We want to share…
We write a lot of code in the New York State Senate Office of the CIO. It’s all released under dual BSD and GPLv3 open-source license, and we post most of it here on GitHub (and some on Drupal.org). We fervently hope that our peers in government– whether in other legislative bodies, at other levels of government (e.g.: city, county, federal), or other branches of government (e.g.: NY State Agencies in the Executive Branch)– will find and make use of this code.
For example, our NYSenate.gov website, into which 1000s of hours of developer work have been invested to customize the Drupal CMS platform for our legislative body, should ideally be able to meet the quite analogous website needs of the ~98 other State-level legislative bodies in the USA, not to mention the 1000s of City Councils in the USA. I’d like it to be easy for an IT Director for any legislative body in the US to quickly and easily get capabilities analogous to NYSenate.gov without needing to reinvent our wheel.
But It’s Hard…
However, most of the time for most of our applications to date, the promise of delivering more utility more efficiently by sharing the costs of developing and maintaining application code between peer public sector entities– better leveraging investments of precious tax dollars in public sector IT– is merely a good idea.
We try to go the extra mile to document our code well, posting it on GitHub, making our peers aware of it through speaking engagements, participation in unconferences, and use of the social Internets, and by liberating the data at least through our extensive support for open data standards and APIs. We field calls about approach and policy and precedent frequently from our peers. However, there is no question that it remains a heavy lift for a peer institution to actually take our application code and efficiently drop it into their own use case, because our code is necessarily customized to our specific nuanced needs in the Senate. Like most government entities these days, we’re understaffed relative to the workload we have. We don’t have the resources to easily contribute the custom Drupal modules we’ve written to Drupal.org and commit to becoming the maintainers of these modules. Nor do we have time to refactor our code into a more generic “Legislative CMS product” for our peers to more easily find and make use of.
Another example– this time one in which another government entity could help us here at the NY Senate: friends of mine at NASA recently built an online application (in their own free time) that leverages NASA’s publicly accessible LDAP server data (which is a subset of the internally accessible data from the same LDAP server), and then allow NASA employees to “claim” their profile on this public website, and add additional metadata like your Gravatar, your skills and interests, etc. Here in the NY Senate we’ve been researching enterprise social networking & collaboration applications that could help our internal staff to discover and leverage specific interests and skills of their colleagues. The application written for NASA, which is open-source and posted on GitHub, could help meet this need for us, but at the moment we’d have to rip out and replace code in order for us to be able to use it; it would take the NASA crew some significant additional work to refactor the code to turn it into more of a product in which we could merely edit a config file to deploy it for NY Senate…
In both our NY Senate example and the NASA example, if the refactoring work on code were done, and some level of ongoing maintenance and support of the code were available, more value would have been returned from tax dollars invested in our respective work, and we would also in turn be able to leverage off of the additional refinement of our code done by our peers in government to further benefit our institution. In other words, we could share the workload of software development and maintenance with our peers in government.
In the case of some business applications, vendors do offer productized SaaS solutions that meet the needs of a customer vertical like “legislative body.” But these solutions are rarely open-source, and are often either prohibitively expensive, or are not really built to suit the nuanced needs of a niche customer group, because such a niche may not be a large enough potential business market to warrant development of a highly nuanced product. Therefore, I think that the public sector needs to innovate in terms of self-support of technical collaboration.
I can envision a thematically-focused technically competent organization like the Sunlight Foundation doing the heavy lifting on a specific application that might be of specific interest to them, like our Open Legislation application. Others of our peers, like the NYS Department of Labor, have taken explicit steps to increase technical collaboration with other Labor organizations in the public sector by creating online hubs like LaborForge. Forge.mil sets a strong precedent at the Federal level, and I’ve heard rumors of a Forge.gov to be launched in the future.
Open standards help. Unconferences and virtual peer communities of practice help. Open APIs help. Open software licenses help. Code repositories help. Cloud computing infrastructure in which a machine image can be cloned with a few clicks helps. Developing applications within a service-oriented architecture (SOA) helps. The promise of online government app stores and feature servers help.
However, all of these approaches today are being implemented piecemeal if at all, and, when they are, often within a narrow niche where a motivated group of peers seeks help in solving a single problem they’re focused on at a single time.
I believe that citizens and governments alike would be well-served by an international non-profit entity (or perhaps a consortium of non-profit and for-profit entities committed to open-source?) charged with putting all these techniques together in a way that any public sector institution can easily contribute to and extract value from software development by their peers. Unlike some existing efforts, this organization would NOT be limited to moving the ball forward in a single thematic arena like “transparency” (e.g.: Sunlight Foundation) nor to a single geographic purview (e.g.: US Federal Government or New York State Government), nor be focused within a specific government service sector (e.g.: Labor, Motor Vehicles, Tax), nor that is focused on a particular technology stack (e.g.: Drupal for Government). Rather, such a non-profit technology organization or consortium would:
- Map the the virtual space of government IT applications (ranging from procurement and contract management, to payroll and other business applications, to citizen identity management, to CMS to CRM to more niche applications like news clippings services and legislative research tools);
- Find open-source licensed code that is being developed within the public sector to address these application needs and that could deliver value for a wide range of public sector entities;
- Do the hands-on work to generalize the code so that it is broadly useful and productized as much as practical;
- Publicize the availability of the new open-source products within the government (as customer) and public sector IT consulting services communities;
- Provide support and maintenance of the code for its public sector consumers and contributors;
- Maintain roadmaps for, and convene multi-institution developer communities around, specific major areas of opportunity for ongoing collaborative software innovation.
- Play a role in the planning and build-out of government clouds, app stores, and feature servers.
Such an entity, once fully realized and effective over a period of years, I believe could yield hundreds of millions of dollars of tax dollar savings worldwide on public sector software development and licensing annually. It could also significantly expand the size of the public sector market addressable by small and medium-sized IT consulting firms by moving dollars spent from software licensing to customization and support, and by allowing small engineering teams from separate companies and institutions to more easily collaborate to achieve large scale engineering capacity when required.
In other words, such an effort might well reduce government IT costs, increase the capacity for government entities to collaborate with one another through technology, increase competition in the private sector for government IT contracts, and increase the likelihood that smaller businesses can provide services higher in the food chain of government IT contracts.
Organizations that I hope will look at taking on part or all of this opportunity include Code for America, the Open Planning Project, the World Bank, ExpertLabs, and perhaps the Ford Foundation’s Effective Government Program.