Subject: Re: MS continued misconception on OSD #6
From: "B Galliart" <bgallia@gmail.com>
Date: Thu, 27 Mar 2008 18:24:49 -0500

Quoting Tzeng, Nigel H:
> But I would think that "open source" as a generic term (no TM) is a little more inclusive
of code that is exchanged so folks can do research and learn from it.

It is hard to support a "generic" definition without it also creating
misconceptions about the proper definition.  While a more loose
definition would include more code, the additional code is still under
exclusionary terms.


Quoting Tzeng, Nigel H:
> Sorry, but I can't get too worked up that MS is releasing code for AIDS research but
not letting someone turn their contributed code into a business.

Normally I would agree with you.  It is the fact that MS has spent the
last 6 months providing a document that claims there are benefits to
providing these tools under open source.  I agree with Microsoft's PDF
that there are merits in providing MSCompBio under a open source
(inclusionary) license.  My disagreement is with the bait and switch
of then using a MS-OSD exclusionary license.  If MS want to advocate
that MSCompBio project benefits from being open source then we should
be able to demand they follow their own advocacy.  Anything less
should be a red flag towards MS credibility.


Quoting Tzeng, Nigel H:
> I guess they could call it "shared source" or something but there are objections to
that as well.

I already tried calling it shared source since MSR-LA is similar to
some SS licenses in it exclusionary method.  Jamie Cannon quickly
objected to the SS term being used for MSR-LA.  Hence, the only term I
can think of that covers all of Microsoft's authored licenses on
Codeplex that contain non-commerical terms is "MS-OSD."


Quoting Tzeng, Nigel H:
> And where but on CodePlex are they going to stick it anyway?

Considering that other non-commercial licensed source-available works
are provided for download outside of Codeplex (such as the Windows CE
code), I would suggest they treat their other non-commercial works the
same way.  Otherwise, what was the whole point to Microsoft employees
claiming on this mailing list that MS would be willing to take steps
to make a distinction between OSI approved and non-OSI approved
licenses?


Quoting Tzeng, Nigel H:
> You refer to it as a continued attack.  Presumably intentional rather than "misconception".

I think the MS PDF claiming MSCompBio to be open source might have
been created and posted as part of a growing misconception at MS.  I
think allowing that misconception to continue to grow and leaving the
document and license both unchanged after the January wave of updates
was intentional.  Even if someone attacks you because they don't know
better doesn't mean it stops being an attack.

Even if MS is failing to make a distinction between OSI-approved and
not approved licenses out of misconception, their ability to honor the
claim that such a distinction would be provided ends up still being an
empty offer.  The resulting bait and switch still remains in place.


Quoting Tzeng, Nigel H:
> Well, one of MS' responses was actually correct.  Academic sharing of code predates
both "free" and "open" source.  Often there was a non-commercial requirement so it would
more or less stay unencumbered from financial issues.  No commercial value and no one
really cared if you gave it away.  If there is a commercial value you tend to get bean
counters involved and sharing is reduced.

And if that was the point the MS PDF document about MSCompBio was
making then I would accept it.  But it does not say anything about the
merits of a non-commercial requirement.  Instead, consider what MS
does say:

> And open source is already established as a successful approach through which distributed
researchers can share and collectively understand data.  "Open source is very common
in biomedicine," Heckerman says.

Jamie Cannon back-pedals on this document's point by claiming that
non-commercial terms is common in research.  But the document never
states any benefit to non-commercial terms and never claims that
non-commercial terms are common.  Instead, it favors instead referring
to open source as "very common."  Despite this, Microsoft still has
yet to release MSCompBio under an OSI approved license or submit the
MSR-LA for OSI approval.


Quote Tzeng, Nigel H:
> As I said, if someone from the OSI asked I would think they would update their outdated
FAQ.  Which likely has a bunch of other errors too since it seems unmaintained.

It will be interesting to see if that is true.  While members of
Microsoft clearly read this mailing list, there does not seem to be
any such offer being made.


Quote Rick Moen:
> Here's an example of a bogus "public domain" offering, Wabburami, which is is David
M. Archer's variant of the quaint 1970s console game "Hamurabi":   http://sourceforge.net/projects/wabburami

Thank you for pointing this out.  It sounds like SF is taking the
stance of "common carrier" in this case.  I could see ways in which
this could also be problematic for the FOSS community.  But, the
example differs from the MS case in that David M. Archer is not a SF
employee.  MS is directly contributing to the Codeplex deviation
towards exclusionary licenses.


Quote from Wabburabi package (provided by Rick Moen):
> 3.  An unnamed coder converted that to C.  That somehow ended up at   ftp://ftp.sco.com/skunkware/
.

I guess Darl McBride was correct when he said that there are
people/companies that are attempting to leak other people's IP into
open source.  Considering that SCO claims skunkware is already open
source and freeware, it seems the leaky company is SCO itself.  How
ironic is that?


Quote from Rick Moen:
> He mentions the "AIDS Vaccine Research Tools" (four codebases that include MSCompBio)

MSCompBio isn't one of four parts to AIDS Vaccine Research Tools.
MSCompBio (PhyloD, Epitope Prediction, HLA Assignment, HLA
Completion)[1] is all of the AIDS Vaccine Research Tools described in
the PDF document.    As such, MSCompBio isn't just included in the
"AIDS Vaccine Research Tools," it *IS* the full set of four tools.

[1]http://www.codeplex.com/MSCompBio


Quote from Rick Moen:
>The slide also mentions in passing Windows CE.  Are we supposed to
think Hilf was asserting that
Windows CE is open source, too?

The slide is supposed to show how Microsoft has "changed" over time
such that it "is making broad-reaching changes to ... drive greater
interoperability."  Windows CE and Shared Source Programs are towards
the beginning of the slide.  Also, Microsoft seems to acknowledge that
Windows CE is not open source and it is not included as part of
Codeplex.  But as we move forward in the slide, it seems implied that
there should be charge towards greater interoperability with open
source.  I don't see any such change.  Windows CE was under an
exclusionary non-commercial license and MSCompBio is also under an
exclusionary non-commercial license.  Both equally violate OSD #6 and
is not a change towards contributing under the OSD terms.

As already point out by other, by itself this issue isn't that big a
deal.  However, as part of a trend it is problematic that MS is
operating on the assumption that a non-commercial restricted method is
somehow interoperating with the FOSS community.  Without an agreement
on what open source is, the interoperability structure that MS is
building will be built on a broken foundation.  The result will
continue to be situations like the one Michael Tiemann objected
too.[2]

[2]http://opensource.org/node/257