"fred bertsch"
Andy has been following the large discussion of PQS, so I will not enumerate the more minor criticisms included there. Some of the major criticisms that influenced my difficult decision include the following:
* Many of the reviews liked the documentation, but improvements there would help out for a large number of reviewers.
<snip>
* Since the library includes a large number of compile-time type manipulations, it would be great if the library included standard metafunctions to query these types. In addition, following other template metaprogramming conventions established within Boost such as documenting the concepts used would be a great help.
<soapbox> I'm sure Fred didn't mean to leave this impression, but I just want to clarify something here that could be misinterpreted. Documenting concepts is not really "a tempate metaprogramming convention;" it's a convention that was started by Stepanov et. al. when the STL was invented, and is appropriate for all generic libraries. Aleksey used it to document the MPL because it was appropriate there, too. That means almost any library with templates in its public interface should have concept documentation. </soapbox> -- Dave Abrahams Boost Consulting www.boost-consulting.com