Open Source for/by Government

02 Jan 2018

Update: Barcelona is going all-open. Sounds extreme, but some times you’ve got to…

“You’ve got to spend money to make money”, I once confidently told a business associate, on the occasion of paying him a thousand dollars to manually clean some terrible data for me. In the event, I was right: that cleaned data paid for itself 10 times over in the following years.

I’m still the only person with a GIS file for 1996 BC elections results by voting area, and the jealousy is killing you.

Governments can play the game too, but it seems like they all end up tilling the same landscape. There’s no shortage of governments trying to create their own Silicon Valley clusters, usually through the mechanisms of subsidizing venture capital funding (via tax breaks or directly) and increased spending on R&D grants to academia. Spending money to “hopefully” make money.

There’s an under-recognized niche available, for a government willing to go after it.

@nayafia Thanks for Roads & Bridges and the rest of your oeuvre! I'm currently researching the question of gov funding of OSS from a Mazzucato / Porter cluster development / Tire Tracks point of view; may I ask for advice? I'm looking for success stories, preferably quantified.

— Meng Weng Wong (@mengwong) January 2, 2018

Venture capitalists are (understandably) interested in having their investments create “intellectual property”, that can be patented and monopolized for outsized profits. By following the VC model of focussing on IP formation, governments are missing out on another investment avenue: the formation of “intellectual capital” in their jurisdictions.

VCs don’t like intellectual capital because it’s too mobile. It lives between the ears of employees, who can change employers too easily, and require expensive golden handcuffs to lock into place. They can monetize intellectual property in an acquisition or public offering, but they cannot monetize intellectual capital.

Governments, on the other hand, understand that by investing in universities and colleges, they are creating intellectual capital that will tend to stick around in their jurisdictions (for all the public wailing about “brain drain”, the fact is that people don’t move around all that much).

Investment in open source technology is a potential gold mine for creating intellectual capital, but governments have been steadfastly ignoring it for years. There is also a big first mover advantage waiting for the first governments to get into the game:

• Instead of “buying off-the-shelf” for government information systems, build on existing OSS, or start OSS from scratch, using local talent (in-house or private sector).
• Deliberately build with enough generality to allow use in other jurisdictions.
• Become the first reference customer for the project. Send your local talent out to evangelize it. Encourage them to commercialize their support and services.
• Wash, rinse, repeat.

Is this risky? Yes. Will it result in some failed projects? Yes. Will it be more expensive than the “safe” alternative? Sometimes yes, sometimes no. Will it result in increased revenues flowing into your jurisdiction? Almost certainly, if committed to and carried out across a number of projects.

When the first library in BC adopted the Evergreen open source library software, they probably weren’t envisioning a Canada-wide open source cooperative, bringing support and consulting dollars into the province, but that’s what they did, by accident. When the Atlanta Public Library started the project, they probably weren’t thinking a local company would end up selling support and expertise on the software around the country.

There is no IP moat around open source projects, but there is a first mover advantage to having a critical mass of developers and professionals who have amassed intellectual and social capital around the project.

Intellectual capital isn’t just built in universities, and the private sector shouldn’t only be looked to for intellectual property. Let’s mix it up a little.

The BC government spends $9M/year on Oracle “maintenance”, basically the right to access bug fixes and updates from Oracle for the software we’re running. It’s not a lot of money, but it’s money being shipped straight over the border. Affilias, the “.org” top level DNS provider built their infrastructure on PostgreSQL – they spend a couple hundred thousand a year having some PostgreSQL core developers on staff. Same effect, different path.