httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Paul Richards <>
Subject Re: vote deadline ?
Date Fri, 15 Sep 1995 15:40:11 GMT
In reply to Robert S. Thau who said
> Paul, as the sole author of many files in this distribution (including
> several crucial files in the server core --- alloc.c, http_config.c, 
> http_core.c), I have absolutely no idea why you insist on cutting NCSA
> out of global credit.  Their server, as a whole, *was* the basis for
> our server, as a whole.  What is the problem with saying so?

I'm not trying in any way to cut out NCSA from global credit. I'm trying
to construct a legally sound license. You *can't* say that portions of
this code are derived from NCSA if NCSA has never been anywhere near
them. Be aware that the Apache license will be more restrictive than NCSA
since they gave away all rights to their code when they placed it in the
public domain. That's not what you (the Apache group) desire since you
want to retain control over the licensing of the product.

To do this you have to have a copyright holder, which will be
yourself for those files that you wrote. This actually raises a
little issue, if there are completely untouched NCSA files still
in Apache then they should stay in the public domain since there's
no-one I can assign the copyright to, unless the other Rob steps
forward and says he wrote certain files. If there are files that are
basically NCSA files with some patches then someone needs to be
assigned the copyright or those patches will also have to go into 
the public domain under the original NCSA notice. This is testing
my understanding to the limit since I've never dealt with public domain
code before but I believe that since the original NCSA code was in the
public domain then the derived work (which it is if you apply patches)
can be copyrighted as a derived work by the person who wrote the patches.

The whole point of the licensing problem is that a restrictive license
can only be placed on code that has a Copyright holder. I better explain
what I explained to Rob H.

This argument assumes there is the means to contest any abuse of the license
which is unlikely at this point since that takes cash but we're redoing the
licensing in order to allow this should it ever become an issue.

If some body takes Apache and sells it and people here are pissed off
about it then you'd have to begin legal proceedings. When you go to the
lawyer the first thing they'll ask is, "Who's the owner of this code?". If 
there is no legal entity who can speak authoratively then the lawyer simply
won't deal with the issue. The lawyer will have to make certain decisions in
dealing with the case (say for instance they agree to settle out of court,
someone has to decide if the terms are satisfactory). If, hypothetically,
Rob H. had initiated proceedings on behalf of the project then he might
decide that the terms are satisfactory, then someone else from the project
might come along and say, hang on, these terms suck, this isn't good enough
keep fighting. The lawyer then doesn't know who the most authoratative voice
is and who he should actually listen to. Because of this they simply won't
take the case in the first place unless there is clear ownership of the

If Apache was incorporated then this person would be the major shareholder.
Since Apache isn't incorporated then the "owner" is the person who holds the
copyright. One of the reasons I argued about having each file individually
copyrighted and licensed was to clear this question of ownership up. 

There are other issues you might want to think about. If people apply
patches to code that is copyrighted by someone else then they could if
they wish, add they're own copyright as a derived work. This gets
very messy when a lot of people submit fixes to someone else's code and
in general, if you send someone a patch then what you effectively do
is place your patch in the public domain and allow the original author
to include it in their work under the original copyright.

This is not a worry for Apache as long as the license is satisfactory. I
better explain this too. Say Rob S. gets an offer for his code, he is
entitled to sell it and thereby have it release under a far more
restrictive license. That would not however, affect Apache since the
code already in Apache is under a license that allows Apache to continue
to use it, even if he was to leave the project and pursue commercial
activities using his code. This is true *even if the code is identical*
since there would be two versions of the same code released under 
different licences, which the author is perfectly within his rights to
do. This is effectively what has happened with NCSA.

Umm, I'm going on a bit, back to the original point, you can't "give away"
code under someone else's name so if NCSA has never been anywhere near it
then you can't list them in the license. Take another hypothetical
example. You have a license that doesn't have a correct disclaimer, someone
sues you, if you've claimed in your license that NCSA actually wrote some
of the code then you make them liable too. NCSA are safe for the code
they actually wrote since they just present their original code and
their original disclaimer, they can't do that if you've incorrectly
listed them as authors of code they didn't actually write, it follows
that you simply can't list other people as authors of the code if they
didn't write it! Incidentally, you're probably thinking, what difference
does it make if I just list them anyway, they can just deny it if it
ever came to court, well, yes, but to deny things in court you have to
prove it and they'd probably sue you for misrepresentation as part of
the process of denying the accusation against them.

Hey, the laws not fun, I didn't write it :-)

  Paul Richards, Netcraft Ltd.
  Phone: 0370 462071 (Mobile), +44 1225 447500 (work)

View raw message