Well, all good things take longer than we'd like or expect. However, some recent progress has helped resolve most of the remaining issues standing in the way of an initial ORCM release. I'm finishing up the equivalent of "mpirun" over the weekend, which was the last required tool.
Once this is complete, we'll need some time just beating on the system to expose the remaining bugs. Not everything will be working (e.g., still won't have a scheduler interface), but the mechanics of job start, execution, termination are all there. In addition, we have the automatic failure recovery of processes and nodes, no single point-of-failure anywhere in the system, and job startup time that is essentially independent of job size.
Still much to be done, but perhaps enough for a 0.1 release. I'll send around a draft release note that contains a list of current features, along with expected features for the next release.
Are there volunteers out there willing to give this a whirl?