Subject: Re: "On Virus" -- get real
From: "Stephen J. Turnbull" <>
Date: Thu, 30 Sep 1999 15:42:54 +0900 (JST)

Unless there is explicit request from an individual to be CCed, I am
going to trim the CC list to FSB and author to whom I am responding
(if I don't know him or her to be on FSB) on my future posts in this
thread, should there be any.

>>>>> "rms" == Richard Stallman <> writes:

	sjt>     Hereditary fails the hierarchy test; heredity is
	sjt> hierarchical, software reuse is not.

	rms> Heredity is not hierarchical in species that reproduce
	rms> sexually.

    For a single characteristic, introduced as a specific mutation, the
    graph of individuals containing that mutation is (nearly) a rooted

    rms> That is one way of looking at heredity, but not the only one.
    rms> (I was referring to another.)  To justify your conclusion,
    rms> you would need to show that this usage is strongly at odds
    rms> with all the natural ways.

Nope.  Remember, we're doing marketing here.  It is _you_ that suffers
the burden of showing that the terminology you advocate that _we_ use
(of course, you may use whatever terms you please yourself) will have
the connotations we wish to project to an audience of those who are
not as informed as we are, and who may never become so informed.

I think that the interpretation I propose is quite common, enough so
that it would be a serious hindrance to getting our point across.  In
particular, the strength of the "mitochondrial clause" (bletcherous
terminology, huh) is not at all apparent to people who use the "family
tree" model of "heredity"; they won't realize that "adopting" a GPLed
child causes all your other children to become GPLed, too.

Of course "child" here refers to the call graph, and the parent is
main().  I think adoption is a reasonable metaphor for incorporation
of a single GPLed function into a body of unrelated code.  And if the
parent isn't main(), but a small subsystem, the effect is even more
powerful, extending to grandparents and second cousins, etc ;-)

As for the rest of Richard's post: he is simply sidestepping the point
by concentrating on the definition of "public good," which is
irrelevant to the issue of my participation in some putative
"partnership" with an author (who I've never met) of GPL code (that I
have nothing like a partner's rights to).


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 two straight lines for?  "Free software rules."