OpenLDAP license

Derek Seabury Derek.Seabury at
Fri Apr 13 20:12:23 UTC 2001

"Ryan S. Dancey" wrote:

> #3:  BSD complies, but is weak because it does not use a copyleft mechanism
> to require that the right to make derived works to be carried forward to
> each recipient.  In other words, I can take a work using the BSD, add a
> modification, but restrict the right to make further modifications of my
> modification.  The BSD does not require me to license my modifications using
> the BSD.

OSI #3 does not intend to require that software restrict it's usage to other OSI
approved projects.  Even FSF has realized that such a requirement is a barrier
to the usage and acceptance of software and created the LGPL.
    This is of course based on my assumption that the intent of OSI is to
encourage open source by establishing a baseline set of requirements.  If the
intent is to battle the concept of software IP perhaps that is the intent, but
that is contrary to #9.

> [ as a side note, I think this is one of the places where the OSD itself is
> flawed.  The language of #2 should say, in my opinion:  "The license must
> allow modifications and derived works, and must REQUIRE them to be
> distributed under the same terms as the license of the original software." ]

Given #3 (I'm assuming you meant #3, not #2) insures recipients of the code the
ability to get the same terms all that a change to "REQUIRE" would do is prevent
people from offering more attractive terms.

For example under OSI I couldn't guarantee software I wrote to work with a given
package or configuration (by #8), something I might want to do.

> #7:  BSD does not comply. (BSD code could be distributed in binary-only form
> with completely different and more restrictive licensing terms than the BSD).

BSD grants rights to "Redistribution and use in source and binary forms" if you
gave a binary only how could I execute my rights to distribute the source?

> > The term "Open Source", applied as an attribute of software, means that
> > the software generally follows the criteria set forth in the OSD.
> Unfortunately, it does not.  The definition of the term is subjective, not
> definitive.  That's why "OSI Certified" is important.

I think it does imply it, but only from the perspective of the party making the
claim.  "OSI Certified" implies an external review has been undertaken that
gives the claim more credibility.  I could tell you the SpeechWorks Public
License is open source even though it is not certified.  It hasn't been denied
by the OSI, and unless they do I'll assume it to be true.

> If the BSD is found not to sufficiently encapsulate the OSD (and in my
> opinion it does not), then the OSI should not certify it.  Otherwise, in my
> opinion, the certification is essentially meaningless.  OSI Certification
> should mean "the rights granted to you WILL comply with the OSD."  Not "MAY"
> comply with the OSD.

Would you agree that source code distributed under BSD is?  Binaries might not
be, but I think source under BSD is sufficient to meet the definition.  Or at
least close enough that it would take lawyers and a court decision to 'prove'
one way or another.

Derek.Seabury at

More information about the License-discuss mailing list