Subject: Re: Must publish vs. must supply
From: "Abe Kornelis" <>
Date: Wed, 12 Mar 2003 22:34:25 +0100

----- Original Message -----
From: Mark Rafn <>

> >       The biggest point in this whole discussion is this simple
> >       fact: if I do not insert either a must-publish or a must-supply
> >       clause in my license they can (and probably will) claim that
> >       their source is available since they'd have to give it to their
> >       customers - who'd refuse to do anything but store them
> >       passively.
> IMO, such an ability is absolutely required by open-source software.  The
> chinese dissident case (Imagine a group of people who want to modify and
> share software among themselves, but who will be executed if it is
> discovered that they are working on this) is one common way to phrase this
> requirement.
--> You raise a touchy point. I'll give you two replies.
       1) Any solution that I would provide would equally apply to
           terrorist groups. Replace the Chinese dissidents with
           Al-Qaeda members - their situations are comparable
           but the way we think about their motives and goals
           are utterly opposite!
       2) I find it hard to believe that I should feel compelled to help
           these dissidents to solve their problems, however sympathetic
           their cause may seem. And if I would I'd still have to solve
           the conundrum above - without discriminating (see OSD).

> This is the choice of such customers.  They have source, so they have
> control of thier systems.  With luck, they're likely to ask you for help
> (being the original author) if they decide something is wrong.
--> They have their software vendors for support. They pay big
       bucks for support - no way are they going to ask for help
       from outsiders.

> >       As Chris said: a license needs teeth, and this one I deem
> >       to be one very important canine.
> It needs teeth to protect the software recipients from the software
> authors.  Teeth that protect an author from the recipients are the
> opposite of free.
--> It works both ways: users need protection from authors,
      *and* authors need protection from users who would
      prefer to both have my cookie (software) and eat it
      (resell without publishing).

> >       Is a must-supply (to copyright holder, that is) clause
> >       preferable over a must-publish (to the public, that is)
> >       clause, or vice versa.
> Neither qualify as acceptible in my book.
--> Not even when there is also an option not to distribute at all?
      I am not intending that *any* change be published - I only want
      to enforce that *if* changes are distributed, *then* they must
      be made available to the public.

>  I'd be interested to hear
> from OSI board members whether this is an area where "free" as commonly
> used by the FSF and Debian differs from "open source" as used by OSI.
--> It would be very nice to know their opinion.

Kind regards, Abe F. Kornelis.

license-discuss archive is at