[PLUG] Forbes: Is Linux For Losers?

Geoff Burling llywrch at agora.rdrop.com
Sat Jun 18 19:10:00 UTC 2005


On Fri, 17 Jun 2005, Paul Heinlein wrote:

> On 6/17/05 12:29 PM, Steve D... wrote:
>
> > Is Linux For Losers?
> > http://www.forbes.com/business/businesstech/2005/06/16/linux-bsd-unix-cz_dl_0616theo.html
> >
> > Theo's a smart guy.  You'd think he'd have enough sense to tone it
> > down a bit for the general media.  It doesn't look good for either
> > the BSD or Linux communities.
>
> I've had *BSD guys give coherent, reasonable explanations why they
> dislike the Linux codebase -- although I think they typically base their
> judgements on stuff they saw back in 2.0 and 2.2 days. They can rightly
> point to some driver code that looks as if it were thrown together by
> amateur programmers during a late-night Jolt session.
>
> The thing is, reasonable folks are usually able to articulate the
> trade-offs: a focus on broad driver support, rapid development, and
> world domination (Linux) vs. stability, audited code, and a much greater
> willingness to be a niche player (*BSD).
>
Sounds like the old Best is better vs. Worse is better arguments that have
been floating around the computer industry for at least a couple of
decades.

The dispute can be summarized as follows:
* Best is better -- you have a group of programmers who write perfect
code that gets relased within a reasonable time frame. The code is
elegant, matches the spec precisely & works with a minimum of flaws.
* Worse is better -- you have a group of programmers who write code
that does most of what you want, & is released fairly rapidly. The
code may be ugly, doesn't match the spec (if one exists), & often has
major problems -- however, using feedback from users (e.g., comments,
proferred code), the code improves.

(I've probably mangled these 2 positions; I'm writing all of this from
memory. Feel free to Google on these phrases & correct what I've written.)

Each POV has its strengths & weaknesses. "Best is better" is defended
because -- well, the code works like it's supposed to: it's solid,
it was written following best practices, & it was written by experts.
"Worse is better" is defended because of speed of deployment (amongst
other points), & the fact that with user input or feedback (analogies
here include the metaphor of the marketplace, & genetic algorthyms)
the code gradually improves in the parts where the users need it to
improve.

Also, the "Best is better" model requires a way to pay its programmers,
who for some reason need a paycheck & benefits; the "Worse is better"
model can be carried to a large degree by volunteers & part-time
amateurs who don't need a paycheck -- although IIRC they don't refuse
the offer of money. Unfortunately, volunteers can be a flakey lot, &
that is why there are a lot of dead projects on Freshmeat & other
repositories of free software: they died due to lack of interest, the
maintainer got bored/went into another line of work/died.

The argument over these 2 approaches will go on until the end of
sentient life because both produce useful software -- just in different
ways. In other words, it's a religious war: you either embrace one
dogma & label everyone who disagrees with you a heretic (or worse),
or be an agnostic, & use what works.




More information about the PLUG mailing list