The Garden and the Horse Race

Strategic Open Source Implementation

The concept of this blog post come from some recent discussions I have had about open source communities but the concepts of the garden and horse race are from another who has the amazing skill to craft analogies that put my thinking in proper perspective. I thank this anonymous person for assisting in today’s blog.

When a new implementation of an already existing feature is submitted to an open source community, the community can take several actions; however, I want to focus on two such choices:

1. Compare the existing implementation to the new version and, hopefully, make a decision based on performance, quality of code, maintenance ability, etc and not based on who wrote the code. I call this the Horse Race
2. Allow the new implementation to co-exist with the previous implementation in a module with a common API set that allows the consumer of the open source technology to select the implementation they want. I call this the Garden.

In the horse race, implementations compete against each other within the development community and one is selected over the other based on technical merit or politics (something that should be avoided at all cost). This concept seems reasonable enough; however it limits the amount of contributors to a given project as only one solution will be selected and the other authors will have to move along.

In the garden, multiple implementations are made available within the project allowing various contributors to continue to work on and promote their variation. As long as the various offerings all work from a specific API then consumers are given more choice. Of course, the argument against this can be that having multiple groups working on the same problem within a project is a waste of resources.

So, why am I even writing on this – consider the case of an open source project with various corporate entities submitting implementations that support their ecosystem solutions. As a community manager, I want many ecosystem partners in my community to expand the opportunity for my open source solution to be adopted. A horse race approach limits the size of my ecosystem and turns away possible customers, as the losing implementation will not be included. In the garden approach, all ecosystem partners are welcome to the project and given space for their solution which allows them to reach out to more customers and expand the overall project’s customer base.

As I said in my last post, more and more corporations are taking part in open source and finding a way to ensure they can actively participate is critical for the entire open source movement and following the horse race model will jeopardize the current opportunity for open source to gain further momentum as the standard methodology for creating software.

Readers – what do you think? Am I missing something fundamental in this comparison? Do you have examples of communities following either process?

As a final note, I am attending OSCON next week in Portland, OR and have my handy Flip video camera with me. I will film several short interviews with various open source projects that I find interesting and post to my blog. If you have any project that you are looking for more details on, let me know and I can visit their both.

• Software