[UCLA-LUG] open-source

[Neil Nelson]

Anne Holohan wrote:

> I am researching a newspaper article on open-source programming. I would
> like to hear from a Linux programmer on why he or she spends a portion of
> their time coding for free.  What motivates you to work on open-source
> code? Two or three sentences, along with your name and regular/paid
> position (e.g. computer science grad student) would be very helpful.

This is a very good question.  There appear to be several factors
involved.

(1) Many of the contributors are university based such that their
funding for their open-source contributions--payment or motivation--
derives from the several academic possibilities.  FFTW (Fastest Fourier
Transform in the West), a superior FFT package, was assembled by two
advanced students with additional funding from, I believe, a
government agency and possibly a large corporation.  It becomes a
form of academic research that is published on the Internet
instead of in a journal.

(2) Companies trying to gain market share or to set a standard in some
area, may use open-source to attract adherents (market share).  When
a product is free and works well, it will be adopted more thoroughly
than a similarly capable product having a price.  E.g., if you wanted
to sell applications based on a particular operating system or
language, you might make those core pieces open-source and then charge
for the additions.

(3) And then similarly with (2), a basic or new application from a
company might be started in open-source and then later improved with
proprietary additions after some market share had been obtained.

(4) In looking at the different projects at SourceForge, there are a
large number for which someone, apparently, only had an idea and started
a project without any additional work being done.  And then there are
projects in varying states of completion with a few in working order.

Many people have programming hobbies and personal interests that they
spend time on because it is something they like to do.  And many of
the projects at SourceForge seem to fit in that category.  And a
number of the projects are in support of essentially non-computer
hobbies; e.g., a database to keep track of CDs.

(5) The use of open-source contributions to show computer abilities
by job-inexperienced programmers that will give them greater
credibility when applying for a programming job.  E.g., if a program-
mer organized and completed a project at SourceForge, any potential
employer could log into SourceForge and quickly see that applicant's
abilities.

(6) There is a growing recognition by many paid software organizations
that certain open-source projects or applications obtain a high-level
of quality because of the very wide review of the source code.  E.g.,
since the source for the Linux operating system is available to anyone,
bugs can be addressed by anyone and the solution made generally known;
whereas with proprietary binary distributed code, only the vendor can
effectively address such problems.  A paid software organization may
then be willing to use paid software staff to advance a common,
widely supported open-source application because they can trade-off
the reduction in software acquisition and bug expense required in a
proprietary solution against the contribution expense in an open-
source application.  I.e., it is not necessarily the case that open-
source is provided by non-company-paid programmers.

(7) As with, say, mathematical knowledge in general, open-source code
obtains a certain fundamental status because it can be treated as a
standard reference.  I.e., proprietary code is, by definition, not
easily referenced, except perhaps when a monopoly in the software
market occurs, because it is hidden behind a proprietary veil.
Software is more-and-more commonly built on standard components, and
it is more easily built when those components are freely known with
an absolute minimum of copyright restriction.  This fundamental
aspect of software causes many software developers to work in the
open-source arena irrespective of how they might be paid.

E.g., a company wants a software solution.  The software person is
paid for their time in providing the solution however that is done.
At some point the advantage in selling similar solutions to other
companies/individuals becomes merely the effort of implementing/
adapting the standard solution to the particular case.  The ability
to widely sell a standard solution is greatly diminished.  I.e.,
the software market is evolving away from the Microsoft model.
Software people will be paid for their immediate and short-term
add-on benefits and less for re-selling the same solution over a
long period.

I am not paid by anyone.  I am investigating a project to auto-
matically consolidate and then reuse source code from a variety
of currently separate areas (automatic programming under AI).

Neil Nelson