6.6 KiB
Technology Trumps Dogma, And Other Open Source Insights
Marten Mickos is one of the most respected leaders in open source. Here's why.
A few weeks back I asked Marten Mickos (@martenmickos), CEO of Eucalyptus Systems, to comment on the changing face of open source. He did, and with the usual Mickos style. Unfortunately, a whole lot of great commentary had to be cut for space reasons.
Given the brilliant insights Mickos offered, I wanted to share his comments in their entirety here. Mickos helped make MySQL arguably the most popular database on the planet, and is trying to achieve similar success with an open source cloud offering.
With a string of successes—and failures—under his belt, Mickos had multiple pearls of open source wisdom to share. For instance while open-source developers have long eschewed corporate influence on open-source projects, Mickos starts by reminds us that money is critical for funding development, not to mention marketing, documentation, etc. The myth of a peace-loving, cashless open source existence is just that: a myth.
On the importance of money to open source...
Without money, open source will die.
On the foundational principle behind open source business strategy...
Some people will spend any amount of time to save money. Some will spend money to save time.
On the changing face of the open source developer...
Back then it seemed that open source developers were true cowboys—out on their own, following their own individual paths, valuing their nearly unlimited freedom. Today, many open source developers are happy to be salaried employees of companies that don't really stand for open source on a corporate level (Google, HP, IBM, Oracle, etc.). When they make public presentations, they have to state that what they say is their own opinion and not necessarily an official statement of the company they represent. There is a voluntary submissiveness today that wasn't as common before.
On the role of copyleft licensing and governance...
The purpose of the FOSS license and the governance model is not really to enable like-minded people to collaborate, although that's a benefit too. It's about enabling unlike-minded people to collaborate. The beauty of open source is that people who dislike each other can produce code for the same product.
On leadership...
Even in a meritocracy, even in peer-production models, people look for leaders.
On critical feedback...
If you, on a sustaining basis, can truly love harsh feedback and if you can truly show enthusiasm and appreciation for contributions of whatever magnitude and type, you can be wonderfully successful in open source.
When people complain about your open source project, you need to hear them as saying "I would love to love you, but right now I cannot."
If nobody is opposed to your open source product/project, you are not really being popular. [This jibes well with my own observations of haters being a leading indicator of success.]
On the role of branding...
More than a question of licensing, it's a question of branding. Red Hat took their open source brand "Red Hat" and made it commercial only. Then they established Fedora as the non-commercial brand. MySQL and JBoss did the opposite: they kept one unified brand for both community and commercial use. When you fork, you must use a different name, because branding is not included in the open source licenses.
On apparent inconsistencies in open source "theology"...
Open source people can be dogmatic, especially about others. They will eagerly demand that some project behave in this or that way for reasons of orthodoxy and purity. But they will at the same time merrily use closed systems such as iBooks because they admire those products. Technology trumps dogma. Coolness is key. All of this I say not as a complaint, but as an observation. To succeed in open source, you must learn to live with it and make the most of it.
On changes to open source in the past 10 years...
People didn't know what it was, how it worked, why people did it, how it could produce great software, why it wouldn't self-die, etc. That's why the LAMP stack made it onto the front page of Fortune Magazine—it was so new and intriguing. Today people know open source and they know it's an essential part of the software world. Incumbents fought it. Now they embrace it (or at least pretend to).
Those who did open source just did it. There were very few people blogging about the meaning of open source, thinking about the business models, etc. Today you have those who code, those who lead communities, those who test, those who use, those who make money, those who write about it, etc. Licensing was a big issue then, for good reasons. Now it's much less of a topic.
Back then it was relatively few projects with relatively few people in them. Today there are probably 100-1000X the number of projects.
Back then the infrastructure didn't exist. Today we have Wiki, Github, Jira and other services that make it obvious how to run and govern an open source project.
Ten years ago people would download distributions. Now they upload images (to the cloud).
On what hasn't changed in open source over the past 10 years...
Still a lot of unbridled enthusiasm, often bordering on naïveté—with all the amazing upsides and inevitable downsides that this will bring. Open source still attracts outstanding talent.
The most successful open source projects are those that target developers. Products that are supposed to be used by consumers or other non-technical people generally don't do as well. But there are notable exceptions, as always, such as Firefox, Android and perhaps OpenOffice.