RTLinux and GPL-friendly patents

Posted 14 Feb 2000 at 16:55 UTC by raph Share This

LWN reported last week that the basic concept behind RTLinux is now covered by patent 5,995,745, by RTLinux core hacker Victor Yodaiken. The use of this patent will be free for RTLinux users, as well as developers of other GPL operating systems. The actual text of this patent grant is still underway (the developers do, after all, have a lot of other important things to do). Developers of non-free realtime systems will have to pay a license, however.

For a more detailed description of Victor's goals and plans, see his letter reprinted on LWN. This letter describes a license still being worked on with Linux International. Apparently, it will be free for other GPL'ed realtime plugins for Linux, but with some additional restrictions. In particular, you'll have to say whether or not you're compatible with RTLinux, as determined by a supplied regression test.

I've heard that there are a few similar RT systems for Windows NT that use the same technique - essentially having a small and simple RT kernel in control of the system, and having Linux or RT run as a process of this kernel. It's a clever idea, and I hope it goes into widespread use, not only in the traditional RT arenas of flight simulators and embedded controllers, but also to provide completely smooth, synchronized, micro-latency multimedia support on consumer PC platforms. It will be fascinating to see whether these NT-based RT systems postdate the patent and are covered by the claims, and thus require a license.

I'm working on a patent grant for my own stuff which will be significantly less restrictive. Basically, any GPL software automatically gets a license. I'll consider free software on other licenses on a case-by-case basis. I considered giving a grant to all DFSG-compliant software, but decided against it because it may make working around the patents too easy for proprietary companies - simply create an X-licensed libraphspatent.a.

LWN closes with the warning "Software patents are a double-edged weapon at best, and any embracing of them by the free software community is likely to lead to trouble." I agree that software patents require great care to do correctly, but it seems that the RTLinux people are at least aware of the issues. With luck, they will work things out so as to be benificial to RTLinux, to GPL software in general, and to themselves.

What do you think?

raph's patents link, posted 14 Feb 2000 at 18:42 UTC by Snorfle » (Observer)

The link 'stuff' referring to raph's patents is incorrect and should be: http://www.levien.com/patents.html.

Patents link is fixed, posted 14 Feb 2000 at 18:56 UTC by raph » (Master)

Thanks, Snorfle. Yes, I know, Advogato should support post and diary editing. It's on the list.

not certain if other GPLed projects can use it, posted 15 Feb 2000 at 03:59 UTC by atai » (Journeyer)

It is not clear if GPLed projects other than Linux kernel can use the RTLinux patent.

Is any restriction "open"?, posted 15 Feb 2000 at 05:28 UTC by fatjim » (Journeyer)

Is it really in the spirit of helping your fellow man to restrict the patent at all? Simply holding the patent and allowing free use of it seems open enough to me.

I understand that a proper patent represents a lot of work for an inventor/developer; and even a lot of money if it appeals to commercial interests. But my gut tells me that restriction based on openness or closedness of a project is a division of the fellowship of humankind; and thus reduces the freedom and quality of everyone's lives.

(sorry for the extremeist view; I simply haven't seen it stated elsewhere and hoped it might figure more prominently in our minds)

The "are limitations open at all" argument is very fundamental, posted 15 Feb 2000 at 19:18 UTC by Radagast » (Journeyer)

That is, it's at the core of the conflict between copyleft-style licenses and X-style licenses as well. Personally, I'm of the rather selfish opinion that if someone won't make their software available under good terms, why should they be able to utilize my code? I say make them bleed a little. Of course, applying the same system to patents is slightly more fishy, since there's no way to clean room implement, a patent covers the method. Of course, I'm one of those zealots who would like to see proprietary software die screaming, covered in flaming jet fuel, but I digress.

Too vague., posted 15 Feb 2000 at 20:45 UTC by prozac » (Journeyer)

I am not a reader of rtl mailing list, so Victor Yodaiken's letter may be clearer in the context of the list, but taking the letter alone it is not only vague, but disturbing in that it can be taken as possibly fracturing the Free Software movement, in that is appears as if he is advocating that it will be okay to buy Free Software (or a portion of) for use in a proprietary manner.

Please correct me if I am wrong to see it this way.

Is it the patent he means when he states, "If you want to use my idea for a non-Linux or non open project, you should think about how to pay."? Or is it more, like how the patent is implemented in RTLinux? The patent will be, or is, apparently, part of RTLinux.

How much of RTLinux could be "sold off?"

The statement, "If you use a different "realtime" component for Linux you also pay no royalties..." is a bit eerie. If I use a different realtime component for Linux it should have nothing to do with RTLinux. Perhaps he means "... component for RTLinux..."? Even then, it reads like, "If you do not use the "patent" component..." And I did not know that the GPL specified where one must make source available.

I will not be thrilled to send money to such an effort, especially if it is mainly aimed at "collecting fees from people who want to do this on Windows."

Making money off Free Software, always okay, can suddenly take a new direction down the corporatism/capitalism road. Making money from servicing, packaging and documenting Free Software is one thing, but spinning-off patents from the Free Software development process to make money is quite another. It will certainly fracture the community.

Any "Patent-GPL" should "guarantee your freedom to share and change free software--to make sure the software is free for all its users."

I hope that this really is a defensive move and that I am mis-reading the entire thing.

"I would rather be not understood than mis-understood." -- Buckminster Fulller

Interesting, but may be trickier than it looks, posted 16 Feb 2000 at 22:10 UTC by alan » (Master)

At least one IBM lawyer I talked to was sufficiently concerned that the GPL's rather flexible 'no additional restrictions' as well as implicitly granting patent rights you own but are needed to use GPL code for GPL licensed use might effectively put the patent in the public domain.

Other countries?, posted 28 Feb 2000 at 04:37 UTC by jennv » (Journeyer)

What about those of us for whom the GPL is potentially expensive to implement?

Example: A company in - say - Australia writes open source, and produces a company licence which is essentially the GPL run through an Australian lawyer and referring to Australian, not US, laws. Would this be accepted?

What if someone made a Finnish Public Licence? Or a Japanese Public Licence?

The GPL is great if you happen to be in the US, or conveniently able to access the US court system. How many GPL translations will the patent holders be willing to look at before they scream 'enough'?

Jenn V.

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!

Share this page