In this article, RedHat's Michael Tiemann says:
"One of the mistakes we made when we launched this Enterprise Linux product was we focused so exclusively on this enterprise market that we left this (early-adopter customer) square uncovered," Tiemann said. "It insulted some of our best supporters. But worse, we lost our opportunity to do customer-driven innovation."
I was surprised when RH made the announcements around the Fedora/RHEL split. I thought it was long term suicide for the benefit of short term profit. Let me explain. OSS businesses thrive on the distribution reach OSI licenses afford them. In other words, JBoss incurs a 20c cost of sales for every dollar in maintenance, where proprietary vendors incur a $2.5 COS for every dollar. This, by the way is why the model works, we can do R&D without the licenses.
Limiting access to the "real distribution" seemed non-sensical. Let people use your stuff and then turn them on. The way I rationalized the decision by RH was based on branding. That essentially RH had enough of a brand and market presence, 85% in the US. I think they thought they could get away with it. I still saw it as the best way to alienate your user base as Michael now acknowledges.
On the positive side, since RH went to a proprietary distribution model (you really have to buy their distro to use it) they made tons of cash and became profitable (GAAP) which is an impressive feat given that 5 years ago they were still in the packaging/barnes and noble distribution model.
Another point Tillman makes is around innovation. We couldn't agree more. Keeping your projects viable and alive is something we focus on at JBoss. All the projects under "professional open source" management that are part of JEMS, be it Tomcat, JBoss AS or Hibernate, are the REAL products, no split, I think this is a lot cleaner. We want the open project to yield enterprise open source quality though our "Professional Open Source" methodology, it is that simple. We sponsor the actual projects as our main expense in R&D.
We selected our VC about a year ago, and one big component was "did they get it". Some would walk in and start talking about "taking the product proprietary" or offering "enterprise versions". We would walk away. I am glad this article by Tiemann validates our choices,