The marketplace’s 2.4 dilemma

36

Author: JT Smith

By Jack Bryar
Open Source business

Now that the Linux 2.4 kernel is finally actually here, the biggest
question
among commercial distributors is what effect this will have on their
business.
This may become another case where the rules of the Linux market are
different
than the traditional software marketplace. If so, many commercial
vendors
may not think much of those rules.
One of oddities of the Linux boom of the last year or so was that the
Linux
kernel itself hadn’t evolved along with the marketplace. It all seemed
rather
odd. In most commercial software environments, any meaningful growth in
the
market requires frequent upgrades of the software. And yet, despite the
explosion
of new applications and complex distributions over the last year or so,
not
to mention a series of new vendors, the core Linux kernel hadn’t
changed a
bit.

More than one of the various mainstream info-tech pundits I read
commented
on the phenomenon. One suggested (despite the lack of any evidence)
that
the Open Source marketplace had been relegated to a kind of holding
pattern
because of the delayed development of the kernel. Others suggested that
while
the market was growing it was doing so despite the lack of improvements
being
made to this critical core element of the platform.

Now, crabbing about 2.4 , and its delays and its functionality is not a
new
phenomenon. I’ve written previously about rumblings among some listserv
and
user group members that Linus Torvald was insufficiently concerned
about

a generating a timely 2.4 roll-out, or he was insensitive to their
various
requirements. Many of these gripes seemed focused on tweaks needed to
transition
Linux into a platform for back-end enterprise applications. Many more
groused
about repeated delays of the kernel’s release date. Over the last
couple
of months, the rumbles grew louder. Wired even put 2.4 into its list of
the
great
vaporware
non-products
of the year 2000.

While enthusiasts pointed out that, for a vaporware product, 2.4 has
been
available (in beta purgatory, admittedly) for months. But it takes a
brave
corporate IT manager (or one with a sizable trust fund) to hang key
applications on a platform that hadn’t been officially “released.” So
conventional wisdom suggested that getting out a “final” release would
have
a major impact on Linux adoption by larger corporations.

Last week, the wait was over. Torvald announced that 2.4 was now “out
there.”
Why? In an email, Torvald suggested the release was a “sign of
incipient
brain damage,” and that he released the kernel because “things don’t
get
better from having the same people test it over and over again.” What a
kidder.

Or maybe not. Based on a look at the complaints and patches featured on
one

mailing
list
, the definition of “release” in the Linux world varied
considerably
from what one might have expected in a commercial software environment.
Complaints surfaced almost immediately concerning device drivers and
support
for various controllers, networking devices, etc.

This type of thing may not matter to core enthusiasts. For one thing, a
lot
of the complaints turned out to be bad, homegrown code that
was
exposed in 2.4 And despite telling new users not to “bother reporting
any
bugs for the next few days. I won’t care,” Torvald did mix it up in
person
with several of the more confused would-be adopters. It was also pretty
clear
that many of the “complaints” were easily fixable.

Nevertheless, the whole release process must have felt very foreign to
many
corporate IT managers, and may have done very little to generate
enthusiasm
for early adoption.

A more important problem may have to do with the other, related
softwares
that make up the suite of applications that drive adoption of the new
platform.
As commercial developers, particularly Microsoft, learned long ago, a
new
operating platform isn’t enough. In order to generate demand there also
needs
to be a set of new or “improved” programs that make full use of the new
platform.
Such co-releases are critical to hyping demand for either product set.

In fact, it’s critical to avoid “cratering” demand altogether. The
prevailing
wisdom among commercial software developers is that is that
uncoordinated
releases can actually retard demand and severely damage a product
rollout.
As the bits and pieces of a new platform start to come out, demand for
the
old platform evaporates, and yet few buy the new platform until all the
pieces
of a complete solution are available.

So will the rollout of 2.4 actually cut into demand for commercial
distributions
like Red Hat and SuSe? It’s a little early to tell, although most
testers
I talked to found that there were relatively few problems integrating
2.4
into existing distributions. Still, will corporate customers wait until
new
distribution packages are ready from the major vendors before they
adopt
the new platform or buy new software? We’re likely to find out soon.

NewsForge editors read and respond to comments posted on our discussion page.

Category:

  • Linux