Goals

[Patrick Maupin <pmaupin@jump.net>]

To complement the license feasibility discussion
(where I am waiting on a couple of email responses,
 and trying to figure out how to find a lawyer),
I thought we could reopen a goals discussion.

We all have slightly differing goals, and will
always, because of our backgrounds, temperaments,
desires, needs, etc., continue to have a diverse
set of goals.  Individuals even have different
goals at different times for different projects.

I think that, just as in free software, there will
be several different free hardware licensing schemes
to accomodate the different goals of the different
authors, but, again, just as in free software, I
think the largest "market share" for free hardware
will go to a BSD-like license (a gift with no strings
attached) and a GPL-like license (you can use it however
you want, but you have to share everything you've
done with it), and, of course the LGPL-like license
(you can use it however you want, but you have to
share your modifications and enhancements).

Perhaps as a group we should go ahead and try to
develop all three licenses simultaneously, to avoid
fragmentation, and to insure that the licenses
"interoperate," as BSD without the advertising
clause can interoperate with GPL.

I see the following breakdown:

The BSD-like license is relatively uninteresting,
consisting primarily of an exceptionally long
legal disclaimer.  I think we should work very
to make sure that these disclaimers are sufficient.

The LGPL-like and GPL-like license build on the
BSD-like license (may as well share the same disclaimers),
and add restrictions to what can be done with the
hardware source code.  Most of the added language has
to do with _disclosure_, e.g. with the LGPL, you need
to disclose modifications, and with the GPL, you need
to disclose all linked code.  Additional language has
to do with licensing the disclosed source without
additional restrictions, with when the disclosures
must be made, etc.  Finally, there should be clauses
to aid in finessing copyright law to force compliance.

I actually see the BSD/LGPL/GPL like licenses as
belonging on a protection continuum, and would like
to try to structure them so that there is a single
license, PLUS a set of "exhibits" which discuss
exactly where on the protection continuum a particular
author wants the particular hardware design.  I think
this concept could remove the need for a lot of
redundant free hardware licenses.  I think that the
entity developing the license could also develop
several standard exhibits as examples of good legal
contract construction, but that the exhibits would
be small enough that it would not be too much effort
to have to read a particular one.

In the spirit of the GPL, the exhibits would be
allowed to describe how much of the final design
must be disclosed (and relicensed to all comers),
and when it must be disclosed (external distribution,
in-house use, etc. -- this may be much more important
than for software, as building a chip in-house may
be considered an in-house use even if the chip is
sold on the open market), but would not be allowed
to add additional restrictions other than disclosure
(the main license could say that anything required
to be disclosed by the exhibit must be relicensed
under the same terms).

This would mean that someone who discloses and allows
relicensing of their entire design does not even need
to read the exhibits of source code they incorporate
into the design, because the exhibits only affect
people envisioning some amount of non-disclosure.

For example, in the source code:

// Copyright (C) 2000, foofram inc.  All rights reserved.
// This design is licensed in accordance with the
// opencollector.org OHL version 1.0 or later, used in
// conjunction with the foofram.com strange_widget
// disclosure exhibit version 1.0 or higher.

If this concept is useful, it might even be worth "back-porting"
to the GPL, e.g. version 3 of the GPL could claim that it, in
conjunction with a particular FSF exhibit, is the successor
license to GPL version 2, and that it, in conjunction with a
different particular exhibit, is the successor to the LGPL,
and the FSF could also supply several additional standard
exhibits, for people who don't want their code dynamically
linked with other programs, etc.

In this way, the tastes of individual authors could
be catered to with finer granularity than the current
choice of GPL or LGPL allows, while preserving the
ability to link together large collections of source
code as long as the aggregator is willing to abide by
the most demanding disclosure exhibit in the collection.

So, I guess there are a couple of questions about this:

1) Is the license/exhibit idea both feasible and useful?
2) Will this license actually prevent a lot of license
   "forks" or are there other concerns besides disclosure
   and relicensing which will be shared by a large number
   of authors?


In the interest of "disclosure" I think I should state
that my personal goals are to be able to dual-release
HDL code (primarily Verilog) I write under a
maximum-strength, full-disclosure hardware license,
and also under a private label license.

For extremely small development shops, one of
the best free software development models I have
seen (best as in allows the author to support
free software and still make a reasonable living)
seems to be that used by Aladdin (ghostscript)
and, more recently, TrollTech.

This model lets the developer be paid in Karma,
bug-fixes, suggestions, and availability of other
free software by similarly situated developers and
users, and to be paid in cold, hard cash by
companies unwilling or unable to incorporate free
software into their designs.

This model can also simplify my life.  I don't
have to worry about NDAs or secrets.  My resume
can include a portfolio of actual code I have
developed.  I'll have people help me debug what
I write.  My customers should be happier -- if
somebody pays for a commercial license, I can
be pretty sure they understand the capabilities
of the design before they give me money,
and that I won't have to give the money
back.  If I'm violating someone's patent,
I'm likely to find out sooner, rather than
later, because they can see my design.
My advertising and sales expenses should be
reduced, because I don't need to pay some
oily salesman in a pinstripe suit to explain
how great my product is to people who haven't
seen it.

Finally, the model helps support the goal,
which Mr. Stallman has articulated so clearly,
of putting building blocks into other creative
peoples' hands, and letting them share the
results of their derivative creativity with
the world, without fear of legal troubles.

Of course, this only works if we can actually
develop a reasonable copyright-based free
hardware license.  Otherwise, it's back to
the NDAs and the oily salesmen, because I
have to work for a living.

Thanks,
Pat


-- 
Patrick Maupin
5216 Crooked Oak Cove
Austin, Texas 78749
512 891 6037