Manual contributions and copyright assignments

Manual contributions and copyright assignments

[Gerald Pfeifer]

RMS asked us to implement the following policy:

  [P]lease don't install changes in the manual on the strength
  of copyright assignments signed before Jan 2002.  If a person signed
  before that, assign@gnu.org can look at the papers and tell you if
  they explicitly cover documentation.  If not, we should get new papers
  from that contributor.

As far as I understand, obvious fixes and minor changes (where we didn't
need a copyright assignment before) are not affected.


RMS, how do you suggest to proceed practically?  This seems to affect more
than 100 developers with CVS write access, any of which can commit patches
of her/his own or others, and about 700 contributors overall (according to
/gd/gnuorg/copyright.list).

Gerald
-- 
Gerald "Jerry"   pfeifer@dbai.tuwien.ac.at   http://www.pfeifer.com/gerald/

Re: Manual contributions and copyright assignments

[Richard Stallman]

    RMS, how do you suggest to proceed practically?

I wouldn't want to start micro-managing you ;-).

						     This seems to affect more
    than 100 developers with CVS write access, any of which can commit patches
    of her/his own or others,

You surely have some policies about which areas any given person can
commit changes in.  Right?  So it is a matter of changing these policies.

			      and about 700 contributors overall (according to
    /gd/gnuorg/copyright.list).

What problem do they raise?


I am worried about the idea that all 100 people with write access can
install others' changes.  That means we depend on all 100 of them to
check properly for legal papers.  It is unreliable to have 100 people
doing this!

I have no objection to your allowing 100 people to install their own
changes--if you trust them, that's good enough for me.  But unless you
have very effective procedures for making sure all 100 people check
papers properly, you should greatly reduce the people who are allowed
to install changes other than their own.

One idea is that each of these 100 people can have a list of those
whose changes he can install.  Some of them, you could trust to check
papers just as you do.  But most of them would have a specific list of
names.  If a person wants to add a name to his list, he asks you,
and you check the papers.

Re: Manual contributions and copyright assignments

[Joe Buck]

On Thu, May 22, 2003 at 04:33:49AM -0400, Richard Stallman wrote:
> I am worried about the idea that all 100 people with write access can
> install others' changes.  That means we depend on all 100 of them to
> check properly for legal papers.  It is unreliable to have 100 people
> doing this!

Don't worry; only a very small number of people have global write privilege.
The others only have permission to modify very limited areas of the compiler,
or to check in a patch that has been approved by someone with greater privilege.
Generally speaking, these people are checking in their own patches, and
all have papers on file.

> I have no objection to your allowing 100 people to install their own
> changes--if you trust them, that's good enough for me.  But unless you
> have very effective procedures for making sure all 100 people check
> papers properly, you should greatly reduce the people who are allowed
> to install changes other than their own.

That's pretty much the process we follow; when an unknown person submits
a patch, the very first question asked has to do with the status of their
papers.
 

Re: Manual contributions and copyright assignments

[Nathanael Nerode]

Gerald wrote:
>RMS asked us to implement the following policy:
>
>  [P]lease don't install changes in the manual on the strength
>  of copyright assignments signed before Jan 2002.  If a person 
>signed
>  before that, assign@gnu.org can look at the papers and tell you 
>if
>  they explicitly cover documentation.  If not, we should get new 
>papers
>  from that contributor.
>
>As far as I understand, obvious fixes and minor changes (where we 
>didn't
>need a copyright assignment before) are not affected.

Let me state for the record that I consider all my changes to the manual 
to date to be uncopyrightable.  So at least you don't have to worry 
about me for the past.

>RMS, how do you suggest to proceed practically?  This seems to 
>affect more
>than 100 developers with CVS write access, any of which can commit 
>patches
>of her/his own or others, and about 700 contributors overall 
>(according to
>/gd/gnuorg/copyright.list).

For the *future* there's no real problem; there's just a major 
annoyance. Someone has to check the paperwork status of everyone in
MAINTAINERS w.r.t. documentation, and ask each of the no-documentation
pre-2002 people if they wish to sign a new documenation assignment.  
For those who don't, they're marked as 'no-documentation' in MAINTAINERS
and notified. :-/  For those who do, their documentation changes can't be
committed until their paperwork is complete.

For the *past*, however, there's a serious problem; people may have 
checked in documentation changes under the assumption that they were 
paperwork-complete, and those people may no longer be maintainers, may 
not want to sign new paperwork, or whatever.  But their changes may not 
be trivial, or indeed practically removable at all (if other people 
have built on them).  If we are not to install their changes (which are 
already in CVS) what are we supposed to do?

--Nathanael