I have now released version 0.3 of the Probability library, which is available in the Boost Vault. As usual the full documentation is available at http://biology.nmsu.edu/software/probability/ Although this version expands the interface by adding a few mixed-argument arithmetic operators, the main focus is the addition of a generic algorithm to identify the most likely paths through a graph. This algorithm finds use in a variety of scientific and statistical models. More importantly, though, it illustrates how well-suited this library is to the development of generic algorithms that ensure type safety and correctness even though template arguments may be as different as probabilities and likelihoods, defined with different value types, and even defined on different domains (e.g., linear vs. log). The library makes such algorithms very straightforward to develop; using native types to represent these quantities within a generic algorithm would introduce artificial constraints, convoluted syntax, maintenance difficulties, or all three. Please review the current version of the Probability library. I think the motivation is much stronger now that it is demonstrated to handle generic algorithms well, even when a diversity of types are mixed together. I would greatly appreciate feedback on how to further improve the library. I would also appreciate feedback from people who can (or cannot) compile the library on other systems. Thank you very much. Cheers, Brook