Open source community - friend or foe?
I spent Wednesday at a small event called Open source community - friend or foe?, organised by OSS Watch to discuss strategies for developing communities around development projects.
To quote from their website, OSS Watch:
OSS Watch promotes awareness and understanding of the legal, social, technical and economic issues that arise when educational institutions engage with free and open source software. It does this by providing unbiased advice and guidance to UK higher and further education.
OSSWatch are now ‘drinking their own koolaid’ (so much nicer than ‘eating their own dog-food’) - they’ve created an OSS development project (Simal) and begun to assemble a development community around it.
A good portion of the day was spent outlining the tools necessary for maintaining a community-based development project. We heard that ‘every tool should do three things’, and that, at a minimum, the core set of tools you really need are a website, issue-tracking, version-control and a discussion list.
OSSWatch were careful to avoid recommending particular tools for development. But which tools are they using to help them manage their own development project? Turns out, they’re using Google Code! This is an important thing to know about OSSWatch - despite the plethora of readily available open-source tools available for managing development communities/projects, they have picked the one they think is best, which happens to be decidedly closed-source. There are two important messages here I think:
- OSSWatch are not open-source bigots - they offer a balanced and considered view
- You should use the best available tools for managing your community-based development project
OSSWatch are actively engaged with helping projects in our community develop their communities. Some such projects were represented at the event on Wednesday and their feedback was very positive.
I learned plenty from this session and look forward to working more closely with OSSWatch. I took away a few specific ideas which is always a good result.
Anyone in the UK HE and FE communities considering embarking on a open-source software development project should consider developing a community as a priority, and if they need help and advice about this then OSS Watch should be their first port of call.
However, the big question for me is:
While the community approach to development is (sometimes) demonstrably successful (Apache, Eclipse etc.), can it work if it’s scaled down to a typical 18month JISC-funded development project?
OSSWatch tell us that they are currently wrestling with this very issue. Watch this space…
(After the event we raised a glass to Randy Metcalfe, who is returning to Canada after years in Britain - bon voyage Randy!)
Technorati Tags: community, GoogleCode, OSS, OSSWatch, software development
June 21st, 2007 at 2:36 pm
My thanks to Paul for attending our workshop and more importantly, many thanks for Paul’s excellent input on the day.
I would first like to reiterate that projects struggling with any aspect of setting up an open source project should contact OSS Watch with your questions (contact info on the about us page of http://www.oss-watch.ac.uk). We are, in most cases, happy to come and visit your project (at our cost) to discuss your specific concerns.
With respect to OSS Watch choosing the “best solution” when we chose Google Code for Simal, it is worth pointing out that we have not yet decided that Google Code is the best solution. OSS Watch is a learning team, we already have experience of a number of other hosting environments, however, we do not yet have experience of Google Code. Hosting Simal there is an experiment that will allow us to evaluate its functionality. it is not an endorsement of the facility (although we are happy so far).
Another interesting aspect of Google Code which is, as you say, closed source, is that it is a service that Google have built from a considerable number of open source projects. Google are very active within the open source communities that their services depend on. They are a good example of a sensible partnership between closed source, commercial systems and open source projects.
Finally, to your question, can community be applied to a short project such as the JISC projects?
You are right, this is something that OSS Watch is considering. One thing that I personally believe that yes, it can scale. All projects start small and some build from there (many do not). I see no reason why JISC projects should be any different.
However, there is another factor involved here. Cash, or more accurately the source of that cash.
When a project is started up by a commercial organisation they have some long term plan for the work. One would assume there is a need for it within the company and that they will continue to support it for this reason.
For JISC projects this is somewhat different. These projects start up because a potential user base has been identified and the JISC provide money to try and satisfy that users needs. When this money runs out the project team move on to other things. There is often no incentive for the project team to stick around and therefore no incentive to support a fledgling community.
This is the problem that I find hard to address. At this time I believe that the solution is in creating some kind of incubator in which we can take promising looking projects and continue to build the community around that work.
The good news is that I’ve spoken to a number of people in the JISC about this. There seems to be a considerable desire to address this and OSS Watch are actively attempting to formulate recommendations. We look forward to working with Paul (and his UKOLN colleagues), the JISC and existing JISC projects in trying to address this problem.