eaves.ca

Last month I published the following article in the Municipal Information Systems Association's journal Municipal Interface. The article was behind a firewall so now that the month has gone by I'm throwing it up here. Basically, it makes the case for why, if government's applied open source licenses to the software they developed (or paid to develop), they could save 100's of millions, or more likely billions of dollars, a year. Got a couple of emails from municipal IT professionals from across the country

MuniForge: Creating Municipalities that Work like the Web

Introduction

This past May the City of Vancouver passed what is now referred to as "Open 3".This motion states that the City will use open standards for managing its information, treat open source and proprietary software equally during the procurement cycle, and apply open source licenses to software the city creates.

While a great deal of media attention has focused on the citizen engagement potential of open data, but the implications of the second half of the motion – that relating to open source software – has gone relatively unnoticed. This is all the more surprising since last year the Mayor of Toronto’s also promised his city would apply an open source license to software it creates. This means that two of Canada’s largest municipalities are set to apply open source licenses to software they create in house. Consequently, the source code and the software itself will be available for free under a license that permits users to use, change, improve and redistribute it in modified or unmodified forms.

If capitalized upon these announcements could herald a revolution in how cities currently procure and develop software. Rather than having thousands of small municipalities collectively spending billions of dollars to each recreate the own wheel the open sourcing of municipal software could weave together Canada’s municipal IT departments into one giant network in which expertise and specialized talents drive up quality and security to the benefit of all while simultaneously collapsing the costs of development and support. Most interestingly, while this shift will benefit larger cities, its benefit and impact could be most dramatic and positive among the country’s smaller cities (those with populations under 200K). What is needed to make it happen is a central platform where the sourcecode and documentation for software that cities wish to share can be uploaded and collaborated on. In short, Canada needs a Sourceforce, or better, a GitHub for municipal software.

The cost

For the last two hundred years one feature have dominated the landscape for the majority if municipalities in Canada: isolation. In a country as vast and sparsely populated as ours villages, towns, and cities have often found themselves alone. For citizens the railway, the telegraph, then the highway and telecommunications system eroded that isolation, but if we look at the operations of cities this isolation remains a dominant feature. Most Canadian municipalities are highly effective, but ultimately self contained islands. Municipal IT departments are no different. One municipality rarely talks to that of another, particularly if they are not neighbours.

The result of this process is that in many cities across Canada IT solutions are frequently developed in one of two manners.

The first is the procurement model. Thankfully, when the product is off the shelf, or easily customized, deployment can occur quickly, this however, is rarely the case. More often, larger software and expensive consulting firms are needed to deploy such solutions frequently leaving them beyond the means of many smaller cities. Moreover, from an economic development perspective the dollars spent on these deployments often flow out of the community to companies and consultants based elsewhere. On the flip side, local, smaller firms, if they exist at all, tend to be untested and frequently lack the expertise and competition necessary to provide a reliable and affordable product. Finally, regardless of the firms’ size, most solutions are proprietary and so lock a city into the solution in perpetuity. This not only holds the city hostage to the supplier, it eliminates future competition and worse, should the provider go out of business, it saddles the city with an unsupported system which will be painful and expensive to upgrade out of.

The second option is to develop in-house. For smaller cities with limited IT departments this option can be challenging, but is often still cheaper than hiring an external vendor. Here the challenge is that any solution is limited by the skills and talents of the City’s IT staff. A small city, with even a gifted IT staff of 2-5 people will be challenged to effectively build and roll out all the IT infrastructure city staff and citizens need. Moreover, keeping pace with security concerns, new technologies and new services poses additional challenges.

In both cases the IT services a city can develop and support for staff and citizens is be limited by either the skills and capacity of its team or the size of its procurement budget. In short, the collective purchasing power, development capacity and technical expertise of Canada’s municipal IT departments is lost because we remain isolated from one another. With each city IT department acting like an island this creates enormous constraints and waste. Software is frequently recreated hundreds of times over as each small city creates its own service or purchases its own license.

The opportunity

It need not be this way. Rather than a patchwork of isolated islands, Canada’s municipal IT departments could be a vast interconnected network.

If even two small communities in Canada applied an open source license to a software they were producing, allowed anyone to download it and documented it well the cost savings would be significant. Rather than having two entities create what is functionally the same piece of software, the cost would be shared. Once available, other cities could download and write patches that would allow this software to integrate with their own hardware/software infrastructure. These patches would also be open source making it easier for still more cities to use the software. The more cities participate in identifying bugs, supplying patches and writing documentation, the lower the costs to everyone becomes. This is how Linus Torvalds started a community whose operating system – Linux - would become world class. It is the same process by which Apache came to dominate webservers and it is the same approach used by Mozilla to create Firefox, a web browser whose market share now rivals that of Internet Explorer. The opportunity to save municipalities millions, if not billions in software licensing and/or development costs every year is real and tangible.

What would such a network look like and how hard would it be to create? I suspect that two pieces would need to be in place to begin growing a nascent network.

First, and foremost, there need to be a handful of small projects. Often the most successful source projects are those that start collaboratively. This way the processes and culture are, from the get go, geared towards collaboration and sharing.  This is also why smaller cities are the perfect place to start for collaborating on open source projects. The world’s large cities are happy to explore new models, but they are too rich, too big and too invested in their current systems to drive change. The big cities can afford Accenture. Small cities are not only more nimble, they have the most to gain. By working together and using open source they can provide a level of service comparable to that of the big cities, at a fraction of the cost. An even simpler first step would be to ensure that when contractors sign on to create new software for a city, they agree that the final product will be available under and open source license.

Second, MISA, or another body, should create a Sourceforge clone for hosting open sourced municipal software projects. Sourceforge is an American based open source software development web site which provides services that help people build cool and share software with coders around the world. It presently hosts more than 230,000 software projects has over 2 million registered users. Soureforge operates as a sort of market place for software initiatives, a place where one can locate software one is interested in and then both download it and/or become part of a community to improve it.

A Soureforge clone – say Muniforge – would be a repository for software that municipalities across the country could download and use for free. It would also be the platform upon which collaboration around developing, patching and documenting would take place. Muniforge could also offer tips, tools and learning materials for those new to the open source space on how to effectively lead, participate and work within an open source community. This said, if MISA wanted to keep costs even lower, it wouldn't even need to create a sourecforge clone, it could simply use the actual sourceforge website and lobby the company to create a new "municipal" category.

And herein lies the second great opportunity of such a platform. It can completely restructure the government software business in Canada. At the moment Canadian municipalities must choose between competing proprietary systems that lock them into to a specific vendor. Worst still, they must pay for both the software development and ongoing support. A Muniforge would allow for a new type of vendor modeled after Redhat – the company that offers support to users that adopt its version of the free, open source Linux operating system. Suddenly while vendors can’t sell software found on Muniforge, they could offer support for it. Cities would not have the benefit of outsourcing support, without having to pay for the development of a custom, proprietary software system. Moreover, if they are not happy with their support they can always bring it in house, or even ask a competing company to provide support. Since the software is open source nothing prevents several companies from supporting the same piece of software – enhancing service, increasing competition and driving down prices.

There is another, final, global benefit to this approach to software development. Over time, a Muniforge could begin to host all of the software necessary to run a modern day municipality. This has dramatic implications for cities in the developing world. Today, thanks to rapid urbanization, many towns and villages in Asian and Africa will be tomorrow’s cities and megacities. With only a fraction of the resources these cities will need to be able to offer the services that are today common place in Canada. With Muniforge they could potentially download all the infrastructure they need for free – enabling precious resources to go towards other critical pieces of infrastructure such as sewers and drinking water. Moreover, a Muniforge would encourage small local IT support organizations to develop in those cities providing jobs fostering IT innovation where it is needed most.  Better still, over time, patches and solutions would flow the other way, as more and more cities help improve the code base of projects found on Muniforge.

Conclusion

The internet has demonstrated that new, low cost models of software development exist. Open source software development has shown how loosely connected networks of coders/users from across a country, or even around the world can create world class software that rivals and even outperforms software created by the largest proprietary developers. This is the logic of the web – participation, better development and low-cost development.

The question cities across Canada need to ask themselves is: do we want to remain isolated islands, or do we want to work like the web, working collaboratively to offer better services, more quickly and at a lower cost. If even only some cities choose the later answer an infrastructure to enable collaboration can be put in place at virtually no cost, while the potential benefits and the opportunity to restructure the government software industry would be significant. Island or network – which do we want to be?

Related posts

• Why StatCan is (or could be) like Google (18)
• Wedding Open Source to Government Service Delivery (13)
• PublicVoice Interview on “Open,” Government and Citizen Engagement (3)