My good friend David Eaves wrote an exciting post wherein he talked about his vision for StatsCan and how it should act like Google.
I think that StatsCan, on its own, could provide some compelling and valuable data for citizens of all stripes by implementing simple, cheap and scalable web services via conventional APIs. However I think StatsCan alone fails to capture the potential in technology.
Taking the Google analogy further and perhaps more literally - I think the Government of Canada needs a centralized managed repository for all its open initiatives. From APIs to software projects and packages to documentation.
Models abound for this kind of service - code.google.com being the most salient and directly applicable.
By providing robust (bilingual) documentation of the APIs proffered by Statistics Canada (and other ministries) Ottawa could spur innovation nationwide. Value-added services, mashups, statistical models and the like would all be available to anyone with the determination and the skills required to make use of the data provided.
Taking it a step further, the Government could provide an eco-system for these projects. A greenhouse for projects that make use of government data or contribute or expand government services.
The savings this could bring to the government’s TCO for software would be tremendous. The cost of staffing an open source community management bureau would be a fraction of the expense paid out to extortionist software vendors. Even if the government outsourced the maintenance of this project to a company, the cost savings would still be significant.
If the government developed robust and transparent standards for security, provided code and tools to develop applications - the issues of security and usability would be minimized.
By providing incentives for development of specific projects and fostering a community of recognition and respect the government can use its most valuable assets, trust, knowledge and influence to drive innovation.
The infrastructure required to build this is peanuts by governmental standards - a rack of servers, a high bandwidth internet connection and a horde of politically minded software nerds (some of whom speak and read French).
To start out, the working group should look like this:
6 developers with experience in open source. Two should know .NET, two should know Java EE and two should be web app experts - PHP, RoR, Python etc. They write foundational code, submit updates to the code bases and handle bug fix delegation etc.
1 project manager - this is our big-picture thinker, tie breaker and bean-counter. The Boss-Person.
4 community managers to wrangle the community projects and facilitate community leaders.
2 government liasons. One should be techie and the other political but both should be competent in the other discipline.
2 UX people, these people ensure applications comply with usability and accessibility guidelines and give guidance to the designers in the community.
3 Tech writers. These people document code and APIs, write tutorials and construct videos and other media.
So that’s 18 full-time, high-quality tech people - that’s probably about a million dollars per annum in costs all told, the value generated by the team would be tremendously greater than that.
This is an excellent idea. I do have to say that governmental (IT) projects are challenging for so many reasons:
- Teams do not have the required skills (it would be ideal to follow the structure of your project team, but my guess is that this is not gonna happen).
- Timeline is not realistic
- Risk Analysis is limited in depth: which is something that this article, IMO, suffers from. It is over-optimistic as it’s not listing the challenges/risks that this project might encounter.
For those interested, this is the whole series of governmental project management mistakes.
Left by PM Hut on February 9th, 2009