Discussion on:

OpenOffice: A legal Trojan horse--but for whom?

The article: "OpenOffice: A legal Trojan horse--but for whom?" has much good discussion, but it missed several points, and came to an entirely wrong conclusion.

Missed points:
* "OSRM published a study concluding that Linux may infringe on up to 283 patents." I think that what OSRM meant was that there are 283 patents that Linux might be accused of infringing, causing lawsuits that will be expensive to fight. OSRM is trying to sell a form of insurance, and this study is advertising its product. Still, I think that the main value of OSRM will be its study of prior art to defend against patent suits.

If Microsoft files a patent lawsuit that is related to a Free and Open Source Software (FOSS) program, the patent will have to be identified. Further the first question that will probably ask is a list of all patents that Microsoft thinks the program infringes. At that point I think that OSRM will publicly release its prior art database for those patents, whether or not the defendant has purchased insurance, and, for an insured defendent, will help fight the suit. Meanwhile, GROKLAW and other members of the FOSS community will be searching, organizing, and publicizing prior art, so the most expensive part of fighting a patent suit will be handled free by volunteers.


* "chilling affect on the adoption of OpenOffice" Super-large companies, like IBM, can self-insure. IBM is already protected by the largest patent portfolio in the world, and the most extensive cross-licensing of patents. If desired, it could help out OSRM and mitigate the risk of a Microsoft patent suit that Microsoft would not settle by cross-licensing by purchasing a small policy from OSRM (not enough to bankrupt OSRM and its secondary insurers before IBM had to use funds in addition to its petty cash to settle a suit). Large companies, like Redhat, might want to self-insure some of the risk, depending on volunteers to do the expensive prior art search, and use OSRM to insure the part it does not feel comfortable self-insuring. Medium size and small companys might want insurance with an appropriate size deductible, to trade-off the premium cost with the risk of the deductible. There will always be a few companies, of whatever size, who feel that the cost of continually paying Danegeld to Sun or Microsoft will be more affordable than the cost of risking a lawsuit or paying premiums to OSRM (especially since Microsoft's rates keep going up, while OSRM's premiums might decrease as its prior art database grows).


* "For a desktop Linux platform to even have a shred of hope, ... it will have to be compatible with Microsoft Office." Not exactly. It will have to be compatible with the parts of MS Office that use Microsoft's enforcible patents, and are so important that a significant number of potential users of OpenOffice would not install it. Microsoft would have to refuse to license the patents to distributors and users of OpenOffice to keep them from licensing the patents and distributing or using OpenOffice legally, which would be an obvious antitrust violation. (Of course, if they license the patents, to satisfy the GPL, the distributors could not require users to pay the royalty or refrain from re-distributing, but no indemnification is required.) A patent license that forbid redistribution of OpenOffice without prior agreement to license the patents would also be an anti-trust violation, but less obvious. If Microsoft does sue, it faces the same prior art problem described above, and parts of OpenOffice that appear to violate the patent and appear to be important enough to deter usage of OpenOffice will have plenty of volunteers to delete the offending functions and to write replacements avoiding the patent.


* "any distribution that comes bundled with a fee-based product such as StarOffice would not be freely copy-able or redistributable under the open source license that comes with the rest of the Linux distribution." Not quite. I am not sure, but I think that there are currently distributions of Linux that are bundled with StarOffice. At any rate, the GPL allows for "mere aggregations" of proprietary and GPL'd programs. If the proprietary programs are not based on the GPL'd programs, the distribution is legal. The proprietary programs can even link to LGPL'd programs, and can make normal kernel calls. Of course, part of the cost of the distribution would be the license fee for the proprietary programs, and the user cannot be restricted from re-distributing the GPL'd programs.


* "At this point, if I were Red Hat, and I knew that Microsoft's team now boasts ex-IBM-patent portfolio architect Marshall Phelps -- who could probably prove that OpenOffice infringes on a Microsoft patent or copyright -- and that the provisions of the stand-still agreement pave the way for Microsoft to seek " back pay" on all copies of OpenOffice distributed to date(a copy of OpenOffice is distributed with almost every copy of Linux), I would be worried -- very worried. And if Red Hat has something to worry about, then one can only imagine the pickle that might put IBM in, not to mention a few other hardware and software companies that compete with both Microsoft and Sun and that also rely on the Red Hat ecosystem. Can you name them?"

This is the conclusion that is entirely wrong. Redhat and IBM and the others have nothing to worry about: IBM can take OSRM premiums out of petty cash, and reserve a little more petty cash for the chance that Microsoft would refuse to cross-license a patent that IBM needed. Redhat and the others can afford a premium to OSRM to ensure that the costs of a patent suit will stay below a comfortable value. I buy earthquake insurance with a large deductible to ensure that the cost of an earthquake will be manageable. It is possible that an earthquake would cause so much damage that my costs would far exceed the deductible. However, I don't worry about it, and I don't think that users and distributors of FOSS and hardware with FOSS pre-installed should worry about their self-insured risks.