Slashdot Mirror
Real Time Linux, Now Patented
This week's Linux Weekly News is reporting that
Victor Yodaiken, the developer of RT-Linux (Real Time linux) has been granted
a patent
on method used by RT-linux. He intends that
Linux users be granted
a no royalty license, users of closed OS' may have to. It's unclear whether
Hurd or *BSD would be granted a royalty-free license. While this could be
heralded as the beginning of a new
defensive patent trust
for free software, it also jars somewhat with the
hacker ethic. What do you think? Is
Victor's idea one which is technically original, and which would not have
been published had the author not had the protection afforded by a patent?
Was RT-Linux's status as prior art not sufficient to keep RT-Linux free from
other patent claims?
The preamble to the GPL isn't the GPL. And so isn't part of the licensing terms. Not that I disagree with the sentiment, but that particular pseudo-legalism doesn't help.
--
Normally I have to say I'm against most software patents (really any overly broad patent) but I know Dr. Yodaiken and if anybody can be trusted to do right by the community with this it's him. He's not only one of the coolest professors I've known he's one of the coolest people I've ever met. I wasn't his best friend or anything so I may very well be wrong but I have to think that he won't be somebody we need to flame over his patent portfolio.
As for the content of the patent, running an OS as a process of another OS is nothing new, and making the host OS a real-time OS is, IMAO, an obvious extension of that concept.
--
I don't disagree with the idea of free software developers acquiring software patents. But they should allow other free software developers to use these patents without royalties---regardless of the operating system platform.
That is to say, if someone creates free software under a GPL-like license which prohibits proprietization, then let them use the patented idea!
On the other hand, I don't see a problem with makers of proprietary software being made to pay royalties. After all, they don't like freely sharing their ideas with others, so why freely share with them?
I hope Yodaiken takes on a more enlightened attitude, and users the power inherent in his patent to do good. The main purpose should be to prevent a hostile organization from acquiring the patent, not to act as a hostile organization.
I have a feeling that Yodaiken will probably grant use of his idea to anyone developing GNU style free software. but they shouldn't have to ask, know what I mean?
On the other hand, I think it's great that the Open Source community has finally cottened on to the fact that the commercial sector has zero ethics and even less morality.
Does that mean the patent is "good"? "Bad"? Indifferent? Purple? IMHO, the only way to find that out is to talk about it, and that's something that a lot of people have either hedged around, ummed and aaahed, or ignored completely. Well, now that can't happen.
The actual conclusion, though, on whether RT-Linux should be patented or not is not the important part of the discussion. That's a side-show, as good as RT-Linux is. What's important is that we talk about how to keep software Open and Free, whilst preventing abuses -of- that Openness and Freeness, by people who neither understand nor care for anything but their bank balance.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
The GPL does state that any patented code you use must be licensed free for everyone's use. So I'm afraid this does violate the GPL unless the author licenses the code for free to everyone.
No. You don't have to defend a patent in the same way that you have to defend a trademark.
"The invisible and the non-existent look very much alike." -- Delos B. McKown
I also dispute that the claims of this patent are non-obvious. I proposed a similar scheme when I worked at Telebit and we wanted to merge our router and modem code bases for a low-cost product. The modem had hard real-time requirements, so we were going to run the router operating system as a task under the modem's operating system. Neither I nor my coworkers thought this was a super-advanced non-obvious technique that should be patented.
As it turned out, I came up with an even better method for building the product, but we didn't end up using it because the price of modem chipsets had dropped enough to make it unnecessary.
Imagine the following scenario:
Now, had FSP patented X, they could have defended themselves agains EvilCorp.
I believe that we need to do this, but establish a holdings company to keep the patents and then make sure than any group which licenses it's product under X, Y and Z licenses can automatically have access to the patents... which neatly ties up the problem of other Free Software presently available and future ones yet to be created having access to such patents.
While the patent is probably bogus, your message is meaningless, because you include the description, and what matters for patent purposes are the claims.
Patents are a paradox in computers. And why? Because they try to cope an ancient "mechanical" view of enginnering with an quantum-artistic world of
computers. A long time ago it was easy to make a patent. Classic mechanics are, by nature, quite deterministic. And so every appliance of its laws is quite
straightforward. Besides Mechanics still lays in the capacity of humans to build instruments, mechanisms. Most of these things are palpable, some may require
a microscope or something similar to be "seen". However a careful examination can always determine how the "mechanism" was made. And this was the
fundament of the patent system. In one way or the other you can determine of one or other mechanism "copies" a patented one.
This does not happen anymore with computers. Yes, you still can determine something on them. But you most of the time can fall in two problems. Two
problems that make the classical patenting systems quite awkward in the Wonderful New World.
First most of the computer "mechanisms" are not palpable or either visible. Most of them are "abstractions" of the behaviour of quantum laws in a relatively
defined environment. Such environment can be a processor, a computer, a network or Internet. No matter the initial conceptions to create a functional structure for
a specifical purpose, such creation has a high potential to overcome the expectations of its creator. You cannot always predict on computers where your idea
may lead to. It is a problem that Intel had to deal with in the time when i286 came up. Protected mode was an "idea in the air" that was left on the 286 for a future.
Most aspects of its functionality was never expected to be used in the way that later people used it. When it was occasionaly leaked out to the public, it created
a different series of protected mode systems. Every of them depended on a inner nature of the processor. Some followed the "original idea" Intel had to create
protected mode. But others went so much to the bottom that they exploited every single feature of the processor in this mode. Exactly in one of these "strange
modes" one friend of mine used 286 for real-time control of system for observations of electro-magnetic ressonance.
Intel tried to stop this thing and turn the "tides" back to where they thought things should go. However we have to note that some of the functionality of modern
Intel processors carries somehow the remains of the ancient Flood...
The second problem comes from the first and it is what puts in cause the use of patenting system. For most of computer uses, you have to use an "abstraction".
Something that translates your ideas to the computer world. That is what for computer languages are used. But as every "abstraction" you loose some of the
determinism innerent to mechanics. In one way or the other, the computer languages determine also an "art of programming". And so it is natural that they
become abstract, that one "style" copies "another", that you may find two similar ideas in two different programs. More than this, as time goes, programs start to
cite "old sources", programs start to embed over each other. Sometimes plagiates start to happen. Good or bad. Conscious or occasional.
How one can determine what is correct or wrong in such world? Well somehow a programm is still a piece of text in most cases. So one may determine, u to
some extent, if anyone violates the copyrights of the original author. But can this be used for patents? Certainly not. You cannot determine the full extent of the
mechanism. Or if try to determine it then you cannot fully acknowledge its potential application. It would be the same as trying to jump over Heisenberg's Law.
You may try to find either the position or the velocity of the electron. But you can't know both at the same time.
1. Noone needs to fake a "general purpose" OS if you have a real-time OS. A real-time OS has all the features of a "general" one and more. This is exactly like running DOS apps in Windows 2000, you can, but WHY?
2. Didn't I learn about this method in college only for the general not linux specific case? To quote Tweety bird: "I did! I did!". Just becasue it's done with Linux doesn't make it new. Any company that has the interest will be able to challenge this patent, and get it tossed. But again, noone will care, becasue there is no need to do this.
3. The problem with the patent system is that they are granting completely bogus patents on things in textbooks. When someone invents something new, they deserve a patent. Invents new, not discovers old. Everything in computer science was done in the 70, get over it already.
4. This patent is no better then any other. Many people will not be able to use the method without paying the grantie lots of money. The only acceptable "defence" patent license is to place the method into the public domain. Anything else is no better then what every other patent holder does.
- Adam L. Beberg - The Cosm Project - http://www.mithral.com/
I wonder if IBM knows about this. The VM operating system has been doing this EXACTLY for over 30 years.
VM consists of two parts -- CP (Control Program) is a virtual machine emulator that runs on the bare iron (or can run under another copy of VM).
Other general purpose operating systems, such as MVS or CMS, run as "virtual machines" under the real-time CP.
Among other things, CP intercepts the hardware interrupts, and emulates them to the guest operating system, EXACTLY as described in the patent.
CP also allows the guest operating systems to use the regular hardware instructions to enter a "virtual interrupt-disabled" state so that the guest operating systems can safely proceed as if interrupts were actually disabled, while the actual hardware leaves the interrupts enabled, to allow continual real-time interrupts/data collection.
Apparently the patent examiners had no awareness of this VERY long-standing prior art.
- John
BSD can be repackaged as close source. Linux cannot. If the patent usage licence resembles GPL there is no such problem. The question is that... Yes... Yet Another License...
Baker's Law: Misery no longer loves company. Nowadays it insists on it
http://www.sigsegv.cx/
UNISYS, too, stated that some software could use the GIF patent free and later changed their mind. Promises and intents are just that; what matters for patent licensing is a contractual relationship between the licensor and the licensee, and merely posting one's intent on some site isn't sufficient. Also, the particular license chosen for RT-Linux looks like it imposes some serious burdens on other real time Linux developers. Besides a contractual relationship, the only other choice for patents is placing it in the public domain.
The only choice I see for ensuring that open source-related patents remain credible and open is to assign them to an organization whose charter irrevocably binds them to particular licensing conditions.
As for the specific patent, there is ample, specific prior art.
I'd recommend to the author to abandon this patent and work with the open source community on a better approach for the next patent.
Hey ya jagoff, why the hell can't I use your code with my BSD-flavor-of-choice?! Fear the Daemon, he's so much more righteous than Tux!
Perhaps I should have said "politely" rather than carefully.
I'll admit to not having followed LinuxRT from conception to present, so I don't know the origin of this code, but if he wrote it all on his own, then that code is his to license or patent however he chooses within the bounds of the law. If he wants to patent his invention and license it free of charge to GPL'ed code and tell people using the BSD license to pound sand - as much as I would disagree with him - that's his choice to make. The BSD people would probably come up with their own stuff, be more friendly in the licensing, and the world would probably follow.
D
2.) I think, if approached carefully, that Hurd/BSD users would probably have little difficulty getting "free use" clauses.
3.) While I understand that in some things, you have to "rush forward" to patent something, I think now is an excellent time to revisit an idea that was talked about once before, that being a "GPP" or General Public Patent... some structure that patent-holders can use as a boilerplate for releasing their patents to any and all "Free"/"Open" use.
I don't have moderator points right now, or I'd spend a good number of them on this thread. I think this discussion is exactly what needs to be held. So I guess I'll just post.
The point is brought up that a patent license could be used just like the GPL uses copyright. And that's true. And it may even be what to work for.
But there's two points to consider:
1) A patent is broader than copyright -- and getting broader (if one can patent things like
build-to-order manufacturing and collecting data in a cookie). Thus, a patent on a widget might extend to ALL widgets that provide similar services. This has been brought up, but it's worth considering, because the point has synergy with my second point.
2) AFAIK, there's no currently accepted open patent license. And, in fact, there's no completely accepted open source license, either. The GPL is close, but even that one isn't universally embraced. But that's OK, because if you want to use someone else's code for the Whizzy Widget, hey, you do it by the license they released. If you don't, you can go write your own Whizzy Widget No problem. Release it under any license you want. HOWEVER, if Whizzy Widgets are patented under a license you find unacceptable, you're out of luck. Write your own, and you face patent infringement.
I think it may be possible to write a license that would get around such problems. But until there's a widely accepted (and recognized as enlightened and effective) open patent scheme, there's gonna be some serious clashes, and probable abuses.
Weston
Tweet, tweet.
Yes, patents aren't free, unlike copyrights. (Nor are they automatic.)
/. and this message... What are you doing wrt patents? Is the FSF working on a GPL-type license for this?
But, the upfront costs are usually the low costs. The big $$ comes in when you defend the patent (or use it offensively.)
And, if you (author) used a product that had a GPL-ish licensed patent, it wouldn't force you to patent something, it'd just require that any patented procedure in that program was also licensed in the same way.
And yeah, the RT patent could be worth a few bucks. I haven't looked at it so I don't know how obvious it is. If it's a 'good' patent then he deserves a few bucks. And a company that wants to use RT Linux for a specific application could probably afford to shell out a few $$ for the privellage.
Where I'm not happy with this is that the GPL doesn't forbid patent encumberance, or bundling the GPLed code and patented algorithms (non distributable and non freely usuable because of the patent). IMHO if you use GPLed code, it should be mandatory that it leave your hands as unencumbered as when you got it.
(By which I mean, if RSA released the an RSA module which they GPLed, it would be patent encumbered, but if they created it from scratch, that would be fine. But if they took the existing network code or GPG code and made it dependant on patented algorithms, they should be in effect granting free use of that patent in any way, as might be derived from that code.) (ie, if someone took that theoretical GPG code and turned it into an encrypted file system, that should be allowed, the same as using any GPL code in your code, no matter the insignificance, turns your code GPL...)
Hey RMS, if you read
Patents are a defensive measure, which really, is what the GPL is. If we could trust everyone to look out for the best interests of the community, or to at least not hurt the community, the GPL would be pointless, people would keep code open simply for the benefits to everyone. We've seen though that this doesn't happen, something like the GPL is needed to keep people 'honest'.
We've seen companies use patents in incredibly hostile ways. Amazon's 1-click patent, that company with the patent on selling any digital information over a network, etc. And they immediately lashed out with these, seeking exorbitant licensing fees or to put competitors out of business.
We need a GPL-type license for patents, something where any company using the GPLed patent in any product has to similarly GPL all the code and patents in the product.
Then a company could patent something, to keep anyone else from using it against them, but by releasing it under the GPL, prevent any fears that they're trying to dominate the industry.
Note though that the article said that Linux users would be granted a royalty free license... nothing about home users, or free users, etc. If this is applied to any open source OS, then it's good in my opinion.
But I would like to see RMS and the FSF come up with a GPL-like license for use with patents.
If there was a standard GPL-type license, and it only forced other people to GPL their patents and code, then it would have the same effect as the GPL does on code, and couldn't be applied selectively. But commercial users who didn't like that provision could still negotiate other licenses. The best of both worlds.
But the GPL isn't a free license (as in BSD). The GPL has restrictions... A patent will be that free as long as the only restrictions are GPL-ish, as in derivative patents and copyrighted materials have to carry a similar license.
The GPL is free in that it makes no restrictions as to who can use the material, or for what. A company is just as free to sell GPLed code as an end user is to use it (provided they follow the free distro rules, etc.) As long as the patents are similar, such that any use in an open-source, open-patent project is allowed, then I fully support it.
I would like to see a standard license developed for this sort of thing. Maybe the FSF can develop one.
Side note: I see patents being used to protect your business against competitors, not against anyone using the idea. As such, eventually having all patents free (as in, companies decide that being able to use the GPLed patent portfolio is more useful than keeping their patents secret) would still serve to prevent a hostile company from stealing research, because it would have to open all of its patents to use the free ones, but it would serve to advance knowledge more rapidly, ala open source.
If 'we' (the OpenSource Community) are going to have to put up with nonsense patents (LZW compression and the Gif format come to mind) then I think this is something I'm all for. Until we don't have to deal with these problems, then I think we have to take a pro-active (shudder... i think I found that word from being around management too long) view and build up our own 'patent portfolio' that we can use defensively if need be.
:)
A better question might be "Who will control the patent?" In order to really have a defensive patent portfolio, don't we need one orginization/group who can barter with them on behalf of the collective whole? (whether they control the individual patents or not)
Now... if only a piece of software would come out of Unisys that we could Sue over
This space for rent. All reasonable inquiries will be entertained at proprietors discretion.
Linus doesn't hold patents on Linux. Secondly, the concept of "protecting" Linux via software patents is more cockeyed than a three legged dolphin.
You don't need a patent in order to 'defend' against patent lawsuits. All you have to do is properly document (date,witness,etc) your invention and you have a very strong claim against anyone who tries to patent the same invention in the future.
I'm not a lawyer, but I'm guessing that you're not one either, given your impression that having documented prior art is a "very strong claim". Unfortunately, that isn't the case, as a patent has a presumed to be valid, and you have to prove otheriwse. A competing, previously issued patent is a very strong claim. Documented prior art is not, at least it doesn't seem to be, from the outside of the legal profession looking in.
When (not if, but when) some 800 pound gorilla decides to extract some money from Open Source (distributors, corporate users, developers, etc...) based on a software patent, the best defense we may have is public sentiment. By showing the world how this closed-source company is extorting money, we might be able to put them in such a bad light that they choose to not continue the suit. Open Source currently enjoys much positive sentiment on the part of the general public (to the extent they have even heard of it).
If members of the Open Source community have patents that are being used against closed source companies, it will be hard not to look disingenuous when we argue that they shouldn't use their patents against us. It will be pointed out that we stand by our convictions only when it is convenient to us, and we will look worse than the 800 pound gorilla. Because the patent situation is one of the few Achilles' heels that Open Source really has, I think this is a potentially dangerous development.
Put me down as "against it."
-Steve
Democracy is a poor substitute for liberty.
Copyleft is an application of Copyright to subvert copyright itself.
Could this be an "Open"Patent...an application of the patent system intended to subvert the patent? Essentially patenting it with the express permission for anybody to use/modify/distribute it...as the Copyleft copyrights something with the express permission for anybody to use/modify/distribute it.
Jazilla.org - the Java Mozilla
It's 10 PM. Do you know if you're un-American?
Make that a really, really freaking HUGE stick, I mean omigod that's a really huge stick. Where in the heck did they get that stick? It's more like a tree than a stick! (-1, offtopic)
However, I think this is A Good Thing. Not because I like patents on software, but, sorry folks, software patents exist. Be in the now. Until the wonderful day when we can strike down software patents, we must arm ourselves to fight on their terms. By patenting RTLinux, and forcing non-OSS (non-free software too, in deference to RMS) to pay for it, we gain a weapon we can use in the eventual fight to end software patents.
www.eFax.com are spammers
if this is so obvious, wheres your realtime implementation of linux?
I just think that no matter who, no matter what, no matter how, no matter why, software patents are simply a Really Bad Thing for programmers.
-=-=-=-=-
-=-=-=-=-
My mom's going to kick you in the face!
You don't need a patent in order to 'defend' against patent lawsuits. All you have to do is properly document (date,witness,etc) your invention and you have a very strong claim against anyone who tries to patent the same invention in the future. The fact that he took out a patent shows an intention to license it, or to obtain monopoly rights.
Funny: Intel did this with Windows 3.1 quite a few years ago. It provided a real-time operating system kernel which ran beneath Windows and treated Windows as a task. It demonstrated it at the Embedded Systems Conference for a few years running and licensed it to embedded systems vendors.
IBM's VM/CMS is also prior art and is even older.
--Brett Glass
--Brett Glass
The abstract:
A general purpose computer operating system is run using a real time operating system. A real time operating system is provided for running real time tasks. A general purpose operating system is provided as one of the real time tasks. The general purpose operating system is preempted as needed for the real time tasks and is prevented from blocking preemption of the non-real time tasks.
This looks a lot like a real-time Mach, or for that matter, a lot like QNX. Not that microkernel=realtime, but isn't that the basic idea behind a lot of RTOSs?
Don't Panic...
Gee, you think the GPL is viral. Patents infect people who don't even see the code.
On one hand, if a sufficiently large block of "Free Patents" were available then perhaps we could make the same kind of patent trades that the big companies do and get access to, say Unisys' patent on GIF, for example.
On the other hand, what corporation is going to deal with a bunch of lawless programmers who would copy "trade secrets" all over the world.
On the gripping hand, if they need the patent, they gotta come to us. Heh heh heh.
Anomalous: inconsistent with or deviating from what is usual, normal, or expected
Anomalous: deviating from what is usual, normal, or expected
Canard: a false or unfounded repor
This is so, so wrong- just _listen_ to how it sounds. "Probably, maybe, other people can use this GPLed software, if they ask _nicely_ and he's not having a bad day- he probably doesn't have any reason to forbid their access to the GPLed software so it shouldn't be much of a problem..." I feel _ill_ :P isn't this exactly what's supposed to never happen? I want to see this situation nipped in the bud, immediately. This is the license I use for my software and I gave no permission for it to be rewritten in this manner...
The ramifications for BSD licensed code, (what's intended anyway), is that a patent in Pool F could be incorporated into that code and (usually) used freely, up until the code was made into a proprietary product. At that point someone would have to negotiate a separate patent license, as the OPL would no longer apply.
I've talked with a Patent attorney about the license in general, and found to my relief that the things I want it to do are in fact doable. I haven't asked him to go over the license in detail yet, as I don't see much point in spending money on the legal debugging when the license itself hasn't fully settled down. Once it seems that the license probably does what most parties would want it to do, or at least their goals are incorporated into it in such a way that a lawyer rewrite the wording properly, I'll take it to the lawyers again and go through a few rounds of legal debugging/general debugging.
If anyone is interested in discussing the topic in depth, there's a mailing list on the site.
if the purpose of a patent is to encourage people to release details about their creation in exchange for protection from people stealign their idea, and thus advance technology in general, than whats the point of patenting somethign I can just go get the source for? (rtlinux is open source, I am imagining)
seems to me that the only real reason to do this is to defend against someone else patenting it, but it exists as open sourced prior art, so that probably wouldn't (or at least shouldnt) fly.
I don't get this at all.
If hackers all around the world started doing this, there would suddenly be no Open-Source anymore! I know I sound a bit like RMS, but you can't have "exceptions" in open source, and to have one "insider" doing it is a bad way to start.
In my opinion, all open source hackers, and people committed to developing open, free software (as in speech, not beer), should also commit themselves that, if they're developing something for the community, to completely release it under the GPL.
If we don't, developers will start like: "This software is free for end-users, home-users... etc, but not for commercial users." How exactly do you define a commercial user? Worse, how exactly do you enforce such a license? The simple, better-for-all way is simply to GPL it.
Finally, any free program is threatened constantly by software patents. We wish to avoid the danger that redistributors of a free program will individually obtain patent licenses, in effect making the program proprietary. To prevent this, we have made it clear that any patent must be licensed for everyone's free use or not licensed at all.
Remember, no matter where you go, there you are.