Editor's Note: A League of Our Own
Feb 02, 2007, 23:30 (5 Talkback[s])
(Other stories by Brian Proffitt)
Re-Imagining Linux Platforms to Meet the Needs of Cloud Service Providers
By Brian Proffitt
In 1994, my eldest was one year old, and my wife and I were in
the process of re-locating from New Jersey back to Indiana. So
things like watching television got put by the wayside while we
uprooted ourselves and got things in order once more.
This is the reason I missed the initial airing of Ken Burns'
documentary Baseball on PBS. Nine nights of two-hour
episodes was just a bit much, back then, even though I am a
This week, I have finally got around to watching it, having
checked it out from the local library. This is, as I wrote, a
lengthy documentary, full of home-spun tales about the early days
of what the movie deems "America's sport." I recommend it for
anyone remotely interested in sports, and now that the first few
episodes are under my belt, I can pretty much recommend it for
anyone remotely interested in free software, too.
Each of the episodes in the series deals with a decade's worth
of history (though episode 1 handled the 1840s- 1900). Having just
finished episode 3, I am up to 1920, when eight players from the
Chicago White Sox were banished from the game forever for having
thrown the 1919 World Series against the Cincinnati Reds to get
some extra money from gamblers who were naturally very interested
in knowing how the series would end.
It was the story of this fix (popularized by the movie Eight
Men Out and somewhat by Field of Dreams) that jogged
my memory when I read the headline from InformationWeek, "Microsoft
Exec Wanted To Mask Linux Report Sponsorship, E-Mails Reveal," this
morning. The fix, I thought, was still alive and well.
Beyond single incidents of shame that mark the baseball and
software industries, there were broader historical themes in the
movie that compelled me draw more comparisons with free and open
Specifically, during the early years of baseball, there were
many leagues established to manage and standardize the rules for
groups of teams, until finally it settled into the National League
in 1876 and the American League in 1901. There were, of course,
other leagues, including--most notably--the Negro National League
and the Eastern Colored League, and a whole slew of "minor" leagues
that included smaller towns and private corporate teams. Though
there were many of these leagues, they did not really interfere
with each other, because of geographical and, stupidly, racial
Even the two main competitors, the National and American
Leagues, ended up being in what we now call "coopetition." When
owner Ban Johnson changed the name of his Western League to the
American League in 1900 and expanded into NL cities in 1901, the AL
became the first major threat to the older NL.
The two leagues initially fought hard over their resources--the
players--so they could gain the most revenue from the fans. It was
rather ugly, with players defecting from one league to the other
based on the best salary offer of the day. The fans were
unimpressed, and overall attendance began to drop.
Sensibly, the owners in the two leagues decided to work together
and form a single, overarching, three-man commission that would
govern disputes between both leagues. (Johnson, conspicuously, was
on that three-man commission and was the de facto leader.) Thus,
major league baseball was born.
The story is rife with parallels to the software industry,
particularly open source. I could not help but be reminded that the
many leagues were almost like today's Linux distributions, with
good and bad reasons for existing. There were even the big two
leagues, just like the big two distros, Red Hat and SUSE. The
players could represent the developers in today's scenario (though
without the indentured servitude of baseball's early reserve
contracts, developers are much more free to move from one
project/job to another).
Distros are like leagues. They all play baseball, but they each
have their own rules. Fans (the users) can choose between them
based on where they are and what kind of baseball they like to
watch. I, for instance, like a good minor league or college game.
The lack of hype and "polish" tends to feel more authentic. Some
leagues are based on others (like many of today's Debian-based
You can take this analogy or leave it, but while I was
ruminating it this week, I heard the story of the short-lived
Federal League, a group of eight teams bankrolled by some wealthy
businessmen eager to make some money from the huge baseball
phenomenon in the early 20th Century. The Federal League set up in
1914, lured some players over who were sick and tired of fixed
salary scales and reserve contracts, and got running. Immediately,
the fan base was divided into three and the other two established
leagues saw an attendance decline.
Ban Johnson raised his leagues' salaries from some of the star
players and threatened to blacklist any player who went over to the
new FL. The FL fired back with an anti-trust lawsuit, but the judge
held off of his ruling, expecting the FL's financial troubles to
force them to settle out of court. By December 1915, that's exactly
what happened. The FL disbanded, and attendance went back up in the
other two leagues.
(Baseball-as-monopoly tends to bring up an analogy to another
certain software company, In know. That's the trouble with
analogies. They aren't perfect, and can be interpreted a bazillion
ways. But, for now, I'll stick with the league/distro model.)
The reason why I am going over the history lesson today is
because of all of the buzz in punditville lately regarding the
sheer number of distributions out there and how Linux needs to
unify. The story of the Federal League tends to support that: too
many leagues put a strain on the fan base and the resources each
league can have. Do we have too many distributions to effectively
support Linux as a whole?
If the number of development resources (i.e., players) were more
limited, then I would say yes. But the numbers are greater, and I
think each project will be able to attract the players it needs to
get the job done. The open source nature of the distros also means
that what's done in one will usually be available in the others.
Leagues (especially in the early days of baseball, tended to be
more proprietary in nature).
There is, I think, a slight risk that all these distributions
are diluting the user base. Less distros would mean more users per
distro and therefore more homegrown marketing and feedback support.
I think that dilution is not much of a concern.
Still, with more corporate funding coming into the open source
field, I think natural evolution will trim down the number of
distributions without any active interference. Simply put, most
developers will go where the money is, and if that means delaying
work on Brian's Linux Distro 1.2 to work on Red Hat, Ubuntu, or
Mandriva for some money, then that may be the way things will
So I wouldn't worry about to much diversification. The natural
ecosystem of the free and open source software community should
settle that issue on its own. Like the slow progress of baseball
over the years, open source will ultimately come out the winner.
And so will we fans.
Program Note: Due to my inattention to the entire
article from SingleHop on RHEL and CentOS posted earlier this
morning, I did not see it was an advertisement for services for
that company. I regret the error, and I have pulled the story.
Sorry for the mistake. -BKP