted again
She seemed convinced that this was an orchestrated attack by Linspire "astroturfers". And when Kevin posted to the forum, she wouldn't talk to him and asked him for an apology from the (imaginary, IMHO) astroturfers. Having said that, Kevin did quote an email he sent PJ which I thought was poor form.
Anyway, I literally sat there for ages watching post after post being deleted which I thought was amazing. A large number of these posts were quite sensible. They just didn't tow the Groklaw line.
When it had calmed down a couple of days later, I posted that here is a place where they discuss free speech, but don't practice it. Quite frankly, the amount of groupthink and censorship I saw left me with a very different opinion of the place.
The best thing about Slashdot's comment system is that it keeps all the posts. Even the trolls.
Who else but the designers of the hardware to produce drivers (open or otherwise)? They have access to hardware schematics, development plans, and the engineers who designed everything from the fabrication plants to the chips you're writing the drivers for. Do you honestly think you're a good enough programmer to fix a driver for hardware you have no knowledge of? I'm not a programmer hardly at all. It doesn't matter to me if it's open or closed. Either way, I can't fix anything.
Assuming that something won't work because it's closed source is as stupid as the closed source camp claiming FOSS is more susceptible to security vulnerabilities. It's absolute BS. And won't get any help from the vendor? I'd say I'm as likely to not get help from a vendor as it is likely that the FOSS community will label my bug Won't-Fix. God forbid I happen to get some rare bit of critical hardware for which the FOSS "community" consists of one guy who's a complete idiot.
Yes, I understand the FOSS model. Yes, I beleive it is superior. Yes, I believe it is the future. But avoiding closed software because of some nebulous bugaboo makes you seem like Chicken Little in a snowstorm.
The road to tyranny has always been paved with claims of necessity.
They become buggy pieces of shit and Andrew Morton has to come out and call for a "bug fix release" because that spaghetti monster that is the linux kernel done borked up. If you look at open source projects in general (GCC, glibc, linux, X) every release relegates some platform/hardware to the scrapheap of unsupported because they've reimplemented their cruftball again and no one wants to rewrite support for everything again. This is known as progress.
And everyone knows that the Linux kernel developers just use MacOS when they want something more sophisticated than emacs on a 80 column tty.
Drivers for fully supported hardware are kept right in there in the kernel tree on kernel.org.
Whenever the internal kernel interfaces change (which is really pretty rarely) kernel programmers also check all those drivers and make any changes necessary.
Once a device is supported, it's very nearly perpetual. It's rare for drivers to be removed, and usually when they are it's because they've been superceded (the hardware still works, the support is just being done more elegantly, for instance when 2.6.16 was released amdtp and cmp had been removed, but that was because the hardware they supported is now supported with libiec61883.)
For a device that was once supported to actually be dropped, there has to be a major kernel change combined with no one in the kernel development community (paid or volunteer) having the motivation to update it. This means, no paying customers of RedHat, SuSE, etc. using it, no kernel hackers have one in a still functional system at home, etc. And even then, you're free to grab the source for the last working version and update it yourself, or pay someone else to do it.
Try to do that with a binary driver.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Friends don't let friends enable ecmascript.
Lack of backwards compatibility would "doom all driver projects to eternal development", except that kernel drivers are updated by whoever breaks backwards compatibility. This is why drivers need to be in the official kernel source tree. Driver authors often sit back and relax while other people make the required changes to all drivers in the tree. Updating the drivers is often a robotic task, often taken on by the "kernel janitors" team.
This usually keeps the less-popular drivers alive for many years, though not forever of course. Linux just recently lost support for the PC-XT hard drives that came in 5 MB, 10 MB, and 20 MB sizes back in the early 1980's. (these never shipped with a 386, but people sometimes put the old drives in newer machines) It is unlikely that any of these drives still work.
Yes.
OK, not 100% of the time, but very often. Perhaps the problem is unrelated to the hardware itself. The code might forget to check if a memory allocation failed. The code might take locks in the wrong order. The code might try to access user memory while a spinlock is held. I can certainly fix these things.
With a bit of hardware documentation, I can do much more.
With plenty of hardware documentation, I can write the driver myself.
If I happen to have the same hardware as you, then you get to use my code.
In fact, lots of people have problems with closed drivers under Linux, much more problems than with the free ones. So there is this little thing called experience backing some of the arguments being made.
That may be well correlated to the way you go about asking for your bugs to be fixed... In any case, I really find this staement of yours to be not true.
Well, in that case, that idiot would be you ;-) Seriously, unless that rare bit of critical hardware is a very expensive thingie and/or you are paying those that make it to write/maintain drivers, then you are as screwed with a propietary driver as with a free-one-wriiten-by-a-complete-idiot.
Who can authorize the release of anything? Well, a CEO can, but he's kind of busy. An engineer certainly can't.
Get a market share above 10% and vendors with consider to release on Linux. Get a market share above 20% and vendors will release on Linux. Get a market share above 30% and vendors can't afford not to release on Linux!
5 .pdf). How to fix this inhibitor? One important action (IMHO the most important) is to declare the guidelines of wyoGuide (http://wyoguide.sf.net/) as the Linux application standard!
How to get a higher market share? Fix the first top inhibitor of the Linux adoption (http://www.osdl.org/dtl/DTL_Survey_Report_Nov200
There might be other actions to fix this inhibitor but I don't know any. Just voice them here and now. But be sure unless this inhibitor isn't fix fast, the market share of Linux will stay low, too low for any significance.
IMHO it's essential that anybody (maybe O'Reilly) starts a Linux conference about this subject to discuss any possibility.
O. Wyss
See http://wyoguide.sf.net/papers/Cross-platform.html
Dell's BIOS was doing some nasty hack to work around an Intel chipset bug. The bios would just grab the CPU in System Management Mode, totally messing up real-time.
Since we didn't have source, we had to solder tiny wires to the CPU just to find out what was going on. We then determined the problem using a digital logic analyser. (looks kind of like a scope)
I find your argument to apply equally well to the open source drivers, as well as the closed source ones.
While in theory I could fix a bug myself in the open source drivers, that likely isn't the case. I just don't have the technical know-how to diagnose such a problem, let alone figure out how to fix it. I could try to find somebody who could fix it, but that will take time and money.
What it comes down to, open or closed source, is that if the driver runs into a problem, I may not be able to fix it. For a non-technical end-user, it's basically the same situation for both types of drivers. The limiting factor is that I do not have the knowledge to fix a broken driver, which renders the fact that it's open or closed source irrelevant.
The LXer article states "...to educate newcomers with respect to the importance of keeping software free" but people, especially customers, don't like to be educated of something they don't care. What would you think if the car seller wants to educate you that 60mph are more than sufficiant for any driving? Would you buy a car there or leave for another shop?
h tml
O. Wyss
PS. See http://lxer.com/module/newswire/view/54009/index.
See http://wyoguide.sf.net/papers/Cross-platform.html
When did jpegs become proprietary? I thought this was the whole reason the FSF was using jpegs when everyone else was using GIFs on their website.
By two tracks I mean one track that has a Linspirish philosophy of just make it work for newbies even if that also means including some closed source proprietory software, AND a purist RMS FSF.
Really these two tracks compliment each other the closed source development track brings in newbewies while the purest camp can defend our freedoms and perhaps save our butts if DRM becomes very prevalent.
The point is though why does each side have to try to convert the other to it's philosophy as my way or the highway? Lets let them both run and see what happens, after all that's what's going to happen anyway, it's very unlikely either FSF or closed source software is going away any time soon.
I use OS X which has closed source software in the OS and I run closed source apps as well, but I also run fink/KDE as well on top of the open source Darwin base at times, as well as running Firefox as my browser. Does that mean one "side" or the other should work to convert me? What nonsense, what a waste of developer time, and above all how immature.
Can't you argue about something of real importance like poverty, or war, or whether peak oil is real, etc?
Tired of all the isms, don't exploit people as an employer, or a government, mmmmK?
This is the problem with FOSS, lack of direction. People need to realize that x branch of y software should have some interoperability and backwards compatability. Various immaturities such as "OMG HOW DARE YOU IMPLEMENT NON-FOSS INTO LINUX LOL!!!!111" and "BINARY DRIVERS ARE TEH SUXOR" will just get more major (and minor) players detracted from linux. If you need to rely on a system, you want to see a solid, unified front, not a disorganized group of hippies hellbent on an idealogy. Compromise is the best and only solution in today's world. That's why Linus will be remembered and Stallman will pass into obscurity.
Do not downmod posts "overrated" simply because you disagree with them.
About a year ago I heard an interview with Kevin Carmony, who at that time has just become CEO for Linspire. Up until that interview I had a pretty negative impression of Linspire (Lindows), but I had never actually tried it. I must admit, however, that I was really impressed with the LUG radio interview. The new CEO explained and addressed most of the misconceptions I had about Linspire up to that time. Linspire has done much for getting Linux pre-installed on desktop computers. They have been around for five years now, and have contributed a lot to open source (http://linspire.com/opensource). I think much of the misunderstanding stemmed from Linspire's target audience being non-technical consumers, which the slasher crowded couldn't appreciate. LUG Radio interviewed Kevin again this week, a year later, and he talked about Freespire. Once again I was impressed by his candor. I run Debian on my computers, but I'm grateful for the past year of education that I've had in coming to understand Linspire and their goals. I even plead guilty to recommending it to my non-technical friends when they say they'd like to try Linux (not sure they're ready for Debian without the ease-of-use coat of paint Linspire and others add).
I'm planning on taking a look at Freespire. I am intrigued.
R. Singfield
Source code would be nice, but it's not what we want most.
We want hardware documentation. We can write our own software. Our software will be more stable, portable, and maintainable. Performance could be a win or a lose.
With hardware documentation, we can turn a WinModem into a telephony interface for a PBX. We can support Linux, OpenBSD, GNU HURD, and eCos. We can port the X server to run on the GPU. Lots of neat ideas become possible.
Do you honestly think you're a good enough programmer to fix a driver for hardware you have no knowledge of?
:)
Actually, yes. Sure, I won't be able to fix the most complex problems as I'm not a device driver developer, but having the source code gives me options. These range from being able to identify bugs, then after determining their root cause work around them, through to hiring a capable programmer with hardware experience who can do the job.
As a real example, I once found a bug in the read-only firmware of a network card. Fortunately, since the driver for this was open-source and I didn't need the feature that was faulty in the NIC, I disabled it along with the corresponding diagnostics the driver initialisation procedure invoked, then recompiled the driver. Presto, a working network card.
I'm not a programmer hardly at all. It doesn't matter to me if it's open or closed. Either way, I can't fix anything.
I'd like to consider this: just because you personally cannot use the source code does not mean that you can't benefit from it's availability. Even if only indirectly by virtue of the fact that real programmers can do things with it you will like.
Who else but the designers of the hardware to produce drivers (open or otherwise)? They have access to hardware schematics, development plans, and the engineers who designed everything from the fabrication plants to the chips you're writing the drivers for.
. . . and budgets they need to stay within, time constraints due to conflicting priority and limited resources, and a profit incentive to sell you new products that might cause their interests to be at odds with those of their customers. Never mind the fact that they may not be around or in the business of providing the products/services they do currently when you have an issue that only someone with the source code can deal with readily.
When your hardware goes end-of-life, unless you're a huge customer of theirs they have little or no incentive to fix that obscure bug that is now causing you major grief. If the device driver was open source, you could pay a thrid party to fix it for you.
People need to realise that Free Software is about giving you control over the computer systems you rely on by removing artificial constraints (in this case the absence of source code), not (necessarily) about making software cost you nothing. It's about making computing predictably controllable and hence sustainable.
I always play my MP3s with a legitimate, patent-licensed player. After I download them from eDonkey.
Where is your video driver with full support for hardware accelerated 3D rendering?
It's installed on my computer, duh. This is possible because I have documentation for it.
Not that I give a shit about 3D, but it came with my computer and documentation was available so I have a driver.
Funny you say I can't rely on the nvidia drivers, they work perfectly, unlike the open source drivers which are a complete joke.
Insightful? Give me a break.
Kevin and Michael are both incredibly nice guys who have a particular angle of insight that no other distro has. That is, that users want something that just works in ways that leave ordinary people knowing what to do, or better, not needing to do anything except the task that interests them. Linspire gets it that most people don't want to do things more complicated than click and run. It takes an enormous effort to make software be just click and it works. That deserves our respect.
All of us are contributing, each in the ways we most understand. This sniping at each other, it is simply harmful.
I think I am going to go install Linspire. Let's face it, I don't have the time to hassle with making mp3s and dvd players and voip work on the big distros either, and I am a Linux developer, I can't imagine what ordinary users do when they want to use Linux on one of these distros that requires you to get libraries that don't just compile and work and somehow install them before your dvds can play. Or have they finally gotten it together recently, someone tell me....
If it is not written by me, it should just click and run.;-) Or at least, make and run.
Oh, and pissing on nvidia is not reasonable. At least they port to Linux, ATI just ignores us.
Charity is something to be thankful for, not to demand. Free software is charity. I like to do it myself, but that gives me no right to demand it of other more sensible persons.
The closest to a clue you have about writing drivers for such hardware is changing your boot parameters to change your console resolution. Faggot poseur.
A couple of years ago, the 'Linux Incompatibility List' was created to track stuff that doesn't work with Linux:
http://www.leenooks.com/
It may not be much, but it has the advantage that it points out what to avoid, and it's community maintained - with all the hardware out there these days, no one person can know about it all.
http://www.welton.it/davidw/
People already must install numerous pieces of proprietary software on their linux systems. Who uses *desktop* linux without any proprietary drivers or software? Even ignoring drivers, what about Java? None of the Java clones are nearly as good as Sun Java... yet linux distros fail to include Sun Java, forcing nearly everyone using java for any serious purpose to replace it immediately at some unnecessary inconvenience.
By taking the hardline "only OSS" stance at the distro level, we're just pushing installing the non OSS software onto the users. It's just an annoyance that accomplishes nothing.
As far as Linux being locked into unchangeable kernel schemes... maintaining binary compatibility for drivers is something they should be doing anyway. It is something that every other kernel I know of does, and it is just plain annoying that I can't swap out the drivers from one linux install to another because of driver breaks between kernel versions. At the very least, driver compatibility should be guaranteed between minor version numbers.
I would never buy a car with the hood welded shut. I'm no mechanic, but I can hire one.
That's pretty good.
I was thinking of that case when a small user (personal or business) might depend on a device that's obseleted, but they've got neither the money nor the skill to fix it themselves.
I guess it's unlikely. Possible, but in the minor list of worries.
In five years of holding out, the Lindows company has made relatively few gains, in terms of convincing customers to buy their repackaging. In other words, the strategy doesn't seem to be working.
"Nvidias unwillingness to simply document their product"
Simply? Are we talking about the same product line here? Leading edge highly competitive constantly pushing the boundaries through advanced hardware and software optimisation graphics renderers? And you want them to "simply" document them... all...?
Sorry but I'd rather leave it to the people who are employed to be skilled in building them. Nvidia/ATI have pretty much left everyone else behind with their graphics processors, if no one's come close to designing better hardware, why the hell would you wanna put a super fast graphics card in the hands of hobbiest driver writers? Could they really do a better job?
The revolution will not be televised... but it will have a page on Wikipedia
The difference in thought here is that I view it as nothing more than any other tool. I want it to work, I don't care how it works. If I have to take notice of it or spend more time to make it work then its not working.
As such, I won't bother with Linux on my desktop because I don't care to have to think to use the damn PC. Its a tool, as such the OS should be as transparent as possible. If I have to do more than slap a CD in to get my latest addition to my PC to work then its not transparent enough. Just as my friend uses OS/X over Linux? Why, because it just works, no thought needed. The OS is not germane to his work and as such it wasn't worth his time to install Linux on a machine. Whats the point?
Convert hell, convert the attitude that only "real computer" users are those who use linux. The whole attitude you espouse smacks of elitism.
Screw that. Its just an OS.
* Winners compare their achievements to their goals, losers compare theirs to that of others.
"So I posted anonymously as I usually do"
Why not stop moaning about it and create an account (free) and post from that? PJ's had SCO astroturfers hitting her sight and has had 'friends' of SCO posting her personal details to the internet at the same time she was getting death threats.
So she might be a little oversensitive. Get over it.
---
We spoke for about a half an hour. I don't recall a thing we said. - Colorblind James Experience
Nobody is expecting you to fix the driver.
When a change is made to the Linux kernel that may break some drivers, kernel developers check the driver code and update it to conform to the new kernel. Of course they can only do this with open source drivers. When you upgrade your kernel, the accompanying drivers are upgraded at the same time. Your closed source nVidia driver, however, may stop working.
This sucks for users, and it's even worse for nVidia. They have to keep close tabs on kernel development and provide drivers for whatever kernel versions they need to support. At the same time, those freeloading companies that opened their driver source are getting their drivers updated and distributed for free. Bastards.
So, the next time you want to claim to understand the FOSS model, you might want to stop and consider whether you actually are.
We had a few developers build a mission critical system based on Kylix to run
on our desktop clients. Now here we are a few years later and borland decides to
end of life the development environment. The runtimes are dependant on the kernel versions. Now guess what I have, a program that I have to compile and or change on a old version of the operating system and I cannot upgrade production to anything newer I am forever stuck in this situation until a large amount of money is spent to replace the application....That gents is exactly why Kevin is just plain wrong I don't need and I sure don't want your proprietary software.
Got Code?
Your PC is a multipurpose tool.
Mine are not. Every one is specialized.
Each does a list of jobs and does them well and every one either saves me money or makes me money.
So on so fourth. Every PC I have is a tool in its own right and every one does its job perfectly well. I possibly can replace them with a single Winhoze box. It may do everything they do, but it will do every single thing worse than they do it now. In addition to that it will cost me more over its lifetime in essential expenses, maintenance and power consumption.
Such is life. Winhoze and MacOS X are multipurpose. Linux is multirole. While Linux cannot be made to do all of the tasks Winhoze does simultaneously, it can be optimised for a specific role to do it much better than Winhoze. Personally, I would rather stick with several tools each of which does its job well instead of a single that does all job at a mediocre level. Other people may make a different choice
Baker's Law: Misery no longer loves company. Nowadays it insists on it
http://www.sigsegv.cx/
... except you picked the perfect example of linux drivers that always WORK FLAWLESSLY. You run the file, and if it can't pull a binary from nVidia it compiles a driver right there and then for you. I've installed it on many, many machines with more than 10 differnt nVidia cards and have never had an issue. nVidia has a vested interest in keeping those drivers working well because, quite frankly, ATI's been sucking (Only supports radeon 8500 and later, DIFFERENT CODEBASE THAN THE WINDOWS DRIVERS, mising features, etc.)
nVidia is the PERFECT example of closed source drivers done right.
One basic problem is that this is part of a long-term strategy to proprietarize GNU/Linux. You start by including a few free-as-in-beer drivers and codecs. You make sure that some of those codes only Linspire can distribute. Now, no one else can compete on those codecs. You add a few desktop applications that are again, a free download from Linspire, but no one else can contribute. A few years down the line, you hope people standardize on Linspire. Then, you offer "Linspire/Pro", which costs money, but includes slightly nicer versions of the apps. A few years down the line "Freespire" is a crippleware relic, whereas "Linspire/Pro" is nice, and we're back to Windows. Only initially developed by free labor.
A second basic problem is that it makes binary-only drivers work in the short-term, so there's less incentive to release specs or free drivers. Binary-only drivers are fragile. The Linux APIs and ABIs constantly change, so your hardware works only so long as the manufacturer keeps supporting it (and there, only on supported kernels). If 5 years down the line, your 802.11 card has a driver for 2.8.19, but your 3d card for 2.8.16, you can't use both at the same time. It's insanely fragile. Free drivers work forever.
It's a slippery slope. Let's not start down it.
So nVidia has a / a few kernel monkeys that keep tabs on the kernel and update the codebase. Big deal. If they want to keep their code close to their chest it is their perogative. One of the nice things (from a developer standpoint, I do 3D graphics work) is that their Linux drivers are the same as their Windows drivers, minus the kernel hooks. The same can't be said for ATI. That's probably one reason why they don't want to release drivers.
Have you ever run the nVidia installer? It is very handy. It will go out to the internet, download the appropriate kernel module and install it. If it doesn't exist it will compile one there, on your computer, against your current kernel. So unless they are breaking the kernel every night you really don't have to worry about breaking your drivers...
What? Since when does the "Open Source camp" prefer non-copyleft licenses? What kind of drivel is this?
And why is no one screaming and pulling their hair over the fact that Dell ships their RH Enterprise-equipped machines with closed-source nVidia drivers?
I teach my Open Source Technology students that OS is a continuum, and that everyone falls somewhere along that continuum. ESR embraces the business side of OSS, while RMS (firmly!) embraces the libre side...everyone involved in OSS has some philosophical bent. If PJ has a problem with Linspire, she has every right to rant about it. But since she doesn't speak for the OSS movement, we have every right to ignore her (or pick up the pieces we agree with and discard the rest).
The beauty of OSS is that there's room for everyone. Don't like what Linspire is doing? No worries, come up with your own distro that ships with OSS versions of whatever it is about Linspire that rubs you the wrong way.
I have installed the nVidia driver before. It sticks out very clearly in my mind because it's one of the few occaisions when I've had to install a third party driver rather than use a driver in the mainline kernel tree. I'd say that nVidia represents the best case scenario when using closed drivers. If I really depended on the capabilities of a good 3d graphics card, today I would use the nVidia. It's the best of a set of bad choices.
I'd rather have my drivers in the main kernel tree, so that my drivers get updated every time I update my kernel. I update my kernel pretty regularly. If I depended on an nVidia driver, I would have to take the extra step of updating my driver. Occaisionally, that update might fail if nVidia hadn't kept up, so I would have to stick with the old kernel for a while. Since I don't need a high end video card, I choose to forgo that hassle.
In summary, using closed source drivers is doing it the hard way. I don't do things the hard way unless it's necessary. I hope we get a fully capable FOSS driver for nVidia cards, because that's the only thing keeping me from buying one.
Not all software is open source. Its a fact of life. I love open source software. I love free software. but I do not think there is anything wrong with putting non-free software in your linux package. This kind of attitude is holding back linux. So what is 1 distro out of hundreds becomes "tainted" with copyrighted software.
Guess what? I play Neverwinter Nights and myth2, on my linux/bsd boxes! So shoot me because the source is not open. Its open source fundamentalism, at its worst. If you want to put out a good product, you have to come to the realization that not all drivers or software is free, and unless you want to write your own driver for every single bit of 19.99$ hardware out there, or 29.99$ game or productivity tool, its a lot easier to package binaries.
I hope we get a fully capable FOSS driver for nVidia cards, because that's the only thing keeping me from buying one.
And if we did, I'd still use the closed source one. As a developer, the code base they use for the linux driver is the same as the Windows driver. When I write an application, I know that the program will look the same (3d graphics wise) under Windows as it will under Linux, since the low level graphics and GL implementation are the same. I can't guarantee the same thing with an open source driver. I don't know that they won't fudge the GL implementation.
That's a big problem with devleopment because the 2 big players are ATI and nVidia. And things can look very different between the two because ATI linux drivers are a seperate codebase from their Windows drivers. So you have to write code that checks card and OS, and change stuff. Very annoying if you are supporting multiple users/mulitple cards. Fortunately I'm just supporting a few users on nVidia cards...
The point of an operating system is to enable the user to get their stuff done.
You mean... Like solataire?
Look at porn?
Play some games?
Write grandma an email?
Many developers have high hopes for the end user's goal into create Nobel Prize winning works on their PCs.
Really, the end user just wants the operating system to work without fiddling with it.
But the same applies to the what you said about getting stuff done. But let us not kid ourselves about what the end user is actually doing.
"I am the king of the Romans, and am superior to rules of grammar!"
-Sigismund, Holy Roman Emperor (1368-1437)
Isn't it interesting that there is an automatic assumption that if there was a good FOSS driver for nVidia, it would not be derived from the current closed driver? What's in there?
The real question is, (using a car analogy) if you needed to get it fixed, do you always want to go to the dealer, or do you want to be able to go to another mechanic?
Today most customers have never seen, nor heard of, a local software repair shop, so to them there is not a difference -- the distributor of the software is the only place to go to get it fixed if there is something wrong with it.
Most people will never actually see the value of free software until there is a local software fix-it shop, because they aren't qualified to fix it themselves.
- "History shows again and again how nature points out the folly of men" -- Blue Oyster Cult, 'Godzilla'
For some other insights from Kevin check out this weeks LUGRadio interview from the episode "Pigeon! Like your English!" availible at http://lugradio.org/episodes/50.
First, don't post anonymously if you want a thoughtful critique to be taken seriously. Anonymous posting of anything in favor of a commercial product just screams "astroturfer", here and on Groklaw.
Second, if your non-anonymous comments get deleted, do what I do: repost them on your own web site, with explanatory commentary on the censorship. And then don't waste any more time using the site in question.
Freedom of speech is basically reserved for those who own their own publications; the sooner people realize that, the better.
As you seem to realize, you're never going to force a site like FARK or (you allege) Groklaw to stop engaging in invisible arbitrary censorship, so it's not worth trying. To continue to use the site is to bolster their credibility. So quit, and write your comments elsewhere.
GCHQ Quantum Insert installed. If only our tongues were made of glass, how much more careful we would be when we speak
Had you continued to read the article you would have read that the chipsets were STILL SUPPORTED, just not under the unified driver. The Riva TNT is what, 7 years old? I would hope that code would have worked its way out of the unified driver.
I have a difficult time believing that someone without the hardware schematics, who has never worked on the card, who didn't design the card could build a better driver, even given the present driver. Running "valgrind" and "gprof" on a driver doesn't make you a driver hacker...
You probably missed the "real-time" part of the grandparent post.
You likewise missed the "software we use for development". This would include the operating system that the code-editing and compilation machines run, especially as opposed to the "libraries that we license for use in our products" that runs in a real-time environment. This software used on development machines includes BIOS and video card drivers; is their source code available?
Impressive that you have built a solution that works so well for you, and I like seeing happy Linux users. That out of the way, your very argument is what makes me so happy I am so far away from Open Source elitist bullshit these days.
A few years ago, I switched to OSS full time - I would not go near proprietary solutions for work or for play based on the premise that I believed OSS was the end all be all of everything. I loved the philosophy, I like most of the software, and I didn't mind when the software was incapable of doing what I needed to do - I enjoyed the challenge of 'making it work'.
These days I use a Mac for my work. Why? Because at some point, I need to be able to do my work. My computer is not a toy, it is not for games, it is for work.
By trade, I am a graphic designer/web coder. Linux/FOSS does not have the capabilities to do my work. Period. I am reliant on closed source software (Photoshop, Illustrator, Flash, etc) to get things done.
When you argue that anyone who is serious about their work will convert, I think you are correct - you just did not mention what we will convert to.
Do I like Linux and OSS as a whole? Yes. I use a lot of OSS software on a daily basis (vim, openoffice, firefox, etc). But a computer is there to do what you want it to do, when you want to do it. I hope in the future I can rely on Open Source to build the tools I need, but just because I can't at this time, that does not make my work or my needs a triviality as the elitists would suggest.
Your needs are not my needs, and apparently my needs aren't what open source developers on the whole care about. If the community would stop being so rabid and learn to bend a little bit, I think we would all be far better off.
Do you honestly think you're a good enough programmer to fix a driver for hardware you have no knowledge of?
Absolutely. How can I be sure? I've done it, several times. In practice, you don't typically need to know much about the hardware, nor do you really need to be a great programmer. The problems are usually pretty simple to spot and fix. Not always, of course. There are some that you need detailed hardware knowledge to understand and fix, but those are rare.
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
But i still feel like what we unionized people with plenty o' hair on the chest and lower regions call a "dirty little scab". I dont like the feeling, hope i one day will get rid of it...but darn sometimes you just have to listen to Edit piaf - je ne regrette rie.mp3
I took a couple of broadsides over this concerning the availability of apps on given operating systems. The gist of it all was if the Wintel world is so convoluted, why don't you just go use FOSS to do what you want. My original point was a comparison of the state of Apple and Wintel experiences, but the bottom line is it's the end result that counts. Can you be who you need to be and get the work you need done or not. It's about balancing htree dimensions of time, money and headaches.
Case in point. Astronomy software. I want to be able to head outside and decode the sky and use a scope. For a while I paid StarryNight because it was impressive and did everything. Once they jacked the price had to consider if I needed everything or if I could live with something less frilly. For a few dollars I can use shareware Equinox. It will always do what it does, it's way leaner than StarryNight and can control my scope. And for free, I can use Stellarium, which is OSS, beautiful, responsive, no scope, but everyone I know can use the same app across platforms.
However, if I had to involve everyone in builds, or an install so huge it should only be shipped on CDs, or requiring X11 or other layer(s) of techie stuff, it'd be back to Equinox and pay for the upgrade.
Blasphemy, I know, but when I want to go look at the stars I don't want to be the IT guy in order to get it done.
You can multiply that by a bunch if you subsitute millions of WalMart customers for me and everything plain old people need to do for "astronomy".
"Win treats sysadmins better than users. Mac treats users better than sysadmins. Linux treats everyone like sysadmins."
We are "in violent agreement". I understand your motives for going with MacOS. For me it is useless. My trade is network stuff and more specifically design, VOIP and QoS. I am not going to get anywhere close to getting any work done on neither Winhoze nor MacOS. Both suck to that effect. So I do not use them.
Every tool for its job. Some are better than the others, some are worse, but as a general rule a specialized tool will beat a general purpose one.
Baker's Law: Misery no longer loves company. Nowadays it insists on it
http://www.sigsegv.cx/
It is funny when overzealous FOSS people fail to recognize how much more complicated the drivers are for a GPU than for an Ethernet card. It's as if they don't even realize that a huge aspect of why NVIDIA was able to crush 3Dfx was that their drivers, DX, and OpenGL impelementations were vastly superior. The reason they competed so easily with ATI even when ATI was able to produce more powerful hardware was due to the driver and API implementation quality. If ATI could wave a magic wand they would make their driver quality better, but it's a complex task that even with millions of dollars and entire teams of professional developers with access not only to technical documentation but also the hardware designers themselves, they can't just will out of nowhere. Even with some of the technical specifications, the open source ATI drivers are really shitty when it comes to 3D. The idea that if NVIDIA were to release its tech specs and all of these people with SCSI controller programming experience are going to save the day is totally absurd.
Given the billions they're putting into Linux development, they might as well spend a few bucks on letters to manufacturers that'll make it a lot easier for them to package Linux solutions for people.
They are in a position to blow some doors open for us.
Tech Public Policy stuff
NO, Not by the way you are defining 'users'.
Unlike commercial software or commercial anything, Linux and other volunteer type products don't need the same type of customers.
True customers for OSS are people who will tinker with it, and develop for it. Its success is dependent only on these type of people existing (assuming the IT sector was a level playing field).
All other types of users like the desktop and server users are really irrelavent. If they come along for the ride, GREAT (IMO, some disagree), but if they don't that isn't as great, but it is OK... not really a loss for OSS.
Obviously commercial software is different, they need paying customers, who provide the funding to pay to develop software; whose focus is on what the customers want which could be something you tricked them into thinking they want and not necessarly 'better' software defined from a pure CS point of view because few if any customers actually CARE about 'better' other than defined by theyselves.
I run my Linux box with full multimedia, it's the machine I use to do business with. . . as somebody who currently writes Linux tutorials for a living.
I would simply not be in business if I couldn't run Linux with proprietary commercial apps like Win4Lin (which lets me run the nonFOSS Windows OS in emulation) and several other proprietary Linux apps... including the Turboprint package of print drivers that supports the Canon printer print-to-CD capability. Or at any rate, I would be in business as a Windows or Mac user, because the base Linux distro installs are NOT READY for SOHO business use.
The fact that the people who have Linux multimedia are either using Linspire style training wheels distro or are part of a l33t minority is all that one really needs to know about Linux usability out of the box. It isn't supposed to be a technical achievement to be able to watch a movie on your computer.
If the fanatics whining about Linspire were to get usable replacements for Linux proprietary apps and the Winblows legacy proprietary apps like PaintShopPro (yes, I've tried GIMP2... it is teh suxx0rs), I'd be happy to ditch my proprietary apps. I can't even use OpenOffice2.0 to do my final edits of Linux tutorials because it doesn't handle highlighting in Word document files correctly... in the way my editors expect. They can write a shitload of new drivers while they're at it, or find a way to induce the vendors to do so. [not impossible, persuading IBM/HP/other megacorps supporting Linux to lean on peripheral vendors might not be difficult]
The zealots don't get that we people who use our boxes to make a living can't wait for the OpenSource movement to get around to writing everything we need in order to make our computers function as we need them to. We don't live in mommy's basement, we work for a living. They can fix the problems, find ways to get them fixed, or STFU.
Tech Public Policy stuff
Even Linus Torvalds runs a Mac as his personal workstation.
It isn't just drivers, it's the apps that exist on Windows and OSX that don't have usable equivalents in Open Source.
As these usable apps (graphics is the biggest problem... when will a usable replacement for GIMP2 appear?) appear, I'll be nuking my Windows legacy apps I run in Win4Lin (Linux proprietary Win virtual environment) one by one. Though what I actually expect is that CPU hardware virtualation support will make it possible to simply run Windows as just another OS and I'll get rid of Win4Lin first... and keep running legacy apps.
Just because Linux zealots think an app is adequate means that anyone in his right mind will actually use it. The biggest obstacle to Linux on the desktop is the apps zealots think are "good enough" so they've stopped working on the UI or functionality... that suck.
Tech Public Policy stuff
Read this slashdot story and the responses CAREFULLY and then come back and post your response. Many people are PISSED about the usability of GNU software who WOULD like to use it, how hard is that to understand?
/ 1315235
http://apple.slashdot.org/article.pl?sid=05/06/12
Tired of all the isms, don't exploit people as an employer, or a government, mmmmK?
I'm not going to use something that sucks (e.g. GIMP2) if I actually have to work using it and there's a proprietary app that does the job right.
Tech Public Policy stuff
I think so.
Of course it can happen. Another poster remembered a specific example - the drivers for the old pre-ata drives that shipped with the PC/XT were recently dropped. They never shipped on a machine with a 386 processor best I know, but some did wind up in 386s anyway. There might be a few left in use, but it seems doubtful.
Proprietary drivers often go unsupported *much* quicker.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Friends don't let friends enable ecmascript.
- I don't believe in God.
- I don't use Linux.
Religion and Linux zealots have one thing in common. Both believe "You're either with us or against us." If it has to be like that, then I'm against you both. Having R all TFAs, it all reads like a big dick size contest. Pitiful.It's a very dark ride.
The whole point of a codec is to do "calculations on hunks of data".
Read compressed data. Write uncompressed data. Repeat.
Also, prior to setting seccomp, a mmap can be prepared.
this guy's post brought up some intersting points. I belive that their should be a free alternative to everything. But we are not in a perfect world, so I think we could work towards that though education. I agree with what he says, the FOSS zelots are making things worse trying to turn this into an all or nothing game. Some people just dont have the choice right now...