Must publish vs. must supply

[Abe Kornelis]

John,

thanks for your reply. Please find my comments inserted
in between your text.

----- Original Message -----
From: John Cowan <jcowan at reutershealth.com>

> > The BXAPL (see http://www.bixoft.nl/english/license.htm)
> > currently has both - which is definitely an overkill,
> > even though it grants users the right to keep their
> > modifications entirely private. That is: one either keeps
> > all modifications private, or they are published to
> > public in general. Selective distribution is not
> > allowed. Users who fail to comply are obligated to
> > supply their modifications upon request.
>
> As to "must publish", I think it's unnecessary.  Those who wish to keep
> their mods private will do so whether the license allows it or not:
> if they sit on the changed version of the software, no one will know
> that it even exists.
--> And that would be entirely ok with me.

> The GPL and the OSL take what I consider to be a reasonable attitude:
> you must supply changes in source form to people who have received
> the changed version.   If the changed version is published to all, the
> changes must also be; if the changed version is distributed to a few,
> ditto the changes; if the changed version is never distributed, the
> changed version need not be either.
--> I understand that - it is one of my problems with these licenses:
      I'd really hate it if modifications were to be distributed within
      a closed group - not to the public that is. Distributing within closed
      groups may lead to discrimination. I concur with the OSI
      that discrimination is undesirable. Is it a good idea to force
      distribution to the public? Or am I being over-zealous?

>  Of course, the distributees may
> themselves distribute, but at least there is no unlimited liability to
> distribute one's changes to anyone who asks for them.
>
> This is quite separate from the question of whether the change is
> *licensed* to all.  No matter what the distribution conditions, anyone
> who has possession of the change is licensed to use it.
--> I agree with the latter, but don't understand the first remark.
      As far as I can see, all OSI-compliant distros are licensed
      to the receiver, which makes 'distributing to' functionally
      equivalent to 'licensing to' - except where receiver does not
      comply with the license's restrictions, but that's not (yet)
      relevant when the software is distributed.
      So how am I to understand your first remark?

> The trouble with the usual kind of "must supply" (which I take to mean
> "must supply changes to the original author")
--> Right indeed.

> is the burden of doing so
> for those who make small changes to a large number of works -- people
> like Linux distro makers and *BSD groups.  If they have to push the
> changes to the original author, the original author may be difficult
> or impossible to find, or (if corporate) may have gone out of business.
> It is much better, if you are going to require such a thing, to let the
> distro creator push to you a distribution point (such as a Web site)
> from which you may pull the changes yourself.
--> There you have a point. Both the copyright holder and the
      contributor may go out of existence - eventually we all do,
      don't we ;-)
      Anyway, I was rather thinking of making things optional:
      1) keep changes private - no distribution at all
      2) distribute to the public
      3) distribute & supply to copyright owner
      Option 1 is obvious, the main difference between choices
      2 and 3 would be the requirement to make the modifications
      public *yourself* or allow the copyright holder to do so
      in your stead. If you already have a site option 2 might be
      more attractive, if you don't have a site, or you have one
      that is not related to the changes or the software, then
      option 3 might be more attractive since it'd be less of a
      hassle. Is it a bad idea to allow the contributor a choice as
      to how the changes are to be made public?

> In general, though, I think all these requirements are over-cautious.
> Most people do not want to maintain forks indefinitely -- they *want* to
> push changes back to you in the hope that you will integrate the changes
> into the mainline distribution, and they will get them back automatically.
--> Sorry, John. I disagree. I know a few companies who would be
       willing to pick up my code, fork it and distribute it both to their
       customers and to their advantage - and never give anything back.
       As far as I'm concerned, they can make it proprietary if they
       want to - but *not* under an open-source license!

Kind regards, Abe.


--
license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3