John Goerzen <[EMAIL PROTECTED]> writes: > On Thu, Jun 07, 2007 at 12:17:28PM -0700, Walter Landry wrote: >> GnuTLS + libgcrypt + libtasn1 implements everything unless you need >> ECC. >> >> > And why does FSFE disagree with our interpretation? >> >> Michael Poole gave a good answer. > > He didn't address the FSFE -- where are they taking a different analysis > than us on this?
I have not seen a specific statement from FSFE on the question, so I do not know and I had preferred not to guess. I suspect that their analysis was more liberal on -- or overlooked -- the angle of Debian as distributor of OpenSSL in accompaniment with GPLed works that are linked against it. Perhaps the FSF's intent for the GPL is to allow a more LGPL-like option with regards to system libraries; that would make sense and I think not sacrifice software freedoms. However, I have not seen a defensible way to construe the actual wording other than the one that classifies Debian's method of distribution as each work accompanied by the other(s). [As a side note, I think that resolving this in favor of allowing works under the plain GPL to dynamically link against OpenSSL would allow the aggregation of binary firmware blobs into a package of the Linux kernel. The lack of source for those blobs would still bar them from "main", but it would not be a license issue. This is meant simply as an observation, not a judgment of whether that would be a good or bad result.] I am not a lawyer, but I belive that most lawyers urge clients to err on the side of caution when straying from a strict interpretation of license -- or even when interpreting ambiguous clauses. I try to follow that guide when making suggestions on what a license permits. In this case, the wording is consistent and clear, although in an unfortunate direction, and I would not want to rely on extrinsic evidence[1] about the GPL's meaning if I were in court. [1]- http://en.wikipedia.org/wiki/Parol_evidence_rule Michael Poole -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]