Open source and government – A 15 year adventure
Last week, there were a bunch of articles about the US government being newly interested in open source. Specifically, President Obama requested a report on the use of open source from Scott McNealy, chairman and former CEO of Sun Microsystems. Never mind if Scott McNealy is the best person to write such a report, what’s interesting for me is that open source has finally becoming so mainstream that the President is requesting a report about it.
Of course, open source in the US government is nothing new. Almost 15 years ago, I build a bunch of systems on Linux, Apache and other open source projects. One of the first things I worked on was a e-commerce site for the US Travel and Tourism Administration. Interestingly enough, it used to be that you could order publications and run queries on live data on the website, functionality which seems to no longer exist…. A footnote is that the LAP (no M, I think we used BerkeleyDB) machine we built ran for 2 years after we stopped maintaining it, only to be taken offline because someone cut the power.
Other projects involving Linux, Apache and mostly Perl included an inventory control system for the National Institutes of Health, the first intranet of the Department of Health and Human Services, data processing and aggregation for the Head Start program and a unified email system for the Department of Commerce (70,000 employees, 1500 agencies across 130 countries…). My partners in crime in all this were Peter Hartzler and Jim Devaty, we were all working for a small government contractor called Ellsworth Associates.
At the time (1995), when we built systems for various agencies, we would not tell clients that we were using open source, but rather used generic terms like Unix. Both open source and the internet were quite exotic and we had problems sourcing robust enough software, either in open source or running on Linux, and explaining how we were building systems. One classic example of this was our need for a more robust database. We started with Empress, then tried Solid. We eventually settled on Oracle, using Linux’s ability to run SCO binaries. When we called Oracle for a license, they asked how many end-users we had. Since the data was to be published on the internet, we told them ‘millions’. You can see where this went, particularly since the Oracle sales team had never heard of the internet….
I eventually left government contracting for the craziness of the dot com boom in Silicon Valley, but that was not to be the end of my exposure to open source in government. One of the first contracts my new company got in 2001 was to look at the use of open in a naval command, NAVOCEANO. Our findings were widely published and we concluded that open source had very high ROI in the kind of deployments this particular organization had. In the following years, with huge budget pressures on non-warfighting commands, this advice and strategy proved to be invaluable for NAVOCEANO and eventually resulted in Navy-wide strategies around open source.
But it wasn’t just the US government that was interested in open source. As part of a wider project in the UK, we spent time with Minister for E-government Transformation of the UK Cabinet Office. The UK government was embarking on a wholesale restructuring of government in an attempt to reduce redundant departments. Part of this was to be achieved by building extensive web-services framework, whose maintenance would be outsourced to an SI. The fear was that, essentially, the government would create an external department over which they would have little control or ownership. And the question was, could open source mitigate this potential problem? This question resulted in some interesting discussions to say the least…
But this was not the last time we would help government leadership use open source for policy aims. One project we worked on was developing an open source strategy for four cabinet-level US government agencies around a specific piece of software. The Department of Housing and Urban Development (HUD) had build a very sophisticated enterprise architecture and management software that had subsequently been deployed at Justice, Agriculture and Labor, and the GAO was considering as a standard application across all agencies. The problem was that HUD alone was responsible for it’s development and it could not solicit funds from other agencies to support the development efforts (congressional budgeting rules forbid inter-agency funds transfer with congressional approval). So the CIO at these agencies wanted to see if they could fund an open source project that they all could contribute to. In the end, going it alone proved to be too expensive and politically risky, but the problem was solved when one of the main developers setup his own OSS project.
Open source in government has evolved in fits and starts, but it’s potential benefits are enormous. In large parts of the world, from France to Brazil, it was seen as a way of distancing governments from the hegemony of American software companies, but in other place like the UK and US, it has been seen as a way of strengthening government control over it’s institutions.
In either case, it’s interesting that it has finally come full circle after 15 years. Back in 1995, not only did we build production systems using open source, but we also were active participants in the open source community. Looking back on it, I guess you could call us pioneering, even if we didn’t see it that way, it was just a better mouse trap. We don’t do much work with governments anymore, as it seems that open source has become more of a political that a strategy issue, but perhaps that will change in the coming years.
Chris Maresca is a business & technology strategist who, over the last 15 years, as advised over 30 global 500′s and over 60 startup companies. He can be found on the web at http://www.chrismaresca.com