To close off this thread -- After various offline discussions we have adopted the new field Imports: (thanks, Martin Maechler, for the name) which should list namespaces that are required to load the package (via imports in the NAMESPACE file or use of :: or :::) but do not need to be attached as packages.
Depends: should be used for packages that need to be attached before the current package, and should be disjoint from Imports. On Thu, 12 Aug 2004, Deepayan Sarkar wrote: > On Thursday 12 August 2004 10:55, Prof Brian Ripley wrote: > > > I can't think of any other example of one package only needing the > > namespace of another (contributed) package. > > That's probably because few packages have namespaces yet. Another > example is nlme which imports lattice, but that's still within > recommended packages. In the long run, hopefully more packages will > start using namespaces when appropriate, so this might become more > relevant. > > The only issue I see here is a way to mention namespace requirements in > DESCRIPTION (which could have been done with Depends thus far because > it didn't actually do anything else, even though though the intent has > always been there). > > Imported packages are still dependencies in the sense that a package has > to be available for it's namespace to be usable, so leaving them out of > Depends: feels slightly weird, but I can live with it. > > Deepayan > > ______________________________________________ > [EMAIL PROTECTED] mailing list > https://stat.ethz.ch/mailman/listinfo/r-devel > > -- Brian D. Ripley, [EMAIL PROTECTED] Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UK Fax: +44 1865 272595 ______________________________________________ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
