Modularity or: Open source turns on all the available brain power, full blast, on every problem, challenge, or opportunity

Ralph Johnson johnson at cs.uiuc.edu
Thu Oct 5 03:41:21 UTC 2006


This is a typical newspaper article, with enough bogus things in it to
make the whole article worthless.

On 10/4/06, Klaus D. Witzel <klaus.witzel at cobss.com> wrote:
> "Putting Open Source Development Under the Scope"
> Linux Insider (10/02/06); Lyman, Jay
> Quote:
> Computer science researchers at the University of California Davis will use
> a $750,000 grant from the National Science Foundation to examine the
> development of open source systems such as the Apache Web server,
> PosterSQL, and the Python scripting language.

I imagine this is true.

>The suspicion is that open
> source systems succeed where commercial proprietary programs fail because
> they avoid the developmental process where the speed of production is
> determined by the slowest contributor.

I don't believe that the people at UC Davis are dumb enough to believe
this.  No good development process will have the speed of production
determined by the slowest contributor.

> The case of Mozilla suggests that
> the modularity implemented by open source systems increase volunteerism,
> because anyone can contribute at any time.

Open source does not necessarily lead to modular systems.  I would
agree with the statement that modular systems are easier to work on in
parallel, but Squeak is open source and is not modular.

> UC Davis lead researcher and
> computer science professor Premkuma Devanbu says, "The belief in the open
> source software community is that open source turns on all the available
> brain power, full blast, on every problem, challenge, or opportunity."  The
> purpose of the study is to put such ideas to the test, in order to get
> empirical evidence.

Marketing.  Taken literally, it is clearly false.  Taken figuratively,
there is some truth in it, but it will be hard to test.

> Many stress that open source development benefits from
> the fact that contributors are not motivated by getting paid, and can
> choose what they work on.

Many people say this.  However, research has shown that most code
submitted to Linux is written by people paid to work on it.  Though
perhaps that does not mean they are motivated by getting paid.

>As no meetings and the lowest level of
> synchronization are necessary when using open source software, development
> can occur at parallel levels simultaneously, rather than requiring each
> step in the process to occur sequentially.  The researchers will monitor
> emails, message boards, and bug reports for insights into what makes open
> source development projects successful.  Devanbu says the case of Linux
> shows that modularity improves the quality of the software developed, and
> that "good design allows implementation to proceed with maximum parallelism
> and minimum synchronization and coordination."

I sure hope they are getting $750K for more than monitoring emails,
message boards and bug reports!

Open source is great!  I love open source.  However, I bet that a
careful analysis would show that the man hours that it takes to create
a system by open source is more than can be done by a crack
development team.  Of course, you are not paying for those hours, and
a lot of those hours are cointributed by people who are having a good
time doing it.  But mail is a poor communication channel compared to
talking face to face.  Lots of work on open source never gets put into
the repository because the people in charge don't think it is good
enough.

Open source works.  It doesn't work for everything, though.  Rather
than try to prove that it works, they should try to find out when it
works, and how to make it work better.

-Ralph



More information about the Squeak-dev mailing list