Copyright

Martin Pool martinp at mincom.com
Thu Apr 15 05:01:05 UTC 1999


Derek J. Balling wrote:

> I would FURTHER go so far as to allow alteration of the licenses, but that
> the "lineage" must be documented, so that people familiar with [for lack of
> a better term] the OSI-BSD license (whatever they come up with) can say,
> "Ah, this license from Acme is based on the BSD license, but they changed
> it somehow. How did they change it I wonder?" and then they'll find out via
> diff/etc.  They'll know all the rest of it as "tried and true" so to speak,
> and then dig down to the alterations that a particular company needed for
> their particular business model.
> 
> statistically, the fewer changes to the baseline the better, obviously.
> 

> We WANT companies to embrace Open Source as much as their business model
> allows.

That's a fairly broad assertion.  "We" want all kinds of different
things.

> We DON'T want a user to have to read through each and every OS license
> because there's no "Standardization" to it (e.g., Acme Co may have based
> much of their license on OSI-BSD license, but if I don't know the parentage
> and diff it accordingly, then I have to sift through the whole thing to
> look it over for problems, or pay my lawyer to do so).

Indeed.

I imagine saying that a particular contract is based on some other
contract would be fairly sticky in law.  That statement (as I understand
it) would provide context within which to interpret the contract under
consideration, which makes things much more complex.

Imagine the difficulties for (say) GNU/Linux distributors if they had to
check the terms and conditions of every little package before putting it
in.  Much better to be able to handle 95% of packages by saying 'yes,
it's GPL' or 'yes, it's BSD'.  See the Debian FSG.

Imagine the difficulties people could get into by making what they
consider to be small modifications to the GPL that have unseen
ramifications.  (Admittedly "prohibit it because it's dangerous" is not
a very hackish philosophy, but "if you want to do something that
dangerous I'm not going to help you" is reasonable.)  Encouraging people
to stick with a well-debugged license is good.  If you're a large enough
organization to pay lawyers to do it properly, then you will probably
want to draft it from scratch anyhow.

> [OSI license examples]

Another important instance is public domain code, which is still useful
for some purposes.

It'd be nice to document and genericize the Sun, IBM and Netscape
licenses.  I imagine the OSI discussions with these companies are aimed
at developing such a beast.  Hopefully after people generate a few more
${company} Public Licenses OSI can standardize on (say) six different
ones that cover most situations: GPL, LGPL, X/BSD/Apache, public domain,
NPL, ... 

-- 
 /\\\  Mincom | Martin Pool          | martin.pool at mincom.com
// \\\        | Software Engineer    | Phone: +61 7 3303-3333
\\ ///        | Mincom Pty. Ltd.     | 
 \///         | Teneriffe, Brisbane  | Speaking for myself only




More information about the License-discuss mailing list