Subject: Re: For Approval: Microsoft Permissive License
From: "Chris Travers" <chris.travers@gmail.com>
Date: Mon, 24 Sep 2007 15:30:32 -0700
Mon, 24 Sep 2007 15:30:32 -0700
On 9/24/07, Wilson, Andrew <andrew.wilson@intel.com> wrote:
>
>
>
> If you are trying to convince people that your reading of BSD is
> not extreme, citing Theo is perhaps not the best place to begin.


My main point is that the "license applies to all source code recipients for
our copyrights" interpretation is not at all uncommon.  Instead, I believe
it is your opinion which is extreme (based on perhaps deliberate
misinterpretations by people like RMS).  Furthermore, though IANAL, I did
note that the few lawyers who have addressed this question on this list have
given answers which back my argument here that the license always follows
the copyrighted element, and that it is a matter of what you can safely
extract rather than what license is wrapped around the code.

So far, I have not found any open source community project (single vendor
reference implementations don't count, so that rules out things like Intel's
ISCSI Reference Implementation) which interpret the license the way you do.

Since one of my projects depends on PostgreSQL and is under GPL 2+, I
contacted their core team so that I would be better aware of any license
conflicts, real or perceived.   They expressed concern that permission
removal under the GPL might violate the license, but suggested that this
wasn't their problem (IMO they are probably right on both counts).

To paraphrase the PostgreSQL core team:  their position is that commercial
variants amount to enforcing the new copyrights that the vendor adds and
does not involve any sublicensing of the source code under the BSD license
itself.

This is why the permission grant is required to be included on every copy of
the source code.  Again, this does not provide any obstacle to enforcing
one's own copyrights in arbitrary ways.  My suggestion is that you find
major BSD-licensed community projects (not single-vendor reference
implementations) which will agree that wrapping their code in the GPL
without adding any copyrighted elements is allowed.

In short, I don't think your viewpoint is in line with how most projects
which use the license read it.

Best Wishes,
Chris Travers




On 9/24/07, Wilson, Andrew <andrew.wilson@intel.com> wrote:


If you are trying to convince people that your reading of BSD is
not extreme, citing Theo is perhaps not the best place to begin.

My main point is that the "license applies to all source code recipients for our copyrights" interpretation is not at all uncommon.  Instead, I believe it is your opinion which is extreme (based on perhaps deliberate misinterpretations by people like RMS).  Furthermore, though IANAL, I did note that the few lawyers who have addressed this question on this list have given answers which back my argument here that the license always follows the copyrighted element, and that it is a matter of what you can safely extract rather than what license is wrapped around the code.

So far, I have not found any open source community project (single vendor reference implementations don't count, so that rules out things like Intel's ISCSI Reference Implementation) which interpret the license the way you do.

Since one of my projects depends on PostgreSQL and is under GPL 2+, I contacted their core team so that I would be better aware of any license conflicts, real or perceived.   They expressed concern that permission removal under the GPL might violate the license, but suggested that this wasn't their problem (IMO they are probably right on both counts).

To paraphrase the PostgreSQL core team:  their position is that commercial variants amount to enforcing the new copyrights that the vendor adds and does not involve any sublicensing of the source code under the BSD license itself.

This is why the permission grant is required to be included on every copy of the source code.  Again, this does not provide any obstacle to enforcing one's own copyrights in arbitrary ways.  My suggestion is that you find major BSD-licensed community projects (not single-vendor reference implementations) which will agree that wrapping their code in the GPL without adding any copyrighted elements is allowed.

In short, I don't think your viewpoint is in line with how most projects which use the license read it.

Best Wishes,
Chris Travers