[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: license of smapi



Scripsit Mike Bilow <mikebw@colossus.bilow.com>

> > > 3) If you modify this code, you must keep the message format compatible.

> > That is also non-free.

> I disagree here.  Dudley is trying to prevent people from making
> proprietary "embrace and extend" changes to the message format.

Yeah, that's also what Sun is saying each time they pull this one.

I'll leave open whether this should be attributed to malice or
thoughtlessnes, but I'd like to point out at least two adverse
effects:

One: This kind of clause gives the author (who is free to ignore
his own rules) a de facto monopoly on experiments with alternative
ways of doing things. One of the points of free software is that
everyone should be free to try out new ways of doing things. The
license clause denies the users the freedom to try out and exchange
new ideas. We belive that this freedom is one of the most important
driving factors for progress in computing - and we like progress.

Two: Another very important facet of software freedom is the freedom
to adapt old tools to new problems. If I find a program in Debian
main which I think can be adapted to solve a problem I have (and
which nobody ever thought about before), I expect to be allowed to
make that adaptation - even if that means I need to change some of
the external interfaces of the program. And I shouldn't need to
worry about the fact that my adapted tool does not solve the original
problem anymore if the original problem is not what I need to solve.

It is OK if the license requires me to document what I changed
if I give my changed code to someone else who also needs to solve
my new problem - but that is different from the clause above,
which says that I must not change the data structures, full stop.


If you like section cites better than rationales, please see point
3 of the DFSG. Notice that the DFSG does not allow the license to
restrict which kind of modifications are permitted.

> This would be something like Debian requiring that no one make proprietary
> changes to the .deb file format,

We're not talking about proprietary changes (the license of the
.deb tools does not allow proprietary deriviates at all, whether
or not the file format is changed). We're talking about *any*
kind of changes, and requiring that no one make any kind of
changes to the .deb file format would be non-free.

> or requiring that mail programs comply with RFC822.

Which would be non-free.

Indeed, where would we be if not long ago somebody had taken a piece
of mail software and changed it to use @-style addresses instead of
bang paths? That would have been illegal under the kind of license
you advocate.

> The whole point of Fidonet is that it defines a technical
> file format which permits interaction between systems;

That's irrelevant to software freedom. If someone wants to modify his
system such that he can only communicate with his cousin in Australia,
he might have to leave Fidonet - but if the sheer act of doing so
gives the author a reason to sue him, the software is not free.

-- 
Henning Makholm                                           "Check the sprog."



Reply to: