Future of open-source (Re: Windows virtualization and Linux)

[Rich Braun]

Jerry Feldman wrote:
> But, I think we have pretty much shown that the OpenSource model is not
> only economically viable, but also can produce better products. On the
> other hand, a purely community based product can have some downsides,
> such as lack of continuity. Let's say we have an open source product
> that was supported by a developer, but the developer loses interest, and
> no one else steps up to the plate.

Open source has changed *so* much since the era that spawned "rock stars" of
the movement (think RMS, think Linus).  Is there any way for someone to
achieve personal recognition by launching or contributing to an open source
project these days?  It seems to have gone the way of the Silicon Valley
garage startup:  too many millions of people have crowded what was once a
wide-open field.

Open source isn't dead by any means but I don't think it's safe the rely on
assumptions that what worked in the past will work similarly in the future.

At my job, I'm shocked at the Total Cost of Ownership calculations for the
Linux systems here.  They cost us about 3 times as much as Microsoft!  Why is
that--because the QA departments at application vendors only have the
resources to validate their offerings against one or two vendors.  If we buy
an application and a maintenance contract that's only valid on Red Hat
Enterprise Linux, we're locked into Red Hat.  The companies which bubbled to
the top of the open source hierarchy have figured out how to play this
anti-competitive game of squeezing out alternatives very well.  I'm therefore
stuck paying Red Hat an astounding percentage of my cap-ex budget:  every
year.  Sun MicroSystems bought MySQL AB, recognizing that the same economic
rules apply to their database software.

On the flip side is the contribution question.  I and my team are building
non-proprietary tools on top of open source tools.  We could publish them as
our contribution to the open source community.  But we face two very real
costs if we do so:  (1) QA verification--taking stuff that's too tightly tied
into our systems and testing it on generic/vanilla configurations that will
work for everyone else; and (2) customer support--answering emails about the
stuff we developed.  The latter is what I ran into when I wrote an open-source
file server package back in the 1990s:  the email got burdensome.

I'm exceedingly happy that RMS got his wish for an open-source world.  Now
that we're there, though, this organizational model will have to evolve into
something that I, RMS and the early organizers of the BLU could never have
envisioned back in the '80s.  Total Cost of Ownership (TCO) is the primary
challenge it faces, primarily from corporate entities that drive market-share
for the most popular open-source products.

-rich