Mediation

[Robert Schuster]

Hi fellow GNU Classpath developers,

for some time now I am participating in this project fixing bugs and adding functionality mainly to the java.beans package. Despite my good knowledge of the Java language, participation in development communities and especially the GNU community, is virgin soil to me. As a result I sensed a steep learning curve when I started helping Classpath. In the last weeks I found myself asking a lot of questions on topics which I think are common knowledge for a fair amount of you.

The problematic cases range from specific tool usage over project plans to general policies. I know there is a lot of tool documentation on the net and a hacking guide for Classpath which is enough for the fundamental stuff like CVS usage or coding guidelines but what I think is still left untouched are questions like:

• What is the outcome of discussions?
• Whom can I ask directly for specific questions?
• What is the general direction of the project? (or: What is considered old stuff which should be avoided in favour to newer decisions?)

Ideas on making this situation better with the intent to make the project participation more enjoyful circled in my head and found their way on a sheet of paper. It was clear to me that the realization of this would need a dedicated effort which cannot be burdened onto someone's shoulders without intrinsic motivation.

After all I am a computer science student who got recently interested in software engineering and was seeking a topic for his semester thesis. I approached the SE group at my department in order to do an academic work around my initial ideas and got positive reponse.

Now I`d like to ask you if you welcome my effort to enhance our project and use the experiences gained from that for academic work.

The following paragraphs describe the planned enhancements (Criticism and comments are welcome):

My basic assumption is that the development process of an unstructured Free software project should be enhanced by non-invasive methods: Any means that make the developer's participation work less comfortable should be avoided. Academic projects with similar goals as mine have largely relied on producing tools which did not get adopted. In contrast to that my approach is based around a role that I call 'mediator'.

In short the mediator's goal in a F/OSS project is to take care that no idea is lost. To be more specific these are some of the things the mediator should do:

• Collect information about project member's interests (e.g. package responsibility).
• Remind of certain events: release, urgent documentation updates, long-term goals.
• Keep an eye on the project documentation and guides.
• Be an active guide for newcomers.
• Dig up or re-introduce ideas which otherwise would get lost in mailing-list conversations.
• Write down the results of decisions and ToDo items.

It should be noted that I consider that some of these tasks require a lot of sensitiveness. A bad formulated ToDo list entry or project decision can lead to unfriendly and heated debate. Furthermore the mediator does not have any higher privileges: Changes to every recorded statement can be made by each project member and the mediator does not make decision, but rather collects them.

The usual work of the mediator will consist of an in-depth study of the mailing-list (archive) but also other communication channels like Classpath's blog area and IRC. Apart from that he stays in touch with the other members and updates the respective documents. The initial effort will be on collecting the existing and upcoming data and finding a suitable way to organize it.

The duration of my thesis is limited to 3 months. At the end of this time we can look at the results of my work and poll whether to continue it or not.

I hope this introduction gave you enough information to get a picture of what I want to do. As stated above criticism and comments are welcome.

Privacy: I respect everyone's privacy but its likely that I will take quotes for my thesis from the mailing-lists (which is already publicly archived) and perhaps from IRC conversations. In the latter case I will address the involved persons and anonymize their statements if they want me to do that.

cu
Robert