I go back and forth on how helpful RMS is. On one hand, where would we be without the GCC? This is something he helped get going.
On the other, I see him as someone who misinterprets a bunch of things because it is convenient for him politically to do so. The two examples are what happens when you dynamically link GPL software to proprietary software (static linking is arguably a bigger deal with the GPL v3 than the GPL v2), and taking extreme positions as to what the GPL versions require of developers bridging proprietary and GPL software.
The second is his view of what the BSD license allows and he makes false claims here he bases his view that the BSD and GPL v3 licenses are compatible on. The BSD license probably does not allow sublicensing in the opinion of most lawyers I have talked to (Eben Moglen being the sole exception!) and so if relicensing/sublicensing is a requirement for compatibility with the GPL v3, the BSD license is incompatible but the MIT license is compatible. Last I discussed this with Moglen he seemed to think it was safe just because the BSD author would lack standing or on other technical grounds. Every other lawyer I have talked to says "don't assume it's safe to take BSD code and add license restrictions without first making significant changes" (this is also the Software Freedom Law Center's official recommendation also) something the GPL v3 explicitly requires. (The counter-argument is that everyone involved in drafting the GPL v3 believed it to be compatible with the BSD license and this was a goal so BSD compatibility should be read into the license. This was what Richard Fontana argued when I brought this up on the OSI email lists).
I actually still use the GPL v2 and BSD licenses almost exclusively, and RMS is a big part of the reason I won't consider upgrading to the GPL v3.
On the other, I see him as someone who misinterprets a bunch of things because it is convenient for him politically to do so. The two examples are what happens when you dynamically link GPL software to proprietary software (static linking is arguably a bigger deal with the GPL v3 than the GPL v2), and taking extreme positions as to what the GPL versions require of developers bridging proprietary and GPL software.
The second is his view of what the BSD license allows and he makes false claims here he bases his view that the BSD and GPL v3 licenses are compatible on. The BSD license probably does not allow sublicensing in the opinion of most lawyers I have talked to (Eben Moglen being the sole exception!) and so if relicensing/sublicensing is a requirement for compatibility with the GPL v3, the BSD license is incompatible but the MIT license is compatible. Last I discussed this with Moglen he seemed to think it was safe just because the BSD author would lack standing or on other technical grounds. Every other lawyer I have talked to says "don't assume it's safe to take BSD code and add license restrictions without first making significant changes" (this is also the Software Freedom Law Center's official recommendation also) something the GPL v3 explicitly requires. (The counter-argument is that everyone involved in drafting the GPL v3 believed it to be compatible with the BSD license and this was a goal so BSD compatibility should be read into the license. This was what Richard Fontana argued when I brought this up on the OSI email lists).
I actually still use the GPL v2 and BSD licenses almost exclusively, and RMS is a big part of the reason I won't consider upgrading to the GPL v3.