On Wed, Feb 23, 2011 at 10:11 AM, Les Mikesell <lesmikesell at gmail.com> wrote: > > Even if no one can come up with a way to extrapolate the dependencies > from the RHEL binary, the task seems like something that would scale > with more build machines and more people to stage the trial-and-error > builds. History seems irrelevant to many; if that is the case with the reader, hit DELETE now. Soon to be 50 years ago, Control Data's SCOPE (simultaneous control of program execution) had 'lgo' (load-and-go) which obviously had to resolve dependencies at run-time, and it produced a load map. It would seem this approach would be useful, but I dunno. Perhaps it is a tool that Jeff or someone would like to investigate further, perhaps not. IMHO, there's nothing unfathomable about a package reaching out to include other packages. IFF this turns out to be a useful approach, I would like to thank Dr. James Browne at utexas.edu for his mentoring of our group, particularly for his guidance in scheduling and proving program correctness.