Subject: UML & copyleft: licensing designs
From: "Manuel A. Castro" <mcastro@andago.com>
Date: 14 Jun 2002 20:58:44 +0200



On other list I've read about the troubles and thoughts about
"licensing" UML diagrams, and how to do it "libre software style". I
would like to know your ideas about this.

The problem, that I don't remember this list has discussed, is what
license (free/libre) should have a "design in paper", in particular, a
complete design as UML diagrams and suplemental info.

In the list there are two positions, that I call the BSD-front and the
GPL-advocates. The BSD-front seems to thought appropiate to just make
public all the designs and, if someone is able to implement them in code
and sell it as propietary software, so it be, as far as the original
authors of the design are recognized as such.

To the GPL-advocates this is unfair, the final product should be GPLish
too.

The thead went on mentioning the similarities among a UML design and
RFCs, and the differences between design and implementation. Then
someone pointed that UML is near to be "almost executable code", and
then a GPL would be required.

I use to be a GPL-advocate, but in this case I feel like the BSD-front
is correct here: it takes work and knowledge to move from "paper design"
to executable code - just take some RFC and code it. But then, an RFC is
mean to be a standard, not just a program, so is not really a good
comparison. And what happens when someone makes trivial or automagical
the compilation from UML to code? (To me this is almost impossible - it
would require such a fine grain in the design that it wont be design
anymore, just coding. And then, GPL would be appropiate)


Any thoughts?

--
happy hacking... man