The Open Source "Field of Dreams"

IMHO, expecting that an open source EMR is the solution is another example of the same flawed thinking that we can't seem to get beyond. It is still about starting with "what is the right technology?". The software is not where the big expenses come in, it is in the process of practice transformation where the expenses tend to rack-up. There are at least 6 fairly good open source EMR's already for the ambulatory EMR space, and then there is Vista from the VA. It simply costs too much to implement free software for most practices. And, no matter how much better the technology in the near future for these "free" software systems proves to be, they will still probably be too expensive because they have usually been designed by geeky types who are very dissimilar to what exists in most practices. They have little focus on designing the software with the proper focus of facilitating a non-disruptive implementation process. This is a far greater challenge than creating software that only does something!
In order for the software to be affordable, the core has to be fine-tuned to be able to be implemented as efficiently (start simple, then incremental evolutions) as is possible. So, the core of the "killer EMR" will not likely prove to be full OSS. However, the core should be completely accessible through an open SDK/API approach, so that it can be extended-interconnected with as few limitations as is possible and without introducing chaos and inefficiencies in the practice.
Implementing technology in a medical practice is very dissimilar to installing an Apache web server system. The latter does not require all the workflow and social re-engineering that is a far greater challenge.
I doubt it will be wise for SOAPware to turn our solutions into a completely open source project for these reasons as well as many others.