Email Sender ID: Exactly what’s wrong with the license?

Larry Rosen, the attorney who has been negotiating with Microsoft on behalf of the OSI for changes to the restrictive patent licensing attached to Sender ID, was asked recently to explain in a concise manner exactly what it is about the Microsoft licensing for its proposed standard that makes it unacceptable to the open source/free software worlds. Rosen’s reply makes interesting reading for all of us.The question posed to Rosen in an email:

“For my part, I keep hearing lots of people try to explain why the MS license is unacceptable, but I’m not fully convinced that it’s significantly different from the licenses that companies like IBM or Sun use when they contribute to open source. And some of it (e.g. the Apache position statement) doesn’t even seem to refer to the current version of the license, but rather to an earlier draft. Maybe I’m just being dense, but can anyone explain, in 100 words or less, what’s wrong with the MS license? (It isn’t that I’m unwilling to read more than 100 words, just that I think an effort at concise precision is rather
badly needed.)”

Rosen’s reply:

The last published draft of the Microsoft license (1) does not allow sublicensing and (2) requires separate execution by those who freely decide
to distribute open source software embodying the standard. This is not compatible with many open source licenses or with OSD #7, and it is
incompatible with the widespread processes of open source development and distribution. There are other subtle wording changes that are needed in the draft Microsoft license, but those two are the main problems.

[That was less than 100 words.]

Eben Moglen and I have made those arguments to Microsoft’s attorney. But if Microsoft executives are unwilling to understand these fundamental
incompatibilities, then I suppose we won’t reach agreement on a compatible license.

The fact that there are also unacceptable licenses from IBM and Sun is no excuse to let this one from Microsoft get by. Point me to specific other
incompatible patent licenses for specific IETF standards and we’ll talk to those companies too if necessary. Perhaps these companies don’t realize that non-sublicenseable patent licenses for industry standards are incompatible
with their own open source licenses. 🙂

Several of us have made it clear to Microsoft that this license will chill open source implementations of this IETF standard and we have suggested that
it should not be accepted by anyone who intends to distribute implementations under open source licenses.

Alternatively, some open source developers and distributors may conclude that the patent will never issue because of prior art, or even if it issues will never be enforced. If you don’t think a license from Microsoft will be needed, then perhaps you shouldn’t ask for one or sign it. I have no opinion on that, except to say it might be risky if Microsoft later decides to come after you, your distributors or your customers as patent infringers.