Well said. I think the EEG work needs to address some of these items, and we are going to be looking into what can be done to not only help developers understand what needs to be done (and how they need to think), but also to help them understand how to build, test, and deploy.

Eric

I started my OSGi life back in 2001 working on embedded applications for the automotive industry, but these days I work on OSGi server applications, so I’ve seen both sides.

I see the big selling point for OSGi as the ability to build highly modular applications, which is something desirable regardless of the target platform.

The biggest thing holding OSGi back now is developers that understand how to design bundles for OSGi and implement them appropriately.

Equinox’s solid DS implementation, excellent DS tooling from Eclipse, and the inclusion of DS in the Eclipse SDK should help matters since now people can use OSGi services everywhere, so long as they understand how.

Training developers to adjust the way they think is going to be the hardest challenge of all.

It’s just an interesting story to see how what people initially envisioned OSGi doing isn’t really what it’s doing now.

However, I think the value of OSGi comes largely due to the frugality that is required in these embedded markets. It will be interesting to see if we can maintain that frugality in the future.

Peter Kriens