Copyright lines, acknowledgement WAS: Re: svn commit: rev 6263 - incubator/spamassassin/trunk/spamc

[Sander Striker <st...@apache.org>]

On Mon, 2004-01-26 at 20:33, Malte S. Stretz wrote:
> On Monday 26 January 2004 20:01 CET Duncan Findlay wrote:
> > On Sat, Jan 24, 2004 at 08:11:44PM -0000, mss@apache.org wrote:
> > > + * Copyright 2001-2002 by Craig Hughes
> > > + * Conversion to a thread-safe shared library Copyright 2002 by Liam
> > > Widdowson + * Portions Copyright 2002 by Brad Jorsch
> > > + * Windows adaption Copyright 2004 by Sidney Markowitz
> > >   *
> > >   * <@LICENSE>
> > >   * Copyright 2004 Apache Software Foundation
> >
> > Isn't this really confusing? Can the same file be copyright by 4
> > people as well as the ASF? Doesn't the CLA give us the right to simply
> > distribute as "Copyright 2001-2004 Apache Software Foundation"
> 
> Actually, what is more confusing in my eyes is the Apache copyright stanza.

The current position of the ASF is as follows:

The number of copyright lines has to be kept down to as few as possible.
A single line with 'Copyright Apache Software Foundation' is prefered.
While multiple copyright lines can coexist, they are undesirable for
the following reasons:

 * Legal protection

    When personal copyrights are mentioned, the ASF can do very little
    when it comes to legal protection (nothing we can do when someone
    is sued on personal title).

 * Social aspects

   Acknowledgement of contributions per file has shown not to promote
   colloborative development.

   There is the pattern where people try and touch as many files as
   possible to get their name in as many files as possible.

   There is also the creation of personal islands within the codebase.
   The acknowledgements make this sentiment stronger.  It will be
   less likely that someone touches a file that is marked as someone
   elses domain.  This is bad.

The ASF obtains enough rights via CLAs, Corporate CLAs and Software
Grants, to re-license contributed works any way the ASF sees fit.
Up to and including, the removal of copyright lines.  In reality it
is more complex, but this is the bottom line.

> I'd personally like to keep a list of major contributors (if they want) in 
> the files because that gives some acknowledgement to their work.

This is highly discouraged.  There are better ways to give credit, take
a look at the CHANGES file in Apache HTTP Server for example:
http://cvs.apache.org/viewcvs.cgi/httpd-2.0/CHANGES?view=markup

There's also this page acknowledging the committership:
http://httpd.apache.org/contributors/

[...]
> And call it a big ego, but I myself like to see my name 
> on files on which invested quite some time :)

And this is exactly what we want to watch out for.

Just imagine the bulk of acknowledgement lines that are going to be
accumulated in a file over time.  If the committer body is turning
over, partial rewrites happen, more lines get added, etc. etc. 
Furthermore, where do you draw the line when to add a personal
acknowledgement line?  A single character change, a line change,
a few lines, time spent on a patch?

The ASF tradition is not to do it.


Sander