I just stumbled upon a good blog post by Ed Burnette.
Without entering into the details (although they are not uninteresting), the issue described comes down to Sun voting no on a Java Community Process specification because, it seems, the techical solution (here OSGi) adopted was invented outside Sun. This is known as the "not invented here syndrome" ("NIH syndrome" in short).
The NIH syndrome is unfortunately very frequent in the software industry. With Orbeon Forms however, we strive to leverage existing software and specifications as much as we can. Most notably, our forms solution is based on XForms; we are eager to see XProc, the standard pipeline language we are developing at W3C, reach the level of recommendation; we use XML technologies in several places; and we leverage an impressive array of open source Java software, in particular Saxon.
It can actually be fun to implement standards. What's even better (outside some very politicized areas such as HTML) is working with other people to define what the standard is, and influencing the standard based on your experience in a particular domain (in our case, form-based applications). Not only do you leverage other people's brain power to provide new ideas, but you also get to test your own ideas against others. This can be quite enriching.
If you catch us having an NIH crisis, please let us know.