Broader AP set ?

[eddelbuettel]

When accepting it on CRAN (with some hesitation), Uwe wrote back

4.     AP <- available.packages(contrib.url(options("repos")$repos[["CRAN"]]), 
                                filters=list())

may be worth to be extedned: Why not include BioC? Dependency structures are somtimes 
in the way that a CRAN package depends on a BioC one that again depends on another 
CRAN package ... 

We definitely want BioC for resolving dependencies. Personally I do not care about the
reverse package recursion up BioC -- but others do.