Resisting the binary blob
This reviewer is not the only one to express this point of view; there would appear to be a rising chorus out there calling on Linux distributors to load up their systems with proprietary code. Some distributors have heeded this call, as witnessed by (for example) Ubuntu's decision to include more binary drivers by default in its next release.
It's not too hard to see where this pressure is coming from. A prospective user with a problematic laptop will be happier with a distribution which "just works." Most of the people who truly care about free software are likely to be using a free system already, so it is easy to imagine that the next wave of users will be less concerned - at the outset - about software freedom. So they will gravitate toward a system which does what they want to do (running on closed hardware, playing patent-encumbered media, etc.) without concerning themselves much about the provenance of the software they are using.
The fact that many of these users worry little about software freedom now does not mean that they will never care, however. Very few of us were born knowing that free software is a better solution, that using free software is an important part of being free in general. Just like most of us have learned, over time, that saving some of the money we earn, while perhaps being inconvenient in the short term, brings long-term benefits, we have also learned that using free software - and helping to improve that software - is better in the long term. Certainly some subset of the new users coming to Linux will come to understand this fact as well.
But it will not matter how well these users understand the fine points of software freedom if, by the time they have figured it out, there are no free operating systems for them to run. If we want free systems then, we have to build and use free systems now. There can be a place for a binary blob which enables a specific bit of hardware to work; your editor would argue that running such a blob is not an inherently immoral act. But it is not necessarily a wise act, and a distribution which quietly installs such blobs on an unsuspecting user's system in the name of "it just works" is not necessarily doing that user any favors.
As a thought experiment, consider how things might have gone if the Linux community had accepted the "just works (most of the time)" non-free Java implementation that Sun made available. Linux distributors, rather than put large amounts of work into making Java code work with free alternatives, could have simply shipped Sun's version. Had they done so, would we have (the promise of) a GPL-licensed Java from Sun now? If we simply accept proprietary drivers in the name of "it just works," when, exactly, do we think free drivers will become available?
So criticism of Fedora - or any other distributor which sticks to free software principles - is, at best, misplaced. There are proprietary systems out there for people who want to run them, but Linux is about free software. It makes no sense to try to push proprietary code onto a distribution which has set a goal of being 100% free, and it is silly to criticize such a distribution for containing only free software. We should, instead, be appreciative of the vast amount of work that has gone into giving us a 100% free system - and help to improve that system.
Along these lines, it becomes natural to wonder why the Free Software Foundation has not recognized the work done by the Fedora Project to make its distribution entirely free. Instead, the FSF has put its energy into promoting obscure distributions like gNewSense and UTUTO. It seems that the Fedora developers and the FSF have been talking about recognition for Fedora, resulting in the posting of this message from Richard Stallman. It covers a number of issues, including firmware, fonts, patents, and more. One sticking point, it would seem, is this:
Given the people involved with Fedora, and the work that has been done to
eliminate packages with problematic licensing, your editor has no qualms in
saying that Fedora is a truly free distribution. It is unfortunate that
the work which has gone into the creation of this distribution is not as widely recognized as it should be. If we want to promote free software, and if we want to live in a world where we can use exclusively free software, we
should not hesitate to acknowledge the work of those who have built free
systems, and who have not given in to those pushing for the addition of
proprietary code. They are doing the work we so very much want to see
done, and we are far richer for it.
(Log in to post comments)
Resisting the binary blob
Posted Nov 14, 2006 17:11 UTC (Tue) by carlabi (guest, #34248) [Link]
Is my editor running Fedora?
Your editor's distribution
Posted Nov 14, 2006 17:27 UTC (Tue) by corbet (editor, #1) [Link]
Your editor runs a different distribution on every system he has. He also really tries to avoid endorsing any particular distribution. A review of past articles will reveal, however, that there is a Fedora development ("rawhide") system in a fairly central position on your editor's network.
Resisting the binary blob
Posted Nov 14, 2006 17:27 UTC (Tue) by pbardet (guest, #22762) [Link]
It's too bad that viewed from the outside world, FC is one of the major linux distributions and that since it's difficult to get anything done with it, "it must be the same with any other Linux distro".
Some other distros seem to have found a way to simplify user's life by making non free packages accessible, even though they're not part of the standard install.
Resisting the binary blob
Posted Nov 14, 2006 21:18 UTC (Tue) by seyman (subscriber, #1172) [Link]
Some other distros seem to have found a way to simplify user's life by making non free packages accessible
- Make rpms of non-free software
- Make a repo of said rpms
- Make a $foo-release rpm that contains the gpg key you signed the rpms with and the yum configuration file for said repo
- Link to this rpm from a webpage
- Click on the link
- Enter your root password when prompted
- Run "Add/Remove Software" from the menu
Resisting the binary blob
Posted Nov 14, 2006 21:52 UTC (Tue) by pbardet (guest, #22762) [Link]
Your point being ?If it's easy for you to do those steps, it may not be that easy for other. If you want to see Linux used by non-techie people, you have to simplify it. I have nothing against FC becoming the techie distro, used by .0001% of the world, while easy distros being used by the rest of us, I just think it's too bad they don't want to be part of the succes story. My goal is to see linux succeed against MS and being accepted as a real alternative.
Regular people don't care about open/closed source. They care about a working system, right here, right now... Until "we" figure this out, Linux will stay away from mainstream, like it or not.
Resisting the binary blob
Posted Nov 14, 2006 22:43 UTC (Tue) by khim (subscriber, #9252) [Link]
My goal is to see linux succeed against MS and being accepted as a real alternative.
...by sacrificing essential freedoms, security, reliability and control in the process. What will be the point ? You'll just replace one pice of junk with another one and one corporation who controls everything (or at least claims that it does) with consortium who'll enjoy finger-pointing.
Those who would give up Essential Liberty to purchase a little Temporary Safety, deserve neither Liberty nor Safety - that's exactly our case: if you'll accept binary blobs today you'll have no control over you system tomorrow. And why will I need another uncontrollable mess on my desktop ?
Regular people don't care about open/closed source. They care about a working system, right here, right now... Until "we" figure this out, Linux will stay away from mainstream, like it or not.
If people really like to have the system which works when the "stars are just right" - they can do it now and will undobtedly have this choice in the future. Why convert Linux to such a system ?
Resisting the binary blob
Posted Nov 15, 2006 0:39 UTC (Wed) by emkey (guest, #144) [Link]
Fundamentally I and a fair number of other people want a real alternative to Microsoft. I happen to feel that as a general rule open source software is a superior model that will win out in the long term, but with one qualifier... It has to gain significant market share first. The battle in the server realm has been largely won. The next step is the desktop. To gain ground there binary drivers are a requirement for now. I don't like this, but I'm a pragmatist.
Don't talk to me about fundamental freedoms when the fundamental freedom I care most about is my right to choose something other than a Microsoft product in twenty years time.
Resisting the binary blob
Posted Nov 15, 2006 2:29 UTC (Wed) by vonbrand (guest, #4458) [Link]
What would be the point at "winning the desktop for open source", only that the system is not really open source, and moreover depends in fundamental ways on closed source pieces to even minimally work?
I, for one, prefer open source because it (mostly) works just fine, and (very important!) if it works with some piece of hardware today, it'll probably do so for the foreseable future, even if the vendor of said device would dearly like me to shell out for a yearly "upgrade", and nugde in that direction by just leaving their binary drivers to rot.
Resisting the binary blob
Posted Nov 16, 2006 18:43 UTC (Thu) by emkey (guest, #144) [Link]
The point is that 95% of and ideal world is a lot better then 0% of an ideal world.
Resisting the binary blob
Posted Nov 24, 2006 6:32 UTC (Fri) by linuxrocks123 (subscriber, #34648) [Link]
To get anything resembling an ideal world, we must be firm in our demands.Binary drivers are not acceptable in the long run, and we should make it
as undesirable as we can for vendors to produce them.
Putting pressure on vendors works, and things are much better now because
of it. You used to basically need the proprietary browser Netscape 4 to
view web sites that had images. Clearly, we've come a long way.
Resisting the binary blob
Posted Nov 15, 2006 8:14 UTC (Wed) by khim (subscriber, #9252) [Link]
It has to gain significant market share first.
Market share is irrelevant if you don't have control. Lotus 1-2-3 and Wordperfect had "market share" in 1990. Microsoft had control. By the 1995 Lotus 1-2-3 and Worperfect become irrelevant. They are still around, but... who really cares about them ? Today it's about MS Office and OpenOffice.org, not about former gigants...
Don't talk to me about fundamental freedoms when the fundamental freedom I care most about is my right to choose something other than a Microsoft product in twenty years time.
The only way to make this fredoom reality is to reject binary blobs today. Sorry. Either you are talking about "something other than a Microsoft in twenty years time" or you are talking about "working system, right here, right now". There are no middle ground. Why ? It's easy. Microsoft's most famous strategy is embrace, extend, extinguish. Do you really think Microsoft will hesistate to apply the strategy it perfected for the last 30 years to kill "something other than a Microsoft product" ? Binary blobs are just begging for this strategy to be applied. Turn blind eye first (that's what the Microsoft is doing), then wait while the same technology is used in Linux and in Windows drivers - and finally ask vendors of few key drivers to "remove their IP" (to make drivers inferior to their Windows ones) - poof no more alternative. Purrfect.
Resisting the binary blob
Posted Nov 15, 2006 14:57 UTC (Wed) by pbardet (guest, #22762) [Link]
MS had control because the had Windows and they could break any API easily to fail other software for unknown reasons...I just hope Linux will prove it's better without using those tactics. That's why I'm opposed to any MS contributions to the Linux world as they're currently trying through Novell.
I don't see why because one distribution uses a binary blob, development will stop on the free version for all other distributions, especially the hard code open-source ones. You can't compare commercial and open-source software development.
The only reason people don't want binary drivers, is because they fear it will actually win in the end. It's up to the open-source programmers to prove them wrong. But it's so much easier to try to prevent things from happening than creating.
Resisting the binary blob
Posted Nov 15, 2006 15:25 UTC (Wed) by tjc (guest, #137) [Link]
I don't see why because one distribution uses a binary blob, development will stop on the free version for all other distributions, especially the hard code open-source ones.On the other hand, just because one distribution doesn't include binary blobs doesn't mean that development will stop on all the not-so-free distributions. I don't see that you have made any compelling arguments why FC should include them, other than you seem want them.
In your original post you mentioned that FC is one of the more popular distributions, and then go on to assume that they would be even more popular if they were to include binary blobs. Could it be that they would become less popular if they included them?
Resisting the binary blob
Posted Nov 16, 2006 18:50 UTC (Thu) by emkey (guest, #144) [Link]
The only way to make this freedom reality is to reject binary blobs today.And you base this assumption on what exactly? Idealism goes nowhere in the real world. I don't like that, but it's a reality. Open Source software has done well to date because there is a good and compelling business case for it. At this point the perception is that this isn't true on the desktop. Linux needs to make significant inroads on the desktop for that to change. And lack of driver support is a significant impediment to having that happen. The good news? Binary drivers exist that solve most of the problems. The bad news? There is no bad news, though some people seem obsessed with pretending there is.
Resisting the binary blob
Posted Nov 17, 2006 7:38 UTC (Fri) by man_ls (guest, #15091) [Link]
Binary drivers exist that solve most of the problems.I wish that was true. Unfortunately it is not. Binary drivers just add more problems to an already fragile desktop market:
- they are brittle, and cannot be repaired easily,
- they don't admit self-maintenance,
- you cannot benefit from others in the community solving the problems (e.g. sending patches to lkml),
- in practice they stop some of these people from working on free alternatives, just as BitKeeper prevented use of free tools,
- when the vendor gets tired of supporting your hardware you are stuck with a buggy solution, just as in the Windows or Mac OS X world,
- efforts cannot be pooled: vendors cannot reuse the code for their own drivers, so the wheel is reinvented for each one,
- you cannot get help on lkml because your kernel is tainted,
- and finally, their legal status is not clear so distributing them could open distros to being sued by their own developers, not by obscure submarine patent outfits.
Resisting the binary blob
Posted Nov 30, 2006 19:38 UTC (Thu) by emkey (guest, #144) [Link]
Binary drivers do not in and of themselves prevent free drivers. Ergo, they cause no problems in that arena. Which pretty much makes every point you raised moot.
Again, there is no issue here. None. What we have is a very obscure and largely pointless pseudo religious objection by some people.
Resisting the binary blob
Posted Dec 1, 2006 3:01 UTC (Fri) by bronson (subscriber, #4806) [Link]
If the binary drivers are installed by default, instead of the free drivers, then every single one of man_ls's points is valid. Binary drivers prevent free drivers when the user doesn't know that the free driver exists.
Resisting the binary blob
Posted Nov 17, 2006 23:03 UTC (Fri) by roelofs (guest, #2599) [Link]
Idealism goes nowhere in the real world. ... The bad news? There is no bad news, though some people seem obsessed with pretending there is.And you call khim idealistic? Whoa, my irony-meter just exploded...
You want a compelling business case? I made the business case for my last job. It went something like this: open source gives you control (binary blobs don't, period); it tends to give you security (binary blobs are much murkier in this regard); and it tends to give you reliability (binary blobs tend not to). It's also cheap, but that was merely fourth or fifth on the list. Those are the features that businesses care about.
As for the desktop: personally, I really don't care if there's a business case for it, compelling or otherwise. The business cases for FLOSS servers showed up long after open source itself did, and large market share is relevant only if it advances the FLOSS cause. A large market share that's 95% composed of binary-blob users is utterly useless as a tool to convince device makers to open their specifications, so who needs it? More power to Fedora and Red Hat, says I.
Greg
Resisting the binary blob
Posted Nov 16, 2006 18:58 UTC (Thu) by lysse (guest, #3190) [Link]
> I happen to feel that as a general rule open source software is a superior model that will win out in the long term, but with one qualifier... It has to gain significant market share first.
If it undermines or abandons its key point of differentiation to gain a few more users now, how is it *ever* going to acquire major market share?
The one thing free software has that can never be taken away is its ethical basis - but when so many people are so quick to advocate surrendering it wholesale for a little bit of convenience right now, perhaps it doesn't *need* to be taken.
Oh, and you want "a real alternative to Microsoft"? Buy a Mac; for if Microsoft were to release Windows under a free software licence, the community would welcome their conversion with open arms. Free software is about principles, not alternatives; choice is merely the shoddy counterfeit of freedom.
I refuse to subscribe to the view that principles are worth less than the sacrifices necessary to maintain them, which seems to be what is being advocated here. For those who think they are, there are two perfectly good operating systems out there in widespread use, and likely one or the other of them came gratis with the last computer you bought. But consider the adjectives usually used to describe people who postpone their purported principles to purchase petty popularity.
Resisting the binary blob
Posted Nov 17, 2006 0:34 UTC (Fri) by emkey (guest, #144) [Link]
The one thing free software has that can never be taken away is its ethical basisWhat in particular makes free software more ethical than for pay software? And can those rule be applied generally, or are they specific to software only? Is free software actually free?
I refuse to subscribe to the view that principles are worth less than the sacrifices necessary to maintain them
I refuse to be sacrificed for a principle that A, I don't agree with and B, I don't believe to have a factual basis. I've elaborated more on my point in other responses to the up thread post I made. I won't repeat myself here.
Today's quiz
Posted Nov 17, 2006 7:59 UTC (Fri) by man_ls (guest, #15091) [Link]
What in particular makes free software more ethical than for pay software?Nothing, considering that free software can also be for pay software.
And can those rule be applied generally, or are they specific to software only?If you refer to free vs proprietary software, yes: the rule is that helping others is good, while not sharing good things (like your knowledge) is bad. Not even primary school knowledge, this is kindergarten stuff.
Is free software actually free?A very old question; the best answer is to define what you mean by free. This is why the GNU project publishes the free software definition. If you agree with the definition then it is free. Others have a different definition, or an even more different definition. So you get to make the decision.
I refuse to be sacrificed for a principle that A, I don't agree with and B, I don't believe to have a factual basis.A principle which you don't seem to understand, since it is pretty practical in nature. And yet you benefit from the fruits of said principle... Well, whatever one says of your position, you are in good company.
Today's quiz
Posted Nov 17, 2006 16:13 UTC (Fri) by bronson (subscriber, #4806) [Link]
You imply that the kernel guys don't understand a principle that normal people are supposed to have learned in Kindergarten? You can't be serious.
Today's quiz
Posted Nov 18, 2006 10:47 UTC (Sat) by man_ls (guest, #15091) [Link]
Not really. It is a bit confusing because there are two sets of principles here. The first one is the set of principles embodied in free software: that software must be free (just as speech, since it is another form of human expression), that proprietary (closed) software is bad, that binary blobs have bad consequences.I explained that this set of principles applies generally because it derives from another, simpler set of principles: that helping others is good, and not sharing good things is bad. Not all children learn these ideas; remember the greedy kid in the corner who has all the goods but no friends. What is worse: we tend to forget those principles as we grow up.
What our correspondent here (as well as certain kernel developers) don't seem to understand is that both sets of principles are very much the same; or, being more precise, we do not agree to the extent of the equivalence. Both are practical in nature, since they have immediate, real-world consequences (as the Open Source crowd likes to point out). Both can be theoretically justified (as in games theory or elaborate essays). And both have ethical consequences which you have to consider.
Many kernel devs don't think that the Kindergarten principle of helping others necessarily applies to their work, which is fine -- but wrong. Torvalds in particular seems to think that it is only about sharing code between professionals (tit-for-tat, a famous Kindergarten principle + games theoretical concept [as I have just learned]), but that Tivo-like locked down systems (which do not certainly help others) are OK. Again, not enough thought is given to the consequences of one's actions, which is not so surprising since you have to be a bit obsessive to think things through in this manner.
Resisting the binary blob
Posted Nov 22, 2006 14:25 UTC (Wed) by lysse (guest, #3190) [Link]
> I refuse to be sacrificed for a principle that A, I don't agree with and B, I don't believe to have a factual basis.
As I say, there is plenty of other software from which you can choose; some of it is even gratis. But you appear to be demanding that people who started doing something on a point of principle discard that point of principle to make your life easier, and that's not "you not being sacrificed" - that's you behaving like a petulant child.
Resisting the binary blob
Posted Nov 22, 2006 14:30 UTC (Wed) by lysse (guest, #3190) [Link]
Oops - should have said this before posting my other comment.
> What in particular makes free software more ethical than for pay software? And can those rule be applied generally, or are they specific to software only?
Free software is not "more ethical than for-pay software", and I didn't say it was. I said its ethical basis - ie. the moral decision which led to the creation of free software as a concept - was its differentiator; I didn't imply anything about whether that moral decision was better or worse than others.
I guess, from this, the real problem you have with free software is that *it makes you feel judged*. Except it doesn't - YOU make you feel judged; free software only sets and seeks to protect its own standards, it doesn't seek to bring the rest of the world into compliance with it. That certainly accords with one generally accepted definition of freedom.
> Is free software actually free?
Now that question has to be a piece of FUD right up there with some of Ballmer's finest... Careful; your agenda is showing.
Resisting the binary blob
Posted Nov 24, 2006 3:28 UTC (Fri) by dkite (guest, #4577) [Link]
Sure, lets make everything work out of the box. Let's make it easy toinstall and run on anything.
Explain to me how having binary blobs, or closed proprietary bits of
software can make this happen?
The desktop is almost there. Most of the major components work very well.
Large parts of the desktop infrastructure is nearing the stage where it
can be assigned a 1.0 version number. There will be very large changes in
the coming years to these basic building blocks when improvements are
implemented, the inevitable dead ends are fixed. One illustration is
Ubuntu rewriting the init script system.
The only thing that keeps this whole mess working at all is the fact that
we have code and licenses that allow us to make it work. Remember, Linux
Desktop (tm) is made up of dozens of different projects. The kernel,
printing subsystem, desktop environments, multimedia libraries, graphics
drivers, office software, scripting languages, compilers, etc. How can
closed source blobs work in this environment?
Maybe someone can make things work now. But what about next year when the
progress has given us new versions of almost everything that makes up the
desktop environment?
I speak from experience. I have had to try to maintain a few systems with
different binary drivers. Security updates break the system, and require
rebuilding, hunting down updated drivers from different sources. If the
distro is doing it, all that means is less resources available for
improvements due to the amount of resources required to keep binary trash
working. The more dependance on binaries, the less stable our systems
will become.
And how is that supposed to gain large numbers of users?
Derek
Resisting the binary blob
Posted Nov 14, 2006 23:13 UTC (Tue) by seyman (subscriber, #1172) [Link]
Your point being ?
My point is that adding third party repos is easy. You were claiming the opposite.
If it's easy for you to do those steps, it may not be that easy for other. If you want to see Linux used by non-techie people, you have to simplify it.
Which part of
- Click the link
- Enter the root password
- Choose the "Add/Remove Software" entry in the menu
Resisting the binary blob
Posted Nov 14, 2006 23:52 UTC (Tue) by khim (subscriber, #9252) [Link]
The complicated step is omitted (as usual). It's
- Find the link
It looks strange when these same user will use Google, dig in the archives, ask in japanese forums and do a lot of other things to make hardware work in Windows - but they DO expect it to work "out of the box" in Linux. Why - is anybody's guess... If the user need to find a link - you've already lost...
The question: do we WANT to invite such users - is good, but different question...
Resisting the binary blob
Posted Nov 15, 2006 11:50 UTC (Wed) by seyman (subscriber, #1172) [Link]
Find the link
This shouldn't be needed. The link to a repo containing Nvidia drivers should be on Nvidia's website, a link to a repo containing ATI drivers should be on ATI's website, ....
Resisting the binary blob
Posted Nov 15, 2006 14:37 UTC (Wed) by pbardet (guest, #22762) [Link]
The thing is that those links are not there yet. Also, the last time I checked for a distro I could install, FC did not have links to point you to the binary drivers. You had to google around to find out how to do it, ie spend a few hours understanding why it doesn't work out straight out of the box, or CD.
In theory, it' easy, in practice, it's only easy for techies who want to take the time to do it. Regular people don't want this and stick to MS. When you install it, MS doesn't tell you go there find that, they just do it. Linux should be the same.
Once again, Linux is not only about freedom, it's also about choice and alternatives. Binary drivers are an alternative, that like it or not, will stay as long as Intellectual Property will be recognized in this world.
Sure, I wish I could install a working system only from free software. I don't think it can happen yet, based on the market share of Linux.
If we want to see distros 100% free software, I can't see why we can not see distros that want to use non-free drivers as a base. This is freedom. Right now, Ubuntu is oriented toward the user, I just switched from Gentoo to it for this specific reason. The day they feel the free driver is better, they will switch back to it (if they ever switch to the binary). I'm pretty sure they want to stay away as much as possible from ties with other companies, as any other linux distribution. But in the meantime, if making it easy for your user is the goal, let's use the best as default and let techies change their systems since it's so easy for them.
Only the future will tell us if one way of doing is better than the other. I find it interesting to see how they would succeed, but I may never know.
Resisting the binary blob
Posted Nov 15, 2006 23:32 UTC (Wed) by quaid (guest, #26101) [Link]
The thing is that those links are not there yet. Also, the last time I checked for a distro I could install, FC did not have links to point you to the binary drivers. You had to google around to find out how to do it, ie spend a few hours understanding why it doesn't work out straight out of the box, or CD.
IANAL, TINLA, etc.
Fedora does not have an easy way to find links to these "solutions" because to provide such links may be against the law in the US.
Read this page for more details:
http://fedoraproject.org/wiki/ForbiddenItems
Once again, Linux is not only about freedom, it's also about choice and alternatives. Binary drivers are an alternative, that like it or not, will stay as long as Intellectual Property will be recognized in this world.
Yes, it has always been true that those who are free/libre have the right to choose to make themselves un-free.
You also have the freedom to make or use a distro that is un-free. The point of this article is that Fedora should be celebrated for being free rather than vilified for failing to support someone else's desire not to be free.
Sure, I wish I could install a working system only from free software. I don't think it can happen yet, based on the market share of Linux.
I certainly don't have any problems installing a working system only from free software. Seems like many other people also don't have any problems. :)
Resisting the binary blob
Posted Nov 16, 2006 14:04 UTC (Thu) by pbardet (guest, #22762) [Link]
"Fedora does not have an easy way to find links to these "solutions" because to provide such links may be against the law in the US." I don't care about the reason. I care about the result: It's not easy for end-users to get things done with such a distribution. I've heard it so many times before "oh it's not our fault, it's the law's fault". "Yes, it has always been true that those who are free/libre have the right to choose to make themselves un-free." I feel perfectly free when using Linux, whether it contains free or non-free components to make it work properly. If you don't, it's your problem, not mine. I make a living of producing proprietary software, and both models can interoperate fine. I have no problem buying a software if it meets my need. The fact is that whether I use free or non-free software, I always feel a little screwed by the fact that if I encounter a problem, I'm always dependant on other parties, whether they're open source or commercial since I don't have the time, or the experience to fix drivers or complicated software. In theory, open source is great since you can fix, but in practice, it's not true for most people using the software since they have no capacity to do so. "I certainly don't have any problems installing a working system only from free software." Good for you. I still haven't found a way to get decent video playback on my MythTV box with Nvidia card with the free driver, while the Nvidia driver worked fine right away, no tweaking, no fussing around. When xpdf, kpdf or any other version craps on a pdf file, I'm glad to have acroread installed on my machine, without getting to hunt down links. The opposite is true also (which sometimes raises questions about acroread, but it's another story). I just had to click on Firefox's (free software) window complaining about missing plugin. Weird that a free software can link to proprietary one, but it would not be OK for the distribution to do it by default.
Resisting the binary blob
Posted Nov 16, 2006 16:31 UTC (Thu) by IkeTo (subscriber, #2122) [Link]
> I don't care about the reason. I care about the result
Fedora is a community project. If you don't care the reason given by a community project, expect them not to care you either.
> I feel perfectly free when using Linux
The point is *not* whether *you* feel it free, but instead whether the community building them feel it free *according to their vision*. Using binary blob is not. End of discussion.
> In theory, open source is great since you can fix, but in practice, it's not true for most people
Open source is great because you can fix, and if you are, like most people, too busy or not qualified enough, you can find others to fix. No such freedom in other platforms. The only one who can fix the software you buy is the vendor selling it. If that single entity doesn't think solving your problem is worth their time, bad luck for you.
> My goal is to see linux succeed against MS and being accepted as a real alternative.
There had been alternatives. DR-DOS, OS/2 and Gems come to mind. Now there is just Linux. Don't you understand that each such attempt is very costly, and don't you know that the one thing that keep MS from using the same tactics against Linux is exactly what you are complaining about, i.e., open source?
Resisting the binary blob
Posted Nov 15, 2006 22:47 UTC (Wed) by pbardet (guest, #22762) [Link]
You're very right, it's awfully easy to install Nvidia binary drivers...http://www.linuxforums.org/multimedia/installing_nvidia_3...
Resisting the binary blob
Posted Nov 24, 2006 8:02 UTC (Fri) by linuxrocks123 (subscriber, #34648) [Link]
http://forums.windrivers.com/showthread.php?t=77687
The fact that you bother to quibble about the relative ease of installing
binary blobs is an indicator of your short-sightedness.
Resisting the binary blob
Posted Nov 14, 2006 23:52 UTC (Tue) by bojan (subscriber, #14302) [Link]
> Your point being?
It could even be (if you want it to be):
- make such a repo
- advertise it
- charge subscription for it
- make heaps of $$
Resisting the binary blob
Posted Nov 15, 2006 18:26 UTC (Wed) by caitlinbestler (guest, #32532) [Link]
More to the point, there are a large body of users who expect thehardware vendor to "just make things work", and if the hardware
vendor doesn't make things work they'll pick a different hardware
vendor. Now those users would like their hardware vendor to be
able to support choices other than Linux, but they have no desire
to become their own OEM and assemble the correct working set of
drivers.
So the question is whether hardware vendors are enabled to support
users who *want* a turnkey package to select Linux as their OS.
Resisting the binary blob
Posted Nov 15, 2006 23:29 UTC (Wed) by JohnNilsson (guest, #41242) [Link]
Regular people don't care about open/closed source. They care about a working system, right here, right now... Until "we" figure this out, Linux will stay away from mainstream, like it or not.
This round of the battle for a Free World (TM) will soon be over. Things will stabilize again, and we will have to wait another century for the next disturbance of the equilibrium for another shot.
If we fail to educate people before that, the freedom movement will have lost, again.
Resisting the binary blob
Posted Nov 15, 2006 1:10 UTC (Wed) by grouch (guest, #27289) [Link]
It's too bad that viewed from the outside world, FC is one of the major linux distributions and that since it's difficult to get anything done with it, "it must be the same with any other Linux distro".
That's an assumption and a blanket accusation with nothing to back it up.
Some other distros seem to have found a way to simplify user's life by making non free packages accessible, even though they're not part of the standard install.
Most complaints I've heard from people who tried FC and moved on to other distributions focused on the rate of change, not on non-free packages. (Note that this is just as anecdotal, rather than analytical, as your assertions). Newbies seem to want stability and predictability in a desktop. Those who want to race along the bleeding edge are generally quite content to deal with the extra work they have to do because of the occasional crumbling of that edge.
If all distributions were alike in their goals, and all drivers and codecs were non-free, we'd end up with MS. Those who promote binary blobs and non-free software as necessities for usability are promoting homogeneity, wherein all operating systems work the same with all combinations of hardware for all users. I've seen the user experience that results from attempting to attain that goal. No thanks.
I'll keep using my Debian and poking fun at FC users for their unending instability, while at the same time cheering them on. They, in turn, will continue poking fun at my "old" packages, while at the same time making good use of them. Free software encourages experimentation and growth in multiple directions, by any individual or group that chooses to try. Non-free software restricts all users.
Resisting the binary blob
Posted Nov 14, 2006 17:40 UTC (Tue) by allesfresser (guest, #216) [Link]
"But it will not matter how well these users understand the fine points of software freedom if, by the time they have figured it out, there are no free operating systems for them to run. If we want free systems then, we have to build and use free systems now."
In other words:
"...it is not our part to master all the tides of the world, but to do what is in us for the succour of those years wherein we are set, uprooting the evil in the fields that we know, so that those who live after may have clean earth to till. What weather they shall have is not ours to rule." (J.R.R. Tolkien)
Temporary inconvenience is often part of the price of freedom.
Resisting the binary blob
Posted Nov 15, 2006 1:08 UTC (Wed) by Mithrandir (guest, #3031) [Link]
Except that these days... what weather they shall have IS ours to rule.
Not the point I know, sorry ;)
Resisting the binary blob
Posted Nov 14, 2006 17:54 UTC (Tue) by oak (guest, #2786) [Link]
Maybe this could be a free article, it's more of an opinion-piece(with an opinion I agree with :)), than reseach article like most
of the other subscriber-only articles?
Resisting the binary blob
Posted Nov 14, 2006 19:17 UTC (Tue) by rknop (guest, #66) [Link]
give it a week :)
Resisting the binary blob
Posted Nov 16, 2006 13:50 UTC (Thu) by liljencrantz (guest, #28458) [Link]
Agreed. This article does not have a short half life.
Resisting the binary blob -> Debian
Posted Nov 14, 2006 18:00 UTC (Tue) by ballombe (subscriber, #9523) [Link]
There is one distribution where developers will go out of there way to remove binary blob from otherwise free software package and this is Debian. It is not sufficient to remove packages well-known to contain binary blob.
Resisting the binary blob -> Debian
Posted Nov 14, 2006 20:35 UTC (Tue) by spot (guest, #15640) [Link]
And yet, Debian isn't "Free" by the FSF guidelines.
Resisting the binary blob -> Debian
Posted Nov 14, 2006 21:34 UTC (Tue) by hmh (subscriber, #3838) [Link]
That's only fair given that Debian doesn't consider a lot of the recent stuff from the FSF "free" either. RMS doesn't like the non-free archive the Debian project keeps, and Debian doesn't "like" the DFSG-incompatible stuff the GNU project ships.
The GFDL did more to keep the "non-free" .deb archive and support structure alive within Debian than anything else ever did, IMO... Talks within Debian about decomissioning non-free have gone silent for a while now, AFAIK. IMHO, too much important stuff lives in non-free currently to even consider doing that, thanks to the GFDL.
Resisting the binary blob -> Debian
Posted Nov 14, 2006 21:03 UTC (Tue) by seyman (subscriber, #1172) [Link]
<p><em>There is one distribution where</em></p>
<p>Going further down the road, isn't this the whole point of having distributions in the first place ?</p>
<p>There are distributions that are turn-key in nature, that will come fully loaded with all the binary bits you need to get every last bit's worth out of your computer (well, provided your architecture is approved by all your vendors) and the only thing it will cost you is your freedom.</p>
<p>And existing besides those distributions, there are those that actually try to preserve that freedom and do their best to serve their users without taking their liberties away from them.</p>
<p>Why can't people chose one or the other ?</p>
Resisting the binary blob -> Debian
Posted Nov 14, 2006 21:05 UTC (Tue) by seyman (subscriber, #1172) [Link]
[Sorry about the parent post.Wrong button.]
There is one distribution where
Going further down the road, isn't this the whole point of having distributions in the first place ?
There are distributions that are turn-key in nature, that will come fully loaded with all the binary bits you need to get every last bit's worth out of your computer (well, provided your architecture is approved by all your vendors) and the only thing it will cost you is your freedom.
And existing besides those distributions, there are those that actually try to preserve that freedom and do their best to serve their users without taking their liberties away from them.
Why can't people chose one or the other ?
Blobists remind me of open relay postmasters for some reason.
Posted Nov 14, 2006 18:31 UTC (Tue) by dmarti (subscriber, #11625) [Link]
1998: "I'm tired of all you anti-spam mail admins saying I should shut down my open SMTP relay. I just want mail to work -- can't you understand that?"2006: "I'm tired of all you anti-blob kernel hackers breaking support for proprietary stuff in kernel space. I just want Linux to work -- can't you understand that?"
Blobists remind me of open relay postmasters for some reason.
Posted Nov 15, 2006 0:38 UTC (Wed) by dark (guest, #8483) [Link]
And in 2006, we have more spam than ever, and mail doesn't work anymore.
I'm not sure what you're trying to say there :)
Blobists remind me of open relay postmasters for some reason.
Posted Nov 15, 2006 16:42 UTC (Wed) by jzbiciak (guest, #5246) [Link]
Right, but typically that spam comes from zombied Windows PCs.
Resisting the binary blob
Posted Nov 14, 2006 18:52 UTC (Tue) by smitty_one_each (subscriber, #28989) [Link]
Shopping on Emperor Linux yesterday, I was encouraged that it is quite easy to design a system to use the Intel video chip.Hopefully it gets increasingly easy to 'just say no' to the blob.
Responding to:
>so it is easy to imagine that the next wave of users will be less concerned - at the outset - about software freedom
My wife was given a complementary low-end iPod by her company, and was horribly dismayed that she couldn't just move her existing music (in whatever format RealPlayer uses) onto it.
She was sufficiently disappointed with the lock-in that I may get the gadget for a little mod-action. Mwahahahaha.
Sadly, too many sheep just line up for the shearing...
Resisting the binary blob
Posted Nov 15, 2006 1:17 UTC (Wed) by grouch (guest, #27289) [Link]
My wife was given a complementary low-end iPod by her company, and was horribly dismayed that she couldn't just move her existing music (in whatever format RealPlayer uses) onto it.
She was sufficiently disappointed with the lock-in that I may get the gadget for a little mod-action. Mwahahahaha.
You probably already know about this, but others might not:
um, you are already locked in by Real
Posted Nov 16, 2006 5:52 UTC (Thu) by stevenj (guest, #421) [Link]
If you are using RealAudio format then you are already trapped by vendor lock-in, regardless of whether you use an iPod.
Using open file formats is far more important than what hardware/OS you pick. The latter trap can be replaced, the former trap (especially for lossy compression) is much more difficult to escape.
um, you are already locked in by Real
Posted Nov 16, 2006 9:19 UTC (Thu) by morhippo (subscriber, #334) [Link]
Unless the format is DRM restricted, I think mplayer should be able to recode the realaudio music into any free format.
Resisting the binary blob
Posted Nov 14, 2006 19:05 UTC (Tue) by MathFox (guest, #6104) [Link]
The first piece of Free Software I conciously started using was gcc as my c++ compiler. Later came other nifty tools like gawk and bash, but still on a commercial Unix platform. When Linux grew up and XFree became decently stable I bought my first AMD box for it. Over the years Linux functionality improved, while the amount of closed source software on my machine went down. (Netscape -> Mozilla, StarOffice -> Open Office, etc.) I think we may have achieved Richard Stallman's dream of a desktop system with 100% Free software (Linuxbios/GNU/Linux/Xorg/Gnome/KDE/OOo/Firefox). Enough functionality for me to do my work on.
If there are proprietary packages installed on my machines, it's most likely the "proprietary kernel driver" package, but many of my machines work nicely without. (I'll miss Wifi on my laptop, but that has wired ethernet too.) Flashless web surfing is so quiet :-).
Users need help choosing their computers
Posted Nov 14, 2006 19:11 UTC (Tue) by JoeBuck (subscriber, #2330) [Link]
I disagree with Ubuntu's decision, but one of the major forces pushing them in that direction is that it is so difficult for users to know, or find out, which hardware is well-supported by free software and which isn't. Users need help buying computers, and it might be in Fedora's interest to steer people in the direction of vendors who will sell machines that just work. This might be problematic for Red Hat because of business relationships, particularly with vendors like Dell who seem interested in supporting Linux for servers but not laptops, but third-party sites could do it.The problem is that hardware changes so quickly, and the free Linux drivers improve so quickly, that most of the Linux hardware information on the net is out of date or filled with lists of hardware that is no longer sold.
Users need help choosing their computers
Posted Nov 14, 2006 19:23 UTC (Tue) by rknop (guest, #66) [Link]
it is so difficult for users to know, or find out, which hardware is well-supported by free software and which isn't.Amen.
In ages past, I seem to remember it not being too difficult to find some sort of hardware compatability guide that would help me figure out what would work. Back then, since Linux wasn't on hardware vendors' radar screens, what worked worked with free drivers, by and large. Yeah, lots of things didn't work, but for whatever you wanted, in general you could find something that would work that you could get at a local CompUSA.
Nowadays, for some things it's getting very tough to figure out what works. Video card with 3D support? Well, people tell you to buy NVidia, becuase it works... but not with free drivers. What works with free drivers?
How about wireless cards? It's nearly impossible to figure out which wireless cards that one might actually find in CompUSA will work at all, never mind with a free driver. Doing so requires digging through a slew of different web pages, many of which are a few years old and don't refer to any modern card, many of which are specific to certain classes of drivers, many of which are just forum archives. Last time around I ended up getting atheros cards because they'd work with MadWIFI -- but MadWIFI is one of those things that uses binary blogs, so I'm a little bit sad.
If we really believe in wanting the "common user" -- or even the serious but non kernel-coding geek like myself -- to be able to choose a completely-free system, somebody, somewhere, needs to put the effort into building and maintaining an up-to-date guide as to which hardware has reliable truly free drivers.
-Rob
Users need help choosing their computers
Posted Nov 14, 2006 20:01 UTC (Tue) by drag (guest, #31333) [Link]
The "average user" is incapable of building their own computer. The fact that people like us tend to think that a hardware guide would be usefull for most people just illistrates how ultra-geek we are. Generally speaking Linux community is out of touch with normal users.
The majority of computers sold today are laptop computers anyways and manufacturers typically try to make them as much 'black boxes' as they can.
In that case the hardware guide that people need to know is:
Buy a Intel-based notebook with Intel onboard video.
The trouble with that is that generally speaking those Intel systems come with Intel-based driver thatrequires closed source 'regulatary blob'. But wireless cards are much much easier to swap out generally then video.
That's about it.
The better solution is to get OEMs to support Linux on their hardware by default and encourage and aid the kernel developers in protecting their copyrights by creating distros that shun binary-blobs and binary-only drivers.
So if you get asked a question what hardware should I buy for Linux?
The easy answer is to go 'Get a System76 (or rcubed or whatnot) notebook computer with Intel video'
Your right
Posted Nov 14, 2006 20:18 UTC (Tue) by Felix_the_Mac (guest, #32242) [Link]
I am hoping to get a new system and am currently trying to figure out what the best graphics solution is using Free drivers; And as you say it is not easy.
Can X run the 3D desktop? Is Y faster than X when using free drivers? etc. etc.
If anybody can point me to a decent site I would be grateful.
But I think we need a 'Linux Hardware Quality Labs' that can test and certify hardware and give certification and logos.
Your right
Posted Nov 15, 2006 4:42 UTC (Wed) by drag (guest, #31333) [Link]
Intel is the only company that supports Free software video drivers.
These are supported by Free software drivers:
http://en.wikipedia.org/wiki/Intel_GMA
The GMA 3000 and x3000 are relatively new and drivers are relatively immature at this point.
ATI cards older then x1000 (the R500 series) are supported by Free software drivers. R200, R300, R400 ATI cards are supported by Free software drivers for the most part.
Supported:
http://en.wikipedia.org/wiki/R200
Supported:
http://en.wikipedia.org/wiki/R300
Supported:
http://en.wikipedia.org/wiki/X800
They are basic drivers. Probably give 40-60% the performance of ATI propriatory drivers. They are fairly reliable. There are certain models that aren't well supported..
See also:
http://megahurts.dk/rune/r300_status.html
Cards to avoid.
ATI, no open source support for 3d OR 2d.
http://en.wikipedia.org/wiki/Radeon_R520
http://en.wikipedia.org/wiki/Radeon_R600
Avoid all Nvidia cards. No open source support for 3d.
http://en.wikipedia.org/wiki/Nvidia
Prefereably for best compatability you want to get a Intel motherboard with 945g chipset. If you need better performance then that then get a R420 series ATI card, probably a PCIE x800. Those are decently supported by free software drivers and should work out of the box on most recently released distro versions.
The GMA x3000 in the 965g will probably be attractive also, but aren't to common yet.
Graphics nightmare
Posted Nov 16, 2006 5:41 UTC (Thu) by rakoenig (subscriber, #29855) [Link]
If the recommondation is "buy an Intel GMA" then this means: Buy a system with an Intel CPU based on an Intel Chipset. I'm working at a computer manufacturer and I know that there are a lot of customers, that want to buy non-Intel products because of pricing issues or just because they don't want an Intel monopoly on the market.
Those people don't have any option to get free 3D graphics support. Intel graphic is always on mainboards, but there is no PCIe plugin card that offers a graphic chip fully supported by free drivers.
The main manufactureres for graphic cards are ATI, Matrox and nVidia (alphabetically) where Matrox has a low market share. nVidia is evil by default because there is no free GPL driver available. ATI has some free drivers for older cards, but they won't fit into todays systems that come with PCIe slots instead of AGP.
The main problem why both ATI and nVidia refuse to GPL their drivers are IP issues I think. I doubt that releasing a graphics driver under the GPL will make it easier for somebody to "pirate" the hardware, but there is a high risk that one firm or the other has a software patent that is violated by the driver of the other party. In the long run the price of software patents is that we need to give up our freedom.
I'd love to run a machine with free software only, but I wouldn't like a monopoly of just one vendor. So I have to go with proprietary code even if its a pain in the ass. Sad, but there is too little market share of Linux overall to put pressure on the vendors.
Graphics nightmare
Posted Nov 16, 2006 12:49 UTC (Thu) by wookey (guest, #5501) [Link]
What about VIA? Their EPIA boards have onboard graphics which I think is their own, rather than intel's. The drivers are free. No doubt it's performance is weedy in comparison to the 3 main manufacturers, but for many applications (perticularly the multimedia the boxes often target) it is quite good enough.
So you can buy non-intel computers with free graphics. Still on-board, admittedly, and not at the high-performance end of things. Those people need to follow drag's advice on which cards have free drivers.
Graphics nightmare
Posted Nov 16, 2006 12:55 UTC (Thu) by wookey (guest, #5501) [Link]
Oh bollocks - a grocer's apostrophe in public <FX: hangs head in shame>.
Graphics nightmare
Posted Nov 16, 2006 15:56 UTC (Thu) by pizza (subscriber, #46) [Link]
> ATI has some free drivers for older cards, but they won't fit into todays systems that come with PCIe slots instead of AGP.
You can buy new R9250 (r200) cards in both PCI and AGP.
You can buy new X300 (r400 aka r300) in PCI, AGP, and PCIe.
Granted, their current-gen X1k stuff has zero opensource support, but even those are available in both PCI and AGP form factors too.
> I'd love to run a machine with free software only, but I wouldn't like a monopoly of just one vendor. So I have to go with proprietary code even if its a pain in the ass. Sad, but there is too little market share of Linux overall to put pressure on the vendors.
My laptop has only free drivers on it, including for the Radeon board. It does everything I need it to do.
Your right
Posted Nov 16, 2006 21:23 UTC (Thu) by zooko (guest, #2589) [Link]
I've long been a loyal user of AMD's 64-bit CPUs, but Intel's combination of first-class support forFree Software integrated video and top-notch performance from their new CPUs is making me
reconsider. Unless this situation changes, I'll be building my next system with Intel Inside, for the
first time in... In Ever, actually. (I think my first home-built system used a Cyrix CPU. :-))
Users need help choosing their computers
Posted Nov 14, 2006 21:10 UTC (Tue) by AJWM (guest, #15888) [Link]
> Video card with 3D support? Well, people tell you to buy NVidia, becuase it works... but not with free drivers. What works with free drivers?
Anything based on the ATI-9250 (or 9200, 8200 etc) will, it was after that that ATI stopped providing specs. It's adequate for e.g. Flightgear Flight Simulator, Blender, and such -- I don't know if it's up to say Looking Glass. OTOH, the cards are inexpensive.
But I take your main point -- it is hard to find this information in general, and it does keep changing as drivers are developed and refined, and as new hardware comes out.
Users need help choosing their computers
Posted Nov 15, 2006 5:40 UTC (Wed) by shapr (subscriber, #9077) [Link]
I totally agree with you.
I can't figure out which wifi card has oss drivers, the broadcom I bought didn't have the chip I thought it would have and I ended up giving it to a win32 user.
I tried to replace my NVidia card with something, ANYTHING that uses an open source kernel driver and will let me play tremulous in my spare time. But I can't find a decent performance comparison anywhere.
Who's brave enough to start a tom's hardware clone, something like GNUhardware.com? I'd contribute articles!
Users need help choosing their computers
Posted Nov 15, 2006 23:54 UTC (Wed) by JohnNilsson (guest, #41242) [Link]
You could create some kind of utility that connects to Ubuntus hwdb (or equivalent) to compare the users current hardware with known hardware, and suggest replacements for problem parts.
I guess it would be trivial to do a Linux version from the code that already exists. But to be really useful I guess it should be ported to Win32.
Users need help choosing their computers
Posted Nov 16, 2006 6:34 UTC (Thu) by tnoo (subscriber, #20427) [Link]
We desperately need a certificate "Linux ready" for hardware that isfully supported and has vendor-specs available. Some organization (maybe
OSDL) would be the clearinghouse and publish the results of tests. If a
company sells a product that is "Linux ready" they could pay a moderate
fee to advertise with the Linux Logo on their product.
YES! Users need help choosing their computers
Posted Nov 16, 2006 16:11 UTC (Thu) by dwheeler (guest, #1216) [Link]
ABSOLUTELY. I want to buy computers which work well with 100% OSS drivers, and it's incredibly hard to do, even with a lot of knowledge. This needs to be REALLY easy, instead.
Resisting the binary blob
Posted Nov 14, 2006 20:02 UTC (Tue) by landley (guest, #6789) [Link]
Eric Raymond and I spent the weekend trying to finish a paper on this verytopic. We didn't _quite_ get done, but you can find the snapshot at
http://www.catb.org/~esr/writings/world-domination/world-...
(And before you ask, Eric named it...)
Rob
Resisting the binary blob
Posted Nov 15, 2006 0:15 UTC (Wed) by bojan (subscriber, #14302) [Link]
He, he... I like the ending:
> Linspire may in fact be able to solve our multimedia problem. They deserve the community's support and encouragement for trying.
Some plug, eh? ;-)
Free software can't succeed by becoming non-free
Posted Nov 15, 2006 3:07 UTC (Wed) by bignose (subscriber, #40) [Link]
Thanks very much for working on a document that expresses complex opinions. It certainly makes it more of a meaningful discussion.
> Default installations of Linux usually have poor multimedia support, are missing numerous codecs like QuickTime and WMV, and often lack even basic 3D acceleration. Linux can't even play DVDs without introducing the risk of lawsuits, and multimedia support files are usually hosted on non-US sites for legal reasons. Third party software support (from Quicken to World of Warcraft) is almost nonexistent.
These are huge problems indeed. Codecs and data formats that cannot be implemented in free software produce significant pressure away from free software.
If we want free software to be able to deal with these codecs and formats, it is the restrictions applied to those codecs and formats that must be changed.
> You can't win the desktop if you don't even try. Right now, few in the Linux world are seriously trying. And time is running out.
The free software world is well aware of these issues and is actively working to fix them, by working to relax the restrictions currently applied to codecs and data formats.
Turning free software into non-free software is not the way to win.
Free software can't succeed by becoming non-free
Posted Nov 15, 2006 9:00 UTC (Wed) by khim (subscriber, #9252) [Link]
If we want free software to be able to deal with these codecs and formats, it is the restrictions applied to those codecs and formats that must be changed.
It's long-term soluion, yes. Like GNU was long-term solution to unfreeness of Unix.
The free software world is well aware of these issues and is actively working to fix them, by working to relax the restrictions currently applied to codecs and data formats.
Again - that's long-term solution. But... when there were no other way even RMS used proprietary software ("applying the same reasoning that leads to the conclusion that violence in self defense is justified, I concluded that it was legitimate to use a proprietary package when that was crucial for developing a free replacement that would help others stop using the proprietary package")... So it's all is not a black and white... I think that separate product is best short-term solution: after all even Windows XP does not play DVDs by default. If you'll try to put DVD in your new FC6 install and instead of nothing you'll see message "DVDs can not be played by free software so you need to buy Codex pack" (with links further explaining why this "free" format is not included by default) - it'll look much better for end-user. They will grumble, but they'll get the message: it's not that Linux is "incapable" - it's just licensing problem. Right now it looks like millions of linux programmers just can not solve "simple" problem...
Resisting the binary blob
Posted Nov 15, 2006 14:01 UTC (Wed) by danieldk (guest, #27876) [Link]
IMO this essay, and ESR's other recent rants, give to much values to hardware shifts. Quite often, hardware is just an enabler for changing usage paradigms. E.g., concepts like graphical user interfaces and multitasking were pretty old, but the i386 made this affordable for the average user. Because "hardware enablement" often happens later than the invention of a new paradigm, it is easy to get tricked into thinking that hardware dictates the change.
It is better to look at the paradigm changes than hardware enablers. And I think the fundamental change that we are seeing is that "the web is increasingly becoming the computer". The most things that home users do, can be done over through the web. The last barriers are more complex applications like word processing. But it is just a matter of time before those barriers will be torn down. So, what end-users will end up with are powerhouses (multicore, 64-bit machines) that don't do much more interesting than running an operating system and a web browser. This can easily be done on low-power RISC boxes that can be produced for less than $100 (NSLU2 anyone? ;)). The hardware has been around for ages, but the software hasn't. The marginal cost of proprietary software is too high for such boxes.
In this light I urge people not to listen to ESR ;) (of course, I can be wrong). Invest in free software drivers, and avoid closed source drivers and codecs at all cost. Free software drivers are much easier to tune, change and optimize for sub $100 (or any) hardware. Having free drivers available will speed up adoption of GNU/Linux over Windows CE or other proprietary systems for cheap low-power hardware.
(For these reasons I also think Red Hat is betting on the right horse, and Canonical and Novell are betting on the wrong ones. There will be a lot of profit in providing 1. a complete software stack for the server, 2. a completely free system with good development tools for developers, and 3. a simple and light system for clients.)
Small form computers
Posted Nov 16, 2006 23:48 UTC (Thu) by man_ls (guest, #15091) [Link]
I was talking with some friends and colleagues today about the OLPC project and its cute prototypes. It is my firm belief that its most important benefits will be felt, not in the third world but right here in the developed/developing world. A cute toy like that will put in the hands of kids everywhere a complete GNU/Linux system, hackable by design; at that price point ($100 for governments, but $200 for individual purchasers) parents cannot deny it to their kids.Couple that with the thought-provoking bit in the article about the "killing app" and danieldk's interesting comment; and OLPC machines could be that "killing app" themselves, an appliance cheaper than a PDA with a full keyboard and which can connect using wifi; they can even create a network by themselves ("mesh networking"). Teenagers can easily use it to chat or as a videophone. It is something that cannot be done without Linux (or in any case free software) and cannot be delivered via a browser (since it involves delivering a browser); however, it definitely fits in 32 bits, as you imply. There is the problem of playing games, but consoles are better for that anyway -- and cheaper than a grown computer.
64-bit servers à la Google will serve the tools, and Linux is already strong there. Lightweight OLPCs will make use of that, and at the same time provide the free system to develop with, at least initially. Free drivers are a must for that kind of system.
It could be a true revolution, even bigger than the personal computer. The desktop wars could be over for good. These are indeed interesting times, and can get even more interesting.
Resisting the binary blob
Posted Nov 20, 2006 9:36 UTC (Mon) by massimiliano (subscriber, #3048) [Link]
For these reasons I also think Red Hat is betting on the right horse, and Canonical and Novell are betting on the wrong ones.
FYI, Novell has the same "no binary blobs in the kernel" policy that Red Hat has. They only offer the option to use proprietary user space applications, like the Flash plugin, the Acrobat reader, and so on, which are anyway provided separately in OpenSuse.
Resisting the binary blob
Posted Nov 15, 2006 23:17 UTC (Wed) by man_ls (guest, #15091) [Link]
Raymond is not known for his intellectual rigor. This paper is suggestive, but as usual it goes over unwanted facts just to arrive at its conclusion. It is hard and unrewarding to make a list of all the flaws so I will let others do it. Actually the article becomes boring at about half time anyway.In the end it is, as you would expect, an allegate for making temporary concessions to proprietary software. Linux must be burdened with taxes that will pay for "multimedia" licenses. The conclusion is that Linspire is going to save the world; the last paragraph is actually headed "Can Linspire save us?", to which the authors answer affirmatively. I would expect that a company willing to serve a free operating system to the masses should be encouraged to, instead of worrying about legal troubles, just go ahead and distribute the darn things (in this case reverse-engineered codecs), just as Sony did with the VCR or Google has recently done with books. Great companies are known for doing what others thought impossible.
However, if Linspire truly can make the compromises and create more free software in the process, maybe it is worth it. If it only wants to extract money from all those "Aunt Tillies" then I would think that "world domination" is not a desirable outcome. We will see, I guess. Anyway, good luck, Eric and Robert.
Resisting the binary blob
Posted Nov 16, 2006 21:34 UTC (Thu) by landley (guest, #6789) [Link]
It was "hard and unrewarding" to write the darn thing (and I wrotenoticeably more of it than Eric did, and it took about six months). But
we show our work. You're just trolling. :)
By all means, shoot the messenger...
Resisting the binary blob
Posted Nov 16, 2006 21:38 UTC (Thu) by landley (guest, #6789) [Link]
Sorry, end of the day and I'm grumpy. That was too harsh.
(I'm surprised how much flak Eric gets. And how much flak _I_ get when I
co-author something with Eric, that I don't get when I write it on my
own. Even in articles covering exactly the same material...)
Resisting the binary blob
Posted Nov 16, 2006 23:16 UTC (Thu) by man_ls (guest, #15091) [Link]
No problem. Maybe I was trolling a little bit; at least I didn't take the time to support my conclusions, which you certainly did in the article and it shows. Still I firmly believe everything I wrote.As Eric likes to say, been there, done that. I'm still waiting for Eric's answer to the "Terminology wars" critique, my results seemed valid at the time and his certainly weren't.
(I'm surprised how much flak Eric gets. And how much flak _I_ get when I co-author something with Eric [...])I would think that it is the price to pay when you associate with famous, controversial figures: everything you say becomes automatically suspect, it is put under the microscope, and sometimes torn to pieces. I myself think that several of his pieces are brilliant; others are horrible. This one of yours is quite good, but some pieces of the reasoning and the conclusion are too hard to swallow for me.
Resisting the binary blob
Posted Nov 19, 2006 0:33 UTC (Sun) by landley (guest, #6789) [Link]
> I would say that it is the price you pay when you associate with famous,> controversial figures: everything you say becomes automatically suspect,
> it is put under the microscope, and sometimes torn to pieces.
At least they're listening. :)
> I myself think that several of his pieces are brilliant; others are
> horrible. This one of yours is quite good, but some pieces of the
> reasoning and the conclusion are too hard to swallow for me.
It was hard for us to swallow too. That's why we sat on this paper for 3
years. You'll notice I point out the table at the end of Halloween 9?
That's where I first noticed this trend: back in 2003 before Opteron had
shipped, when Hammer had been vaporware for about three years, back when
Intel refused to even acknowledge its existence (going down with Itanic
forever) and Power (g5) was making PC users envious of macs and PPC in
general was getting enough unit volume in game consoles to be almost as
cheap as x86. Saying "a clear 64-bit hardware winner will emerge in 2005"
was quite a stretch. We didn't believe it, but it happened. That winner
is x86-64. This is now retroactively obvious. Intel shipped x86-64,
Apple switched over to it, and Bill Gates announced his retirement.
The same data says a clear software winner will emerge in 2008. It will
also be retroactively obvious.
Resisting the binary blob
Posted Nov 19, 2006 15:06 UTC (Sun) by man_ls (guest, #15091) [Link]
The same data says a clear software winner will emerge in 2008. It will also be retroactively obvious.Your findings for hardware are very well documented and very clear. But not all of your data suggests the same for software. For example, the last software switch on servers (not on the desktop) has happened on these last years (early 2000), when the industry has mostly abandoned proprietary Unix and embraced Windows and Linux. However, the last shift happened around the early 90s when mainframes gave way to those same Unix servers (including *BSD machines). And I see no explanation in your paper on why servers are different than desktop machines. How come Linux and Windows could carve themselves a niche there, and then fight for the dominant positions, between hardware transitions?
Resisting the binary blob
Posted Nov 16, 2006 13:09 UTC (Thu) by wookey (guest, #5501) [Link]
An interesting article. Thank you for taking the trouble to write it. The problem I have is that its fundamental thesis is that there is a genuine deadline to be hit before the possibility of GNU/Linux becoming widespread is lost to us. Everything else about expediency follows from this deadline.
I have to say that I remain skeptical that there really is such a deadline. I could be wrong of course - I don't claim to be a visionary (and have almost no interest in 64-bit computing).
It does also have a US perspective, and the problem of outrageous patent protection and intimidation is somewhat less serious elsewhere.
Certainly media formats are a big problem. I find Debian + the multimedia.debian.org repository makes everything that I need except youtube work. Your average punter expects youtube to work too. Personally I will concentrate on the legal pressure angle to stop re-implemntations of proprietary formats being illegal. I am happy that linspire are taking the 'include proprietary codecs for a fee' route. Many will appreciate it.
Resisting the binary blob
Posted Nov 16, 2006 21:32 UTC (Thu) by landley (guest, #6789) [Link]
If you don't accept the thesis that there is a deadline, then the rest ofthe paper is irrelevant, sure. If we have all the time in the world
there's no need for unpleasant temporary measures now.
Are you suggesting that the curve is wrong and thus x86-64 did _not_
become the new hardware standard as predicted? Are you suggesting that
win-32 will outlast 32-bit hardware when DOS stayed behind on 16-bit?
Are you suggesting that the market will stop having a dominant standard
platform that conveys such signifcant advantages that even something that
sucks as badly as Windows does can remain dominant despite hosting more
viruses than applications?
Pamela Jones raised a good objection: we have a disproportionate number
of developers who preferentially code for Linux no matter what the actual
market incentives are. (Which is true, and a great advantage.
Unfortunately, it hasn't yet meant that when you get a CD free in
specially marked boxes of frosted flakes, it runs on Linux without Wine.
And my fiancee wants to watch videos, and play Sims II which is on a cd
with copy protection I'd need Cedega for, so her desktop is still
Windows. We don't personally care about DVD watching because we have a
PS/2 for that.) I'm sure there will be other good comments.
We could always be wrong about our thesis, we just don't think we are.
How do we solve the DeCSS issue without becoming the dominant platform
_first_, and thus having a zillion warm bodies on our side to push for
changes in the law? (This is not a rhetorical question, if you have an
answer I'd really like to know. We've been trying to come up with an
answer for something like seven years now...)
Rob
Resisting the binary blob
Posted Nov 17, 2006 10:49 UTC (Fri) by khim (subscriber, #9252) [Link]
If you want some factual complains then it's easy: Microsoft does have "64bit system". It's called Windows Vista. You are asking The biggest question about Vista-64 is: will there be one? but it's already exist! Yes, it's mostly 32bit system with some 64bit addons, but... if Windows95 was "good enough" for may years with all this 16bit legacy code then surely Windows Vista will be good enough
Resisting the binary blob
Posted Nov 17, 2006 18:58 UTC (Fri) by Hawke (guest, #6978) [Link]
The copy protection of Sims II is the least of your worries. A bigger one is (the lack of) support for DirectX9 (maybe 9.0c, I dunno) in wine, and even in Cedega and Crossover Office.
Resisting the binary blob
Posted Nov 18, 2006 15:46 UTC (Sat) by kevinbsmith (guest, #4778) [Link]
I would suggest that people won't move to 64-bit software until a) it is ready, and b) it is needed.
As the paper clearly shows, Microsoft is nowhere near releasing a new 64-bit OS, and Apple has no intention of doing so for commodity hardware. If there are no 64-bit OS's available before 2009, it seems odd to claim that one of them must be dominant by 2008. If the only available system is Linux-based, then it would have to win by default if the decision is made at that point. I believe the decision would be postponed.
Does Aunt Tilly need 64-bit power? No. Heck, many of us barely use 32-bit power as we surf the web and browse email. Games and compilers aside, most of us are swimming in far more computing power than we need already. Sure, some "killer app" may come along that forces the move to 64 bits. No sign of one yet.
And speaking of "killer apps", the article deftly avoided mentioning any killer apps of MS Windows, whether 16-bit or 32-bit. (There is no way I can accept "solitaire" as a killer app.) I would argue that the killer app of 16-bit Windows, if there was one, was that it was "pretty" and allowed much more user friendliness. For 32-bit Windows, it was multi-tasking, fewer crashes, and (again) more user friendliness.
What was the last true "killer app" for desktop machines? Email? Web browsers? If I had to put my money on the next one, it would be something related to video...maybe something in the area of mythTV, or videophones, or podcasts.
My conclusions: Our competitor for the next few years will remain 32-bit Microsoft Windows OS's. 2008 is not a magic deadline. As Apple has proven, even a 5-10% market share can be sustained for decades. And therefore: There is no _compelling_ need for unpleasant temporary measures now. Maybe we should pay for codecs, maybe not. But there is no need to abandon ideals wholesale in a panic.
Resisting the binary blob
Posted Nov 19, 2006 1:21 UTC (Sun) by landley (guest, #6789) [Link]
I've been over-aunt-tillied, I find the old bat annoying. :)
No, most users don't need 64 bit power. Most of 'em didn't need more than
the 640k DOS supplied. But people want an OS that can support the
hardware they've got. Want to drive a user nuts? Tell 'em "This box has
a 10 gig hard drive, but you can't use the last 2 gigs." (I saw this
circa 1999 or so. Drove people _nuts_. They didn't _need_ it, but having
it in the box but not usable would taunting them.)
Meanwhile, gamers will find a use for 6 gigs of ram. They always do.
Something like "Spore" or the eventual World of Warcraft upgrade that
allows unrestricted flying mounts (and a level 90 cap by then, probably).
You have to preload a _ton_ of graphics data (with gryphons they can cheat
and preload the stuff in your fixed path :)...
As for killer apps, don't underestimate solitaire (or wysiwyg word
processing). But we never said a new platform _had_ to have a killer app.
most of the really famous killer apps drove hardware purchases (like
Visicalc or Halo did). People are already buying x86-64 machines in bulk
as faster Windows machines, the same way people the first few years worth
of 386 boxes (and the entire 286 production run) as faster DOS machines.
Getting people to move to a new OS that takes better advantage of hardware
they've already got _and_ runs their old software is a much easier sell,
and then once enough people are running the new OS you get new-os-only
software. (Heck, there was 386-only DOS software. All the DOS games you
had to exit Windows to run back in the early 90's still required a 386...)
Nobody's saying Linux will go away, we've just pointed out that we've had
low single digit market share for 15 years now (as has the Macintosh until
the recent surge, and it's about where OS/2 peaked too).
If we want to avoid another 15 years of low single digit market share, we
need to grab something like 30% of the upcoming surge to even get
into "CP/M vs Apple II" territory where a long-term slog can make
progress. (And we've been slogging. Remember how well 802.11b was
supported when 802.11g came out? Or how our current reverse-engineered
support for R300 doesn't even give us 2D on ATI's new stuff? How many
times have Dell/Compaq/IBM announced and discontinued some way to buy
desktop/laptop Linux from them? As far as I can tell Linux actually had
slightly _less_ desktop presence in 2003 than back in 1998, after Loki
collapsed and Red Hat pulled out of the desktop market and so on...)
Or we can continue our annual tradition of announcing "the year of Linux
on the desktop" every year. (After all, the _next_ version of KDE's going
to be even better.) It'll "just happen", one of these days, we don't
actually need a _reason_ for it to...
Rob
Do the Math
Posted Nov 16, 2006 23:03 UTC (Thu) by GreyWizard (guest, #1026) [Link]
An 8-bit processor can address only 256 bytes of memory with a single word, yet the maximum practical memory was 16 kilobytes. A 16-bit processor can address only 16 kilobytes of memory with a single word, yet the maximum practical memory was one megabyte. A 32-bit processor can address only 4 gigabytes of memory with a single word but you insist that a rapid transition to 64-bit machines will happen when 4 gigabytes "falls off the low end of the market." Are you unaware of PAE? Have you not noticed that people have been deploying 32-bit server systems with up to 16 gigabytes for quite some time with good results?
Do you dispute that desktop and laptop systems could do the same or do you admit that your deadline is at least three years too early? Either way, this might be a clue about the credibility of the rest of the document.
Do the Math
Posted Nov 18, 2006 6:41 UTC (Sat) by roelofs (guest, #2599) [Link]
A 16-bit processor can address only 16 kilobytes of memory with a single wordEr...2^16 == 64K. ;-)
Greg
s/16 kilobytes/64 kilobyes/
Posted Dec 3, 2006 20:50 UTC (Sun) by GreyWizard (guest, #1026) [Link]
Yes, 16K was a typo. Well spotted.
Do the Math
Posted Nov 19, 2006 2:21 UTC (Sun) by landley (guest, #6789) [Link]
Your first question (about why the 8 bit machines could access more than256 bytes) was answered in footnote 8. Your second question (about PAE)
was answered in footnote 13. As for question #3, "Will the laptop slow
things down" is the point of the bit at the end of Section 1, first
paragraph under the title "Is 2008 a hard deadline?", the sentence
starting with "At the time, we doubted our own findings..."
I take it you skimmed a bit?
Doing some skimming yourself?
Posted Dec 3, 2006 20:50 UTC (Sun) by GreyWizard (guest, #1026) [Link]
Actually, I read the entire paper, but forgot that you mentioned PAE without credibly addressing the implications. Contrary to the claims in your paper, there is no need to rewrite user space software to get performance benefits in most cases. The kernel manages the memory through virtual memory hardware. (This would not be enough for applications that require more than 3GB in a single process, but if your claim is that this is the important limit then your deadline is too late rather than too early -- and still wrong.)
On the other hand, you apparently failed to read the relatively short comment you responded to. The first sentence about 8-bit computers did not end in a question mark because it was not a question, so there is no need for a footnote to answer it. The rest of your response is difficult to decipher. Since you seem to regard yourself as too cool for punctuation, exactly which sentence in my previous message are you counting as "question #3"? That you once doubted your findings is no excuse for being obviously wrong, even according to your own othewise dubious logic.
Users will decide what matters to them
Posted Nov 14, 2006 20:06 UTC (Tue) by mspevack (subscriber, #36977) [Link]
Corbet,
An interesting article to read.
Once upon a time, the "binary by default" decision was THE LINE IN THE SAND for the vocal majority of Linux users.
Somewhere, that vocal majority has seemed to lose its majority, or its vocalness. As the article says, and as I pointed out in an interview I did on Slashdot a while back, Ubuntu has shipped binary drivers by default for a long time.
But yet they have enjoyed a position of the "darling" distribution for quite a long time now, and Fedora has received much criticism for its continued stance of not shipping those binary drivers by default -- even when making that ideological choice causes some things to "just NOT work".
Not all free as in beer distributions are equally free as in speech. Fedora is quite clear where it stands. In the past week, we've seen a lot out of where other prominent distributions stand.
Everyone should look honestly at reality, and make their decisions based on that.
Users will decide what matters to them
Posted Nov 14, 2006 22:25 UTC (Tue) by wlach (subscriber, #23397) [Link]
In the absence of an official statement from Ubuntu, I think it's a bit premature to cast judgement on this issue. I'm pretty sure we'll be seeing one in the next few days.
Users will decide what matters to them
Posted Nov 16, 2006 22:04 UTC (Thu) by mdz@debian.org (guest, #14112) [Link]
While Ubuntu provides the drivers in question, they are not used by default. Explicit action from the user (generally after digging on the forums or wiki) is required in order to enable them.
Resisting the binary blob
Posted Nov 14, 2006 20:48 UTC (Tue) by AJWM (guest, #15888) [Link]
> Users do not want to hear reasons and excuses for why the operating environment doesn't work with their favorite Web sites
True enough, but they should also consider that it's the web site that is at fault. For example, Flash websites are so much blank space to me (running 64-bit Linux -- although even on Windows I have Flash turned off).
When I encounter a web site that is unreadable because it isn't following W3 standards or includes binary crap like Flash, I sometimes (not often enough) complain about it -- not to the webmaster who will just shrug it off, but to e.g. the VP of marketing, who hopefully cares more about the fact that a potential customer can't read the website than he cares about the technical details as to why or why not.
Ditto with binary drivers. I just bought a 3D card a couple of weeks ago, an ATI-9250 (last for which ATI released specs) based card. I am sorely tempted to photocopy the receipt and mail it to Nvidia, explaining why I didn't even consider an Nvidia card.
One such letter isn't going to mean much to them, but many might. And it's at least complaining to the right people.
Resisting the binary blob
Posted Nov 14, 2006 20:55 UTC (Tue) by gezza (subscriber, #40700) [Link]
I think there is a bit too much religon here. If I have a bit of hardware which needs a binary blob to run, why is it a problem to include it in the driver if it is free to distribute?
Pick a modern wireless card. There can easily be a bit of configuration which although documented, cannot be changed without breaking wireless transmission regulations. Why is it such a problem if the bit which cannot be changed is a binary blob?
Why should it be any different if I store the program for the on board processor on the card in ROM on the card, or in RAM, loaded by the driver at initialisation? The only difference I can see is that the manufacturer can get the product to market quicker using RAM, because he can make the product in parallel with debugging it, and respins are faster. Am I completely wrong?
Gezza
Resisting the binary blob
Posted Nov 14, 2006 20:59 UTC (Tue) by corbet (editor, #1) [Link]
Here's an article from last year on why binary driver blobs are a problem. "Religion" has little to do with it.
Resisting the binary blob
Posted Nov 14, 2006 21:17 UTC (Tue) by bronson (subscriber, #4806) [Link]
That article makes a convincing argument against binary drivers, yes. But the only discussion on binary firmware blobs that I can find is in the comments.
Does our esteemed editor have an opinion about firmware blobs? Say, a tiny chunk of 8-bit code required to get an embedded 8088 working, or an FPGA netlist only synthesizable using $100K+ toolchains? Is source necessary for these blobs as well?
Firmware
Posted Nov 14, 2006 21:24 UTC (Tue) by corbet (editor, #1) [Link]
I've written about firmware fairly often. Source for firmware would be nice, but I don't see a huge difference between firmware which has to be downloaded and firmware which is soldered onto the device. I guess I don't see downloadable firmware blobs as being part of the operating system. It's just part of the magic incantations that so many devices seem to need before they will consent to perform their function.
Firmware
Posted Nov 15, 2006 8:20 UTC (Wed) by ballombe (subscriber, #9523) [Link]
The major problem with firmware is not the lack of source, but the lack of a license properly allowing ditribution and modification, even for the firmwares shipped in the Linux kernel, see http://doolittle.icarus.com/~larry/fwinventory/2.6.17.html for the list of issues in 2.6.17.
Striving for a fully free distribution imply tracking that issue and I only see Debian working on it so far.
Boundary of abstraction
Posted Nov 16, 2006 8:55 UTC (Thu) by cfischer (guest, #3983) [Link]
Very much in line with Jon's comments about firmware, I think that it is really the boundary of abstraction that matters. If the device offers a well documented programmable interface (the same for all OS'es), it is fairly irrelevant whether it needs a blob of firmware to work, or whether that firmware is soldered onto the device.On the other hand, once that line gets blurred - and that is when the OS starts doing some Voodoo magic to make the device operate better, extend its functionality, use some undocumented features etc. - the blob should stop being "just a blob", and be considered OS territory.
I think this defines very well why I don't mind binary firmware downloaded to my Epkowa scanner, but I do object to Winmodems, ATI, Nvidia & friends.
Resisting the binary blob
Posted Nov 14, 2006 22:51 UTC (Tue) by pjones (subscriber, #31722) [Link]
In Fedora, the position that's been followed is that we don't really have any problem with binary firmware blobs, so long as they're freely distributable, and as the terms by which they're distributable are appropriately transferable to our users.We do have problems with non-distributable firmware blobs, and with binary daemons required for operation even of redistributable firmware blobs. You won't see us ripping out e.g. the qlogic firmware from the distro. Such free-as-in-water blobs might get ripped out of the kernel itself, as that's the trend upstream, and we try to track upstream fairly closely. We'd instead be loading those firmware blobs from userland at runtime. That change, if done right, shouldn't detrimentally affect users.
That is, we're very much in the Free-as-in-liberty camp for software, but we're not spinelessly agreeing with Richard while there's a strong ethical position contradicting his viewpoint.
Resisting the binary blob
Posted Nov 15, 2006 17:25 UTC (Wed) by ballombe (subscriber, #9523) [Link]
So let's look at the qlogic firmware in the file linux-source-2.6.18/drivers/scsi/ql1040_fw.h which is licensed by QLogic under the GNU GPL v2 or later.
Without the source of the firmware there is no way users can satisfy the section 3 of the GPL hence they cannot distribute it.
Resisting the binary blob
Posted Nov 16, 2006 20:01 UTC (Thu) by notamisfit (guest, #40886) [Link]
Considering the size constraints of some of these firmwares, who's to say that there even is any human-readable source code?
Resisting the binary blob
Posted Nov 16, 2006 21:12 UTC (Thu) by pjones (subscriber, #31722) [Link]
IANAL, but I don't think you're actually correct, though I myself dislike the reason. The part that's licensed under the GPL is the text string in the source file. Literally:
...
static unsigned short risc_code01[] = {
0x0078, 0x103a, 0x0000, 0x4158, 0x0000, 0x2043, 0x4f50, 0x5952,
0x4947, 0x4854, 0x2031, 0x3939, 0x3520, 0x514c, 0x4f47, 0x4943,
0x2043, 0x4f52, 0x504f, 0x5241, 0x5449, 0x4f4e, 0x2049, 0x5350,
0x3130, 0x3230, 0x2049, 0x2f54, 0x2046, 0x6972, 0x6d77, 0x6172,
0x6520, 0x2056, 0x6572, 0x7369, 0x6f6e, 0x2030, 0x372e, 0x3635,
...
The compiler takes that, builds it, and produces a binary, which is also covered by the GPL. There's no legal constraint that the data from which "0x0078" was generated need be licensed with the GPL, nor any other license.
Resisting the binary blob
Posted Nov 17, 2006 10:58 UTC (Fri) by khim (subscriber, #9252) [Link]
And that's violation of the GPL by itself. Do you have anyone who can claim that "the preffered form" of the risc_code01 "for making modifications to it" is this long hex string ? Truly ? Really ? Sorry - but I can not see any sane developer/expert who'll claim that. Of course somewhere there exist some other more modifyable form of that risc_code01 - and that is the form you can demand from linux kernel developers: GPL allows it. Unfortunatelly the only result of that will be total removal of driver in question from the kernel (it's the only easy thing to do) - so noone tried it. Yet. But the possibility is there: some "evil company" (probably no Microsoft BTW, it's not the only "evil company" in the world) can easily enough force removal of all such drivers...
Resisting the binary blob
Posted Nov 23, 2006 11:00 UTC (Thu) by forthy (guest, #1525) [Link]
There is a clean and legal way to handle binary blobs: Do not link them into vmlinux. The binary blobs are objects of their own right, they are a completely independent work, and have no copyright relation to Linux except that a hexdump of them is part of the source code of Linux.
The open question then is: Is the data, that's part of the Linux kernel legally a "linking", i.e. generating a derived work? There's a grey line like the ndiswrapper issue: Writing an emulation layer in GPL is ok, and loading the original drivers (written for another operating system) is also ok. It's not free software, but it's not a copyright issue, unless the original copyright holder of the blob objects to doing that.
From a technical point of view, the blobs should not be in the kernel. They should be in /var/blobs/... We have initrd to keep this sort of structure around at boot time. People who want only free software on their system should be allowed to choose on install time if they want /var/blobs to be polluted with blobs, or empty; in the latter case, some devices might not work. Similar things e.g. exists for sound fonts on sound cards - MIDI won't work if you don't load the sound font into the card. Nobody is foolish enough to compile something so large as a sound font into the kernel.
So while it's clear that firmware is software, and software should be free, we can't use copyright to force it to become free - the one who writes the firmware can choose. We can choose not to use it, but the fact that the firmware is completely independent of Linux, and Linux communicates with the firmware only through documented or reverse-engineered interfaces, shows it's an independent work. We can do nothing about independent works. But we should make clear that this work is independent by not putting it into files which claim they are released under GPL. And we should avoid tainting GPL software by embedding these blobs right into the binary.
Resisting the binary blob
Posted Nov 14, 2006 21:49 UTC (Tue) by zoop (guest, #14698) [Link]
Minor error: The second sentence of the last paragraph should read:"It is unfortunate that the work which has gone into the creation..."
Virtualization and Reverse Engineering.
Posted Nov 14, 2006 22:31 UTC (Tue) by brugolsky (subscriber, #28) [Link]
Something that might benefit the community is more work on adding reverse-engineering tools to our virtualization environments, since every new machine will have hardware virtualization. There are various patches floating around for QEMU; Xen 3 seems to have taken a temporary step backwards from Xen 2 in that regard.
Reverse engineering is difficult, frustrating, and often boring work; making it easier for novices to contribute raw data might help things along. Over the years there have been quite a few examples where drivers have been reverse-engineered to the point where basic functionality was working, and then the vendor starts contributing to the driver development. One such driver that I currently benefit from is Manfred Spraul's forcedeth driver for the Nvidia ethernet controller; Nvidia has now contributed quite a few patches.
Also, a tool for Windows that does hardware enumeration would be helpful, especially if it can be run from a USB drive; holding the sales clerk at bay while rebooting the display machine into a Linux LiveCD/LiveUSB can be difficult. :-P
this is why gNewSense is important
Posted Nov 14, 2006 23:47 UTC (Tue) by coriordan (guest, #7544) [Link]
I hit this decision a few days ago. I was discussing a mild PR stunt, and it required endorsing a specific distribution for new users. Which one? Should it be gNewSense (the free, blobless distro), or one of the other distros?
The conclusion I came to was that we had to choose gNewSense, even though it would give a bad user experience for some, because picking another distro implies that adding binary blobs and proprietary software makes a distro "better".
That's quite a bad start if you goal is to spread free software in a sustainable way.
this is why gNewSense is important
Posted Nov 16, 2006 10:59 UTC (Thu) by gravious (guest, #7662) [Link]
sounds like a g-nuisance
Resisting the binary blob
Posted Nov 15, 2006 9:38 UTC (Wed) by riteshsarraf (subscriber, #11138) [Link]
Editor's comment================
So criticism of Fedora - or any other distributor which sticks to free
software principles - is, at best, misplaced. There are proprietary
systems out there for people who want to run them, but Linux is about
free software. It makes no sense to try to push proprietary code onto a
distribution which has set a goal of being 100% free, and it is silly to
criticize such a distribution for containing only free software. We
should, instead, be appreciative of the vast amount of work that has gone
into giving us a 100% free system - and help to improve that system.
Question to Editor
==================
If I go with your statement of "Not pushing proprietary code into
linux/linux distribution", Could you or someone responsible explain why
we have such a framework for such drivers ?
Why can't linux "Just Not Support" such binary drivers ?
Resisting the binary blob
Posted Nov 15, 2006 12:00 UTC (Wed) by seyman (subscriber, #1172) [Link]
Could you or someone responsible explain why we have such a framework for such drivers ?
We don't. Said drivers break regularly, which wouldn't happen if there were a framework for them.
Resisting the binary blob
Posted Nov 15, 2006 12:57 UTC (Wed) by arjan (subscriber, #36785) [Link]
there is no framework for doing binary drivers.There is a framework for doing open source drivers, and the binary guys abuse that. I use the word "abuse" because it involves mixing their binary stuff with GPL code from the kernel in order to use the framework for open source drivers...
Resisting the binary blob
Posted Nov 15, 2006 18:48 UTC (Wed) by pointwood (guest, #2814) [Link]
Linux supports more hardware out of the box than any other OS according tohttp://www.kroah.com/log/linux/ols_2006_keynote.html
That is pretty impressive, however that helps little if some of the most
important pieces of hardware isn't properly supported. For desktops (which
today often are laptops), that includes video and wifi.
I do think the situation is getting better though. According the Ubuntu
developer summit article, Ati might release some documentation and there
is an open source Nvidia driver on the way.
I consider binary drivers to be evil and I avoid them as much as I can,
but sometimes it is simply not possible to avoid them. I recently bought a
Thinkpad (z61t to be specific) and I bought that one specifically because
it have Intel video. Kubuntu runs pretty well on this machine, but there
is some things that doesn't work. The card reader and the fingerprint
scanner doesn't work though they get detected. I have reported these
problems of course. In that regard, more info about how to submit a proper
bug report would be nice or maybe a better way to report it. In case it is
a printer that doesn't work (I have that problem too), what info should I
post in the bug report?
I think it is too early to just the Ubuntu team in regards to shipping
binary video drivers in their next release. I think they will do it, but
do it in a resonable way, educating the user about it. If the user gets a
clear warning about every binary driver and the problems the user might
experience because of these drivers, then I consider it less evil. That
seems to be what Linus would like too see too (as seen in a link posted
earlier in this discussion).
In general I have large respect for the Ubuntu developers and Mark
Shuttleworth. He's done a lot of good stuff for open source and his blog
posts are usually very insightful.
We need binary blob distinctions
Posted Nov 15, 2006 21:20 UTC (Wed) by ebiederm (subscriber, #35028) [Link]
I think we need to make some distinctions when this class of problems arediscussed.
1) Hardware drivers.
2) Separate distributed software.
Binary only kernel drivers just are not going to be accepted, and there
is serious work to make that go away.
There is a lack of support for precompiled software from anyone but your distro vendor. The breadth of applications you need to run a general purpose computer requires support from someone beyond your distro vendor to build the softare.
If distros are to be critized I think it should be on that level. Why isn't there a one true linux interface (per architecture) for supporting third party software?
I think that question even when it comes to 3rd parties distributing binary only software is supportable.
Why can't I get a single RPM that supports all x86 desktop and server linux distros. Likewise for x86-64.
I anything is holding back the linux from world domination it is that problem, and I don't even see it discussed. Am I looking in the wrong place?
We need binary blob distinctions
Posted Nov 15, 2006 22:14 UTC (Wed) by khim (subscriber, #9252) [Link]
I anything is holding back the linux from world domination it is that problem, and I don't even see it discussed. Am I looking in the wrong place?
Yes. You should look at LSB - and then understand that you'll be unable to use it. Why ? Because it just does not work. Even in Windows (how many programs work with Windows XP but not with Windows XP SP2? and that's after billions spent on compatibility testing!).
The only way to gurantee something - is to test your package with all supported distributions. If you don't need hard guarantee - then LSB is enough (and you'll get more-or-less the same lottery and in Windows; no way to avoid it).
Resisting the binary blob
Posted Nov 16, 2006 4:49 UTC (Thu) by ssavitzky (guest, #2855) [Link]
I have a completely different gripe with Fedora -- it simply doesn't include the (free) software that I've been using for fifteen years. Debian does. This includes ctwm and xtoolwait, among others. The problem is that RedHat picks a handful of applications that they think people should be using, and those are the ones they package. Every once in a while they decide that people should be using something else, so the Gnome window manager changes yet again.
Debian is more inclusive.
Resisting the binary blob
Posted Nov 16, 2006 14:01 UTC (Thu) by liljencrantz (guest, #28458) [Link]
It seems to me that Fedora Core tries to pick one tool (or a small number of tools) for each task. If you want one of the other gazillion window managers, text editors or scripting languages, use Fedora Extras. If your pet tool is not available in Extras, become a contributor. (I did. Aside from getting my butt handed for being a lousy RPM .spec writer, it was mostly painless)
I think the Fedora model with a small core of suggested components is the right one. And even better, every Fedora release seems to integrate Extras and Core more than the last one. In FC6, you can install packages from Extras during the system install, which is pretty cool. If I understand things correctly, the plan is that in FC7, doing so won't crash the installer most of the time, which is even cooler. ;-)
Resisting the binary blob
Posted Nov 16, 2006 18:08 UTC (Thu) by davej (subscriber, #354) [Link]
Those two apps got into debian because someone cared for them, and did the necessary work to make it happen. No-one has done the same for Fedora extras because, well, no-one cared enough I guess.
There's no reason that anything in debian couldn't also be packaged for Fedora extras. Like Debian, if someone does the work, and it isn't horribly broken in some way, then it gets in.
I think there's a disconnect where a lot of people are still of the opinion that only Red Hat employees can contribute to Fedora. This isn't the case, and it's becoming even less so with every release.
Resisting the binary blob
Posted Nov 17, 2006 21:46 UTC (Fri) by leoc (guest, #39773) [Link]
CTWM - best window manager evar! :)
Resisting the binary blob
Posted Nov 16, 2006 9:32 UTC (Thu) by addw (guest, #1771) [Link]
What is the right/pragmatic thing to do varies over time.
Early days:
Not use proprietary formats => Linux does not gain market share
Little motivation for vendors to release specs that may be used by Linux
Later days (when Linux has good market share):
Not use proprietary formats => non free vendors loose market share
Strong motivation for vendors to release specs that may be used by Linux
The point is that the painful early days need to be gone through to reach the sunny later days. However: pragmatically we need the market share and so early on binary blobs (BB) were more readily accepted - if we did not then use of Linux would be severely curtailed. As Linux has gained more ''marketing muscle'' we are seeing disinclination to use BBs where there was no discussion previously - a nice example of this is firmware in device drivers.
High market share proprietary vendors will try to use their marketing muscle to control the market and opposition to it's own ends; Linux is doing exactly the same thing, now that it is becoming mainstream it can start to control the market, but Linux's idea of control is not about squashing the opposition with secrecy and lawyers but by making a level playing field where the best wins.
(I have used 'Linux' here to mean FLOSS/Free_software/...)
Resisting the binary blob
Posted Nov 18, 2006 4:35 UTC (Sat) by kmw (guest, #38039) [Link]
A better solution, of course, would be for everyone to get over themselves and realize that there is nothing inherently ethically inferior about proprietary software.
I used to be a hardcore Stallmanite, too, until I realized he was a lunatic. While ESR's pragmatic arguments for free software have merit, there is no particular ETHICAL reason to use free over non-free software.
Resisting the binary blob
Posted Nov 19, 2006 19:27 UTC (Sun) by allesfresser (guest, #216) [Link]
Thanks so much for deigning to share your infinite wisdom. We can all go and abandon our principles now, as you apparently have. How wrong we were, how foolish and deluded. Apparently people aren't meant to share and contribute to each other's intellectual achievements after all; hoarding, suspicion, duplicated effort and endless lawsuits are so much better for everyone involved.
Resisting the binary blob
Posted Nov 25, 2006 10:08 UTC (Sat) by mcelrath (guest, #8094) [Link]
"The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man."
-- George Bernard Shaw, Maxims for Revolutionists
Resisting the binary blob
Posted Nov 28, 2006 3:28 UTC (Tue) by josh (subscriber, #17465) [Link]
playing patent-encumbered media, etc.Please don't conflate these kinds of things with proprietary software. Patent-encumbered media formats can still have Free Software implementations, which we should spread far and wide to the extent that we can avoid getting sued in countries that have such patents or feel willing to ignore them. Support exists in Free Software *right now* to play the latest Windows Media Audio and Video formats, most Quicktime formats, MP3, MPEG-4, DivX/XviD, H.264, Flash video format (flv), and many others. Keep in mind that many novice *Windows* users have trouble getting the necessary codecs together to play random multimedia files (and sometimes end-up grabbing massive codec packs from random sites, with added spyware), whereas most FFMPEG-based players will play almost anything you throw at them. While none of this should stop us from pushing for Free formats like Ogg Vorbis, Theora, and FLAC, please don't equate patent-encumbered formats with a need for proprietary software.