Slashdot Mirror
Red Hat's Secret Patent Deal
Bruce Perens writes "When patent troll Acacia sued Red Hat in 2007, it ended with a bang: Acacia's patents were invalidated by the court, and all software developers, open-source or not, had one less legal risk to cope with. So, why is the outcome of Red Hat's next tangle with Acacia being kept secret, and how is a Texas court helping to keep it that way? Could the outcome have placed Red Hat in violation of the open-source licenses on its own product?"
Copying the first paragraph of TFA, verbatim, does not make a helpful summary.
My Operating Systems instructor (a die hard open source / linux guy) went to testify as an expert at a court case involving patent infringement with some pretty big names involved in this now famous East Texas District court. He was skeptical going out but came back with a different opinion; apparently that court district processes so many IP and software patent claims that the court district that they've become better at dealing with it than any other court. They process the claims faster and the judges are more knowledgeable on the subject than most and run a very tight ship. He said it's easy to see why they might have started those types of cases out there, but the reality now is far different than the picture painted.
How does it feel to be a liar with pants constantly on fire?
IANAL, however it seems that changing the LGPL to specificly deny sealed cases involving patent infringement settlements with out notification of all developers, to be a prudent response. If there is a settlement involving an LGPL'ed software then all developers need to know so that the infringing code can be change, removed, amended, or licensed.
When dealing with a commercial software house there is only one company that needs to license the code and only one set of developers who are aware of the licensing via the
corporate structure. This paradigm is quite different in the opensource community, this is the equivalent of the company having a ruling/ agreement but being bard from informing the developer staff.
IAAL (in fact, IAAPL, I Am A Patent Litigator).
1) Bruce Perens knows nothing of the law. The case was not sealed. You know that because you can use PACER to go on the EDTX docket and look at the case (Civil Action No. 6:09-cv-00097-LED-JDL). If the case itself was sealed, you could not do that. Only really exceptional cases get sealed (i.e. national security cases, some mafia cases). Whether or not the case was sealed has nothing to do with whether the settlement will be publicly available. Nearly all patent cases end with a confidential settlement agreement. Generally, with limited exceptions (e.g. class action cases), there is no requirement to put a settlement of any civil case before the Court or to make it public. The parties simply file a stipulation to dismiss, as was done in this case. Furthermore, Mr. Perens confuses a protective order (which is often a permanant order by the Court not to disclose or use confidential material produced during discovery) with a confidential settlement agreement (which is a contract between the parties, typically enforced by a subsequent breach of contract lawsuit, rather than contempt).
2) The LGPL provision at issue (presumably LGPL 2.1 Paragraph 11) is HORRIBLY DRAFTED. In fact, it is not clear what at all it is meant to do. What does it mean that a judgment or "allegation of patent infringement" "contradict[s] the conditions of this License"? Whether or not RedHat takes a license, a patentee could sue another JBoss user for infringement. Just because the patentee could not sue RedHat again or sue RedHat's paying customers again does not seem to necessarily contradict the conditions of the license. RedHat could not guarantee before or after that 3rd parties would not be sued by the patentee. RedHat is not the one that is impinging on the downstream use of the LGPLed software, the patentee is, so how could RedHat be in breach of the license? The whole section is non-nonsensically meaningless and demonstrate a complete misunderstanding of how patent litigation and licenses work. Sure, if a patent license required Redhat to not distribute the source code to LGPL licensed materials, that would be a clear breach. But that's not how the license works. The license simply would not guarantee that the patentee wouldn't sue individuals who then took that source code and sold a product based on it. No "contradict[ion]" because RedHat is doing nothing to impinge the downstream re-sellers (the patentee is) and the downstream re-sellers are in no worse condition than before the agreement that RedHat entered into.
3) It does not appear from the JBoss documentation that JBoss mentions any particular version of the LGPL. Therefore, under the LGPL, you can choose any verison. LGPL v3 does not have the equivalent non-nonsensical provision.