Thoughts from Oscon: Why diversity is annoying and assholes run large corporations

Oscon is over, I'm home and recovered both from jet lag and just general exhaustion.

Oscon is a very broad conference so there is a lot to learn and many people and projects to befriend. There are many things and angles one could write a blog post about. To me Oscon is above all the conference to meet other open source people and have the deep and inspiring discussions. So in that spirit I will make a few philosophical remarks in this post, thoughts from Oscon 2013.

If you'd want to read a run-through of the conference itself, I recommend Dirk van den Poel's very extensive summary.

My "day job", so to speak, was to man the MongoDB booth in the expo hall. My collague Francesca Krihely blogged a few observations from that point of view.

Community Leadersip Summit

Before Oscon I attended the Community Leadership Summit - my third.

On Saturday I moderated a session Why do foundation governed projects grow bigger than others?. It was simply based on sharing statistics I have published previously on this blog (they are linked from the session notes), but they were news to many of the attendees and sparked an interesting discussion. Josh Berkus made an interesting comment suggesting that maybe PHP never grew to be as large as Perl and Python precisely because they never formalized the governance by incorporating (or at least joining) a foundation?

For the last session slot on Sunday, when everyone were suitable tired, I proposed a philosophical discussion about Post Modern Open Source. This was inspired by my observation when joining 10gen that it's just me and Laura using a Linux laptop and everyone else uses Apple. The other entry point to the discussion was the ongoing discussion about the fact that most Github projects don't apply any open source license at all.

The very philosophical conclusion of this session was that what we see happening to open source is what happens to any religion or ideology as it matures and becomes more mainstream. Those of us who joined the open source movement in the 90's are the fundamentalists who would suffer persecution while defending their pure beliefs. Otoh the open source developers that use Macs are like Catholics that use condoms.

Thanks to everyone who joined and contributed to these sessions. Especially the second one I had low expectations for myself, but the attendees made it a truly great session!

Why is diversity so annoying?

Two of the Thursday morning keynotes went into (pretty much) recommending permissive licensing over copyleft licensing. As you can imagine, it gave people something to talk about for the rest of the day...

I made the observation that diversity is actually an important source of strength in the open source ecosystem, and in fact that is true for any ecosystem. Diversity is how species succeed and survive in a biological ecosystem too, and so forth. It's a law of nature. The same must obviously be true for open source licensing: a certain amount of diversity is not a problem, it is a core ingredient for a successful ecosystem.

The desire to agree on "the one true license" is therefore weird, unnatural even. However, it is very common. When you think about it, we humans instinctively dislike diversity. Racism is the dislike or even hatred of people with a weird skin color, culture or religion. I personally once caught myself annoyed at a group of people talking a foreign language - this is quite understandable: I was annoyed about the fact I had no clue what they were talking about. We go into a communal shock when someone forks a code base, and we are frustrated to have so many licenses to choose from.

It's an interesting observation. Rationally we know that diversity is a core strength for a community, but there's something in the human psyche that would like everything to be very simple and uniform and that makes us annoyed at diversity.

Why the most successful IT companies are run by assholes

One of those days I also had the opportunity to sit down with Florian Haas of Hastexo. I learned that his company now offers support for OpenStack - I came to know Florian as a DRBD and Pacemaker expert, so this was news. (Not that I need OpenStack support, but it was just interesting to hear what people are working on...) Talking about living the life of an entrepreneur and growing Hastexo, we talked about leadership, and at some point I ended up sharing my thesis about challenges in leading really large corporations.

It's my belief that the number one challenge for a large corporation is to maintain discipline, or like Stephen Elop put it at Nokia, "accountability". Discipline is needed to fight the inevitable decline towards mediocrity and complacency.

For example, when we (at MySQL) were acquired by Sun Microsystems, I found out that Sun was full of people that believed (and they still today believe) it was the greatest IT company in the world, who had basically invented the internet, or at least the dot-com part of it. Maybe that was true sometime back in the 80's, but it certainly wasn't true in 2008.

It's not childs play to do a proper benchmark, but it's not that difficult. Many of Sun's customers did, and we constantly got questions why Sun was recommending a Solaris+Sparc server that was twice as expensive but only half as fast as a modern Linux+Intel server. You know, it's not difficult to objectively conclude that Sparc was sucking and had been for years. I've heard from two different people they could get better benchmark numbers out of their Intel desktop or laptop than from a Sparc server.

Yet inside Sun there was still a strong belief that Sparc was superior to anything else. That the rest of the world was enchanted by Intel and Linux servers was just some anachronism in time that would eventually be corrected, as long as Sun just continued to invest more and more in Sparc engineering and marketing. This attitude was nothing else than a lack of discipline and professionalism and it was what eventually brought down Sun.

In a large corporation, there are several structures that nurture mediocrity and lack of discipline. For example, when a manager does the annual performance review of his team, he is strongly biased to come to the conclusion that his team did a great job. After all, if his team is great, that makes him look good as a manager too. You'd think that such inflation would get caught by the next-in-line manager, but no. The same is of course true for him! He is biased to believe that all of his reports are great managers, in charge of great teams. And so mediocrity and even below-mediocre performance is protected and nurtured year in and year out.

The antidote to mediocrity in a large corporation is therefore discipline. The leadership must spend tremendous amount of energy on fighting mediocrity and instilling discipline into every level of their organization.

As proof for my thesis I proposed that there have been 3 really successful IT companies: Microsoft, Apple and Oracle. All of whom, in their heyday, were lead by assholes. (Bill Gates may not be as universally an accepted asshole as Steve Jobs and Larry Ellison are, but according to some accounts his management style was inquisite to the point of being a bully.) Being assholes they are in a stronger position to maintain discipline in their companies, which is - if you believe me - what has made their companies so hugely successful.

Florian was clearly intrigued by this - or maybe he was just amused - and continued to debate this thought with people that joined us at the Burgerville.

It's not that I don't see Hastexo becoming even more successful in the future it's just I can't imagine Florian being an asshole! :D

Around 2004 I was doing performance tests for a new sort algorithm and tried it on as many different types of CPUs as I could access. I had a 6 year old Pentium 3 desktop -- 600 Mhz, maybe 128kb L2 cache and borrowed a brand new Sun server -- maybe 900Mhz UltraSPARC IV, 8MB L2 cache. The sort ran much faster on the much older Pentium 3. Sad, but my experience matched yours.

About the bookAbout this siteAcademicAccordAmazonAppleBeginnersBooksBuildBotBusiness modelsbzrCassandraCloudcloud computingclsCommunitycommunityleadershipsummitConsistencycoodiaryCopyrightCreative CommonscssDatabasesdataminingDatastaxDevOpsDistributed ConsensusDrizzleDrupalEconomyelectronEthicsEurovisionFacebookFrosconFunnyGaleraGISgithubGnomeGovernanceHandlerSocketHigh AvailabilityimpressionistimpressjsInkscapeInternetJavaScriptjsonKDEKubuntuLicensingLinuxMaidanMaker cultureMariaDBmarkdownMEAN stackMepSQLMicrosoftMobileMongoDBMontyProgramMusicMySQLMySQL ClusterNerdsNodeNoSQLNyrkiöodbaOpen ContentOpen SourceOpenSQLCampOracleOSConPAMPParkinsonPatentsPerconaperformancePersonalPhilosophyPHPPiratesPlanetDrupalPoliticsPostgreSQLPresalespresentationsPress releasesProgrammingRed HatReplicationSeveralninesSillySkySQLSolonStartupsSunSybaseSymbiansysbenchtalksTechnicalTechnologyThe making ofTransactionsTungstenTwitterUbuntuvolcanoWeb2.0WikipediaWork from HomexmlYouTube