Subject: Dwyer, FRB, Economics of Open Source
From: kmself@ix.netcom.com
Date: Sun, 28 May 2000 00:06:43 -0700
Sun, 28 May 2000 00:06:43 -0700
Gerald P. Dwyer, Jr., Vice President of the Federal Reserve Bank of
Atlanta, has published a draft article "The Economics of Open Source and
Free Software" [1].  It explores reasons, structures, and motives for
development of free software.

The article has been submitted as a discussion topic to Kuro5hin, an
online weblog I rather like:  http://www.kuro5hin.org/



After a preamble discussing the phenomenon and drawing a (slightly
inaccurate [2]) distinction between free and open source, the author
discusses Apache, and makes as a first point that most free software is
highly modular:

    The first proposition is that software programs can be broken into
    pieces and that communication between the programs can be relatively
    simple.  The programs at issue are not monolithic entities that
    must be completely understood by one person or even a small group
    of people.

This modularity characteristic is one that I've noted several times,
particularly in the July/August  IEEE Software  response to the editor
co-authored with Russ Nelson and Steven Turnbull of this list:

    "Open-Source Methodology:  Ready for Prime Time?",  Steve McConnell
    "Response:  Open-Source Methods:  Peering Through the Clutter",
    Terry Bollinger, Russell Nelson, Stephen Turnbull, and Karsten Self

    http://dlib.computer.org/so/books/so1999/pdf/s4006.pdf
    (Subscription required, copy on request from me).

Modularity (or lack thereof) has been suggested as a failing (or at least
significant obstacle) in the Mozilla project.  Others [3] have suggested
that the MozPL encourages poor program design by dividing software at the
file boundary, which makes for clean law but dirty code.  Well defined
application interfaces would be Don's preferred mode.

Dwyer continues to suggest that a direct-employment model (Collab.net?)
might be the most market-efficient mode of software development.
Without describing it as such, Dwyer also suggests a Nash equilibrium
argument for avoiding the "stupid tax" (mentioned earlier today by me on
this group), that is, it's more efficient and effective to contribute
patches.

In exploring the relative viabilities of free and proprietary software,
pp 11 and 12, there appears to be a typo in the last paragraph p11:

    Conversely, the more divergent the difference in programmers'
    marginal benefits and costs of writing various parts of the
    software, the *more* likely is open source development.

...by my scan, "less" is "more", or "open" is "closed".



-- 
Karsten M. Self <kmself@ix.netcom.com>         http://www.netcom.com/~kmself
  Evangelist, Opensales, Inc.                       http://www.opensales.org
   What part of "Gestalt" don't you understand?      Debian GNU/Linux rocks!
     http://gestalt-system.sourceforge.net/      K5: http://www.kuro5hin.org
GPG fingerprint: F932 8B25 5FDD 2528 D595  DC61 3847 889F 55F2 B9B0


--------
Notes:

[1] Proving once again that the OS/FS split has one verifiable economic
effect:  doubling the consumption of ink (real or virtual) used to
label the phenomenon.  This being the  Free Software  Business list, we
have no such resource-consuming dilemma to face <g>.


[2] Dwyer defines "free software" as "licensed under the GNU GPL".
Though something of a technicality, this bothers me on several points.

For starters, it excludes the software licensed under the GNU LGPL.

Second, we've just had a wonderful instance of a straw-man argument [4]
based largely in a redefinition of free as referring to price, not
liberty, which though Dwyer is careful to avoid, he still changes terms
as laid out by RMS.

Stallman's own definition is of the four freedoms of software, number
0-3, as found on http://www.gnu.org/philosophy/free-sw.html :

    ``Free software'' refers to the users' freedom to run, copy,
    distribute, study, change and improve the software.  More precisely,
    it refers to four kinds of freedom, for the users of the software

A Venn diagram on the "Categories of Free and Non-Free Software" shows
open source and free software as being overlapping, non-identical sets,
neither of which is a proper subset or superset of the other, and both
being subsets of "Free Download", which includes several classes of
proprietary software.

I have to admit that the distinction between free and open is fairly
fine -- I'm inclined to say that "free" is that which meets the four
freedoms definition, whereas "open" is that which has been formally
approved by the Open Source Initiative's board.  While theoretically
different, I'm not aware of any OSI compliant licenses which fail the
FSF four freedoms test.

The matter is a bit long to be treated in a footnote.


[3] Don Marti, post to the Silicon Valley Linux User's Group mailing
list.


[4] [5] Bertrand Meyer, "The Ethics of Free Software"
http://www.sdmagazine.com/features/2000/03/f4.shtml.  An utter piece of
misleading trash.  Several rebuttals of varying degrees of excellence
were written, I favor:

  - "It's about the code, Mr. Meyer, not the money", by Phillip Rulon
    http://www.kuro5hin.org/?op=displaystory&sid=2000/5/23/23310/1431

  - My own comment at K5:
    "False premesis, straw men, ad hominem, and other m[inor problems]"
    http://www.kuro5hin.org/?op=comments;sid=2000/5/21/3219/37999;pid=0;cid=76#76
    
[5] Footnoting footnotes is, naturally, to be avoided.  Considered
harmful.


["application/pgp-signature" not shown]