Should an Open Source Licence Ever Be Patent-Agnostic? (Linux Journal) [LWN.net]

Glyn Moody covers a new license that is being submitted to the Open Source Initiative. "Sharing lies at the heart of free software, and drives much of its incredible efficiency as a development methodology. It means that coders do not have to re-invent the wheel, but can borrow from pre-existing programs. Software patents, despite their name, are about locking down knowledge so that it cannot be shared without permission (and usually payment). But are there ever circumstances when software patents that require payment might be permitted by an open source licence?"

to post comments

Should an Open Source Licence Ever Be Patent-Agnostic? (Linux Journal)

Posted Apr 10, 2009 17:14 UTC (Fri) by tzafrir (subscriber, #11501) [Link] (8 responses)

It is not free software licenses as a whole that tend to be incompatible with patenting limitations. It is copyleft licenses. BSD and MIT licenses have no issues with patents.

The reference implementation might just as well used a BSD license.

Should an Open Source Licence Ever Be Patent-Agnostic? (Linux Journal)

Posted Apr 10, 2009 17:29 UTC (Fri) by yokem_55 (subscriber, #10498) [Link] (2 responses)

Would it be possible to add exceptions to the GPL V3 such that the patent provisions in the license are rendered void? Such a model might guarantee that the code would follow a copyleft model on the relevant copyrights without restricting the ability to license out the patents.

Should an Open Source Licence Ever Be Patent-Agnostic? (Linux Journal)

Posted Apr 10, 2009 17:52 UTC (Fri) by mheily (guest, #27123) [Link] (1 responses)

That sounds rather pointless, as it would probably render the software incompatible with other GPLv3-licensed programs. See here for the FSF's viewpoint.

Besides, this proposed license is for a reference implementation to "promote the extended use of digital media content through increased interoperability". Placing the software under the GPL would prevent it's use in non-GPL software, and defeat the whole purpose.

Should an Open Source Licence Ever Be Patent-Agnostic? (Linux Journal)

Posted Apr 10, 2009 18:11 UTC (Fri) by tzafrir (subscriber, #11501) [Link]

The license they want to use is based on the MPL, which is not unlike the LGPL. Only way more complicated.

Should an Open Source Licence Ever Be Patent-Agnostic? (Linux Journal)

Posted Apr 10, 2009 18:20 UTC (Fri) by jordanb (guest, #45668) [Link] (4 responses)

My understanding is their lawyers believe that the BSD license may carry an implicit patent license. It doesn't mention patents but it does grant a right to 'redistribution;' a parallel patent claim to restrict distribution would seem to be at cross-purposes with that grant.

They want to be absolutely sure they can profiteer through patent licensing while still being able to stick the OSI's increasingly meaningless seal of approval on their software.

This means having a license that *explicitly* declines to grant a patent license.

Should an Open Source Licence Ever Be Patent-Agnostic? (Linux Journal)

Posted Apr 10, 2009 19:35 UTC (Fri) by drag (guest, #31333) [Link] (1 responses)

Remember the Fluendo people are shipping MIT licensed Mp3 that is has patents hanging over it that Fluendo and Novell paid for redistribution rights for, which doesn't extend to anybody else.

So it seems that they feel that MIT fits the bill for patent restricted but copyright open.

Should an Open Source Licence Ever Be Patent-Agnostic? (Linux Journal)

Posted Apr 13, 2009 9:16 UTC (Mon) by AlexHudson (guest, #41828) [Link]

It only obviously works if you're not the patent holder, though. If you're a licensor it really doesn't matter too much what's in your license, you can't automatically give rights to anyone.

This is by design and it's GOOD thing

Posted Apr 10, 2009 20:20 UTC (Fri) by khim (subscriber, #9252) [Link]

If you own the patents in question then you must grant patent license as well. If some other party own the patents - then you are not obliged to fight with them. Looks like a good compromise to me if you genuinely want to popularize something.

implicit grant doesn't work everywhere

Posted Apr 12, 2009 18:58 UTC (Sun) by coriordan (guest, #7544) [Link]

The implicit grant you mention in your first paragraph was debated during the GPLv3 drafting. IIRC it probably does work in the USA but probably doesn't work in the UK.

Should an Open Source Licence Ever Be Patent-Agnostic? (Linux Journal)

Posted Apr 10, 2009 21:00 UTC (Fri) by rickmoen (subscriber, #6943) [Link] (1 responses)

I think Glyn's question muddles a key distinction: A codebase merely being under an open-source licence doesn't suffice, by itself, to make the code open source. Glyn seems to be assuming that it should.

However, that has never been the case. For example, if I claim that a piece of my code is BSD-licensed but never get around to releasing matching source code, then it remains by definition proprietary despite the attached licence. Similarly, any code for which patents turn out to preclude exercise of rights necessary to open source is inherently proprietary within that patent's jurisdiction -- no matter what the licence says. (So, for example, Apache httpd mod_ssl code was open source in Europe during the RSA patent's term, but proprietary in the USA -- because the right to freely use and redistribute was not available in the latter.)

Licences don't need to guarantee patent rights: Despite brave talk about "implicit patent licences" in some permissive licences, they never have guaranteed such rights.

Rick Moen
rick@linuxmafia.com

Should an Open Source Licence Ever Be Patent-Agnostic? (Linux Journal)

Posted Apr 11, 2009 21:24 UTC (Sat) by OLPC (guest, #47981) [Link]

Here's what a lawyer told me (who specializes in intellectual property law). As always, talk to *your* own lawyer for advice.

*If* you hold a patent, *and* you release code that practices that invention for a particular purpose under a software license, you are granting an implicit patent license to practice that invention for that purpose. The further from the stated purpose of the code, the less the implication. He said there is quite a lot of case law on the topic.

Note that if releasing code that practices the invention when someone else owns a patent in no way affects the patent holder's rights: this is the Fluendo case, and a paid up license exists (with some limitations) for that patent. There is no contradiction between the software license and the patent license.... The two are not coupled (unless the software license talks about patents, and you can only talk about patents you own...).

Why expect anything else?

Posted Apr 18, 2009 22:56 UTC (Sat) by pboddie (guest, #50784) [Link]

This supposedly "open" licence is exactly the kind of thing one would expect from a bunch of organisations who each believe that their own "intellectual property" is really valuable, whereas everyone else's is worthless. That people can develop software on their own clearly unsettles such organisations, and maintaining a monopoly and the threat of sanctions is presumably the only way such organisations feel they can compete and still maintain control of an area they regard as their exclusive property. Now that it's fashionable to cut back on research and development, all too readily blaming the financial crisis (instead of, say, badly designed products), I suppose the business plan involves doing none of the actual work and just collecting cheques and retaining lawyers.

That it's a patent cartel involved in this specific case shouldn't surprise anyone in the slightest. Still, I imagine various apologists will regard such a licence as a "win", although they'll obviously be on the "open source" (as opposed to the Free Software) side of the fence.