Subject: Re: OpenLDAP license
From: Ian Lance Taylor <ian@airs.com>
Date: 11 Apr 2001 09:56:07 -0700

"Ryan S. Dancey" <ryand@organizedplay.com> writes:

> To me, this is the credo of the OSD.  If we are to support the OSD, then the
> OSD should capture what the great majority of the software community means
> by the term "Open Source", and I suggest that for the great majority of the
> software community (in addition to many other things) "Open Source" means
> "binary-only distributions are unacceptable."

Yes.  But I don't think that most people think that ``open source''
means ``binary-only redistributions are unacceptable.''  Of course,
such a redistribution would not be open source.  But the original
distribution would be open source, and would remain open source even
if a binary-only redistribution occurred.

> Here is what I see.  I see a requirement in the OSD that source code >MUST<
> be provided with binaries.  Then I see a specific exception provided that
> requires the distributor of the code, if a binary-only distribution is made,
> to make the source available in a "well publicized manner".

Yes, this all applies to an open source distribution.

> If the OSI wants to keep to this concept that licenses which allow
> binary-only distributions meet the OSD, then I think the OSD should be
> changed.

Licenses which allow binary-only distribution do not meet the OSD.

Licenses which allow binary-only redistribution do meet the OSD.  Note
that a redistribution does not affect the original distribution.

(You then have several paragraphs which argue based on what I believe
is a false premise.)

> "Accordingly, an open-source license >MUST< guarantee that the source is
> readily available..."
> 
> Care to explain what part of the BSD license guarantees that the source will
> be readily available?

It doesn't.  The BSD license sort of goes along with an actual source
distribution.  I suppose the BSD license could say something like
``this license must be accompanied by complete source code.''  The BSD
license tends to be used on a file-by-file basis, so that statement is
arguably not required.

In any case, the sentence you quote from the OSD rationale should not
be construed as saying that the source for a modified version
redistributed as binaries must be readily available.  I agree that
that is a possible reading.  But I don't think it is the right
reading.  OSD #4 says that the license must permit distribution of
source code.  It doesn't say that the license must prohibit
distribution of binaries without source code.

> > As such, the OSD was written to define what was and was not open
> > source software.  I'm not comfortable describing the OSD as an
> > ethical position.
> 
> I have tremendous respect for you sir, and have found your postings on this
> list to be uniformly excellent in logic and content.
> 
> However, I cannot read the italicized comments between the sections of the
> OSD, which contain phrases like "users have a right to know who is
> responsible for the software they use", and "Distributors of open-source
> software have the right to make their own choices about their own software."
> without drawing the conclusion that the OSD codifies a set of ethical
> principles.  Discussions of rights are inextricably linked to discussions of
> ethics.
> 
> I must therefore disagree with your interpretation of the content of the
> OSD.  It is as much an ethical framework as the Declaration of Independence
> is.  And like the Declaration which heavily influenced the contents of the
> Constitution which resulted from it, the OSD has to acknowledge that the
> ethical framework it espouses will be encapsulated in the licenses it
> inspires.

You're right.  I guess that OSI is straddling a somewhat anomalous
position: building a pragmatic case based on an ethical stand.  I
guess that's kind of what RMS says, too: that the OSI, by talking up
the practical benefits, neglects the ethics.

Ian