François Elie, you are one of the pioneers and champions of Open Source application development in the public sector. What are the main advantages you see in this?
First and foremost, although it sounds rather obvious, Open Source has liberated IT... just as, once upon a time, mathematics was liberated. But among the commonly accepted technical and economic arguments, I also think two other points are especially important:
- In the first place, keeping a clear separation between solutions and services helps to prevent ‘vendor-lock-in’; it is vital to be able to put service offerings into competition with one another without affecting the technological choices you make
- Secondly, it is a formidable tool to drive the change.
For our part in the public sector, adopting Open Source was an obvious choice; although we quickly realized that there would never be any business applications available in Open Source unless those who were going to use them actually developed them themselves. That is why we founded ADULLACT in 2002, to develop the business software that local authorities needed. Today ADULLACT’s members include some 130 major local authorities. We were the first to implement a software forge dedicated to the public sector and to the cooperative development of these e-government applications by professional communities. Since June 2008, we have been hosting the applications developed by the French government, and we have persuaded the European Community to create a public service forge2 (http://forge.osor.eu/), which was unveiled in November 2008.
What is your vision of software forges?
When it comes to proprietary software, developers are working in ‘bunker’ mode; it’s the business of ‘secrecy’. In the Open Source world, software forges allow for much wider collaboration between developers working together on the same code (using mediation mechanisms). The code is constantly being re-worked; every now and again it is ‘frozen’ as a new version is released, and this represents the state-of-the-art for a given moment in time. Work then continues to develop the code still further.
Today’s forges address the needs of three types of community:
- The ‘user’ forge is the historic kind of forge addressing the community of users who go there to find the open software they need. The economic model is based around publicity, and possibly also a small amount of traceability
- ‘Industrial’ forges – typically the OW2 forge – used by the community of middleware developers. The economic model behind this type of forge is shared research and development. Such forges provide developers with a software production environment known as CDS (Concurrent Design Services), comprising messaging servers, forums and packaging versions, etc. which are the heart of the forge. They involve very heavy security to guarantee access, especially the all-important write-enabled access, to code, as well as to ensure that the forge itself is durable and robust. In this context, the workflow of the enterprise can be modified virtually in real time, with tools such as utilities, and integrity controls which will transfer directly from the forge to the enterprise’s production chain, and which will link the software production itself with its live implementation.
- The ‘customer’ forge has recently emerged, with an economic model based on economies of scale achieved through sharing the demand for business applications. It represents a genuine marketplace.
Over the past few years, we have seen these forges moving away from each other, which I think is very dangerous. There is huge interest if they were working together.
What dangers do you see ahead?
If we look at how the economics of Open Source have evolved, it is noticeable that it has been adopted step-by-step, and linked to the history of communities. At the start, it was fostered by developer communities at grassroots level; networks, for example. An initial ideological impediment against open software failed in the year 2000 thanks to Eric Steven Raymond3 and these low-lying areas of software have now achieved extremely widespread adoption. Next we saw acceleration thanks to the savings enabled by Open Source. IT industry players shared their middleware developments so they could concentrate their added value and their differentiation on end-user applications. These middleware applications are now in the process of being very widely adopted. So now it is the turn of business applications, with their communities of customers. Their broad acceptance has yet to come.
Nevertheless, you could imagine other scenarios instead of the gradual adoption of Open Source. What we have, in effect, is an Open Source software ecosystem which could run up against the problem of keeping control over the roadmap. Let’s not forget that last January, MySQL was bought by SUN for a billion dollars. I foresee two kinds of threats:
- The first danger is that some kind of ‘power’ backed by a huge amount of capital could get its hands on most of the Open Source communities in order to gain control of their roadmaps. Microsoft, for example, is going to be giving away a development environment in order to encourage children to develop collaboratively in schools. So it is conceivable that the power behind the roadmap for different communities will change hands. If we arrive at a monopoly on the communities, who will be ‘doing’ Open Source tomorrow? That’s the real question. We need to remember that revolutions are sometimes hijacked, and not always by allies of that revolution.
- The second danger is that if we outsource everything, and end up with a massive, totally opaque kind of ‘cloud computing’, the question: “Do I have the right to modify my software?” no longer has any meaning, because actually someone is creating software on my behalf. For me, the notion of ‘cloud computing’ which is being rammed down our throats is, in reality, a huge trap which could even go so far as doing away with the idea that we can control our own working tools.
So, do we need to come up with a new model?
What, indeed, is the best solution? One answer could be that we need to speed up the process of cohesion between the three kinds of forges and their communities. To ensure that Open Source will last, and is adopted on a massive scale, three conditions must to be met:
- 1st point: communities need to be paid for what they do. All work deserves to be rewarded! We therefore need to set up economic models that enable us to pay contributors.
- 2nd point: the customer ‘wish-list’ must change. If we want to develop business applications that can be shared with others, it means we must do up-front work to ensure that applications are sufficiently generic. If we don’t do this, we are inevitably creating a specific solution in Open Source, and we have gained nothing in the process.
- 3rd point: IT industry players must reduce the granularity of their components. The smaller these components are, the more robust they will be and the better they will communicate between each other.
In this new model for the marketplace, any services activity will be able to supply code. And then the real interest will be in ensuring that Open Source is sustainable: we must not allow it to become a throw-away commodity! We need to produce sustainable and robust objects to a high standard of quality.
1 ADULLACT: The French association of Open Source software developers and users in local and central government
2 An example of the collaboration between ADULLACT and the European Community in the framework of OSOR: the English versions of openCimetiere and openCourrier were offered on the European forge >>>
3 Eric Steven Raymond, who is credited with popularizing the term ‘Open Source’, was committed to promoting the quality of Open Source code.
ADULLACT has set itself the objectives of supporting and coordinating the actions of local authorities, public sector bodies, and hospitals with the aim of promoting, developing, sharing and maintaining a common software legacy of open applications that are useful when it comes to delivering public services.
Find out more >>