Matthias Schabel wrote:

In general, it's easiest to avoid explicitly creating unit systems by using the ::unit_type approach. For most cases, the unit system ought to be considered an implementation detail since the library supports heterogeneous units transparently. That is, for any given unit, the appropriate system is just the set of units needed to fully describe the unit itself...

Thank you Matthias for the code example. It was exactly what I needed to get back into implementation flow. Yet I still feel the urge to create an md_units_system. For instance, once I create a system including base_units picoseconds, daltons, and nanometers, I can automatically create an energy unit in that system which magically has units of kilojoules per mole, without having to remember how mass, time, length, and energy are related. I am now pondering the creation of a clever set of base units such that Avogadro's constant appears in only one conversion factor to SI units. Either mass or amount would need to be a derived unit. But perhaps this topic is drifting outside the scope of the boost users list... Steven Watanabe wrote:

This was definitely a bug in the library. Fixed in the trunk.

Thank you Steven Watanabe so much for your rapid analysis and fix of this bug. I have updated to the latest trunk and can verify that the compile errors no longer occur. I now have all I need to proceed to more sophisticated use of boost.units in molecular computation. Thanks again for the thoughtful and diligent help Matthias and Steven. --Chris Bruns