Today, I drove all the way up to Sebastopol, CA to meet with Tim O’Reilly, one of our early investors. Tim had supported our initial work with Open Source, and I wanted to give him an update on the early results of our move back to an open source model. The short story goes like that: we generated more qualified leads in the past six weeks that we did in the last six years, we tripled the number of software vendors building their products on top of ours, and we quadrupled the number of system integrators using our product for engagements with their customers. Half of the committers to the code of our process server are open source contributors who are not on our payroll, and this ratio should increase significantly following yesterday’s donation to Apache, while extending the user base an order of magnitude faster than we could on our own.

Overall, Tim liked what he heard, but what got him excited was the idea that we could use our software to implement a tansition from software to infoware, as described in this article. Tim’s idea is to encourage users of our process design tool to share their process models through a public repository, much like people contribute articles to Wikipedia today. With the proper infrastructure in place, a process analyst building a custom order entry process for SAP could immediately get access to the list of processes that have been built using the same SAP BAPI that lets you create the header for a purchase order for example. You might not find the exact process that you were trying to build, but you would benefit from the collective experience of your peers, and you would start from were they stopped, as opposed to starting from scratch. While this approach raises some objections for processes that bring advanced competitive differentiation, as illustrated in the discussions that followed the publishing of our article introducing the concept for On Demand Development, it appears to me that a large community of process analysts would not mind sharing part of their work if that allowed them to benefit from the work of others in return. In fact, on my way back from Sebastopol, I met with a company that builds advanced factory floor automation software for the manufacturing industry, and they offered to contribute some of their processes if we were to build the proposed repository.

In order to better understand how to encourage the creation of volunteer labor, Tim suggested that we read the excellent article from Dan Bricklin titled The Cornucopia of the Commons. We will diligently follow his advice, gather your feedback, and give serious thoughts to his idea of setting up a public repository for business processes. The idea has been suggested to me very many times in the past, and I believe that the community of BPM practitioners might be ready for it today, through the convergence of Open Source, industry standards, and mature-enough products.

Thank you Tim!