Subject: License Question
From: "Stephen J. Turnbull" <turnbull@sk.tsukuba.ac.jp>
Date: Wed, 16 May 2001 17:44:24 +0900

>>>>> "Crispin" == Crispin Cowan <crispin@wirex.com> writes:

    Crispin>      Is our immunized Apache a derived product, and we
    Crispin> may not call it "Apache"?  Or is it actually Apache per
    Crispin> se, and this clause does not bear on us?  Do you have an
    Crispin> opinion on whether our approach is compatible with the
    Crispin> Apache license?

I think this depends on whether these are purely internal compiler
techniques or could be considered "automated source transformations"
(ie, you're automatically doing what OpenBSD does by hand).  If it's
just a compiler, then you can argue that it's just an ordinary Apache
binary, targeted to "the Immunix system".  I think the Apache license
probably doesn't care if you link to special libraries for system
functions such as memory allocation, etc, either.

BTW, the GPL presumably requires publication of source to those
libraries, if any, although if there actually is a full "Immunix
system" product, you might be able to avoid that under the "Motif
loophole."  I doubt that, however; I'd interpret the "system" as
GNU/Linux, and Immunix as an add-on feature -- rms certainly has
tended to lean that way in the past (eg, for Qt).  I'd watch it if I
were you; a smart lawyer might be able to pry open your compiler
technology (or force you to stop distributing GPLed products) by
arguing that the code transformations constitute an implicit library
being linked with the GPL'ed code.


-- 
University of Tsukuba                Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences       Tel/fax: +81 (298) 53-5091
_________________  _________________  _________________  _________________
What are those straight lines for?  "XEmacs rules."