Several packages aren't really operational without the suggested packages, resulting in a lot of manual dependency resolution. The rule seems to be "Don't call my really useful functions without manually resolving all my possible dependencies." Can we move all "Suggests" into "Depends" and deprecate the Suggests metadata? -- output of sessionInfo(): . -- Sent via the guest posting facility at bioconductor.org.

Hi, On Tue, Dec 11, 2012 at 3:12 PM, Andrew Pennebaker [guest] <guest at="" bioconductor.org=""> wrote: > > Several packages aren't really operational without the suggested packages, resulting in a lot of manual dependency resolution. The rule seems to be "Don't call my really useful functions without manually resolving all my possible dependencies." Can we move all "Suggests" into "Depends" and deprecate the Suggests metadata? I don't think I've gotten the same impression, but perhaps we are using different subsets of packages. I guess it would be helpful if you could identify which packages, in particular, you are referring to if you're sounding a call to action, since this clearly isn't a universal something. Cheers, -steve -- Steve Lianoglou Graduate Student: Computational Systems Biology | Memorial Sloan-Kettering Cancer Center | Weill Medical College of Cornell University Contact Info: http://cbio.mskcc.org/~lianos/contact

Hi Andrew, On 12/11/2012 3:12 PM, Andrew Pennebaker [guest] wrote: > Several packages aren't really operational without the suggested packages, resulting in a lot of manual dependency resolution. The rule seems to be "Don't call my really useful functions without manually resolving all my possible dependencies." Can we move all "Suggests" into "Depends" and deprecate the Suggests metadata? This isn't a question for BioC. Instead it is a question for R-core, who are responsible for such things as deciding if the current status of package dependency hierarchy is useful or not. I should note that the general goal for BioC-core has been to keep both the number of downloaded packages required for a given package to a minimum (so we have advocated putting things into suggests), as well as keeping the search path as short as possible. So I guess most of us would disagree with your suggestion. Anyway, there is never any manual dependency resolution required, as this is in your hands. Note that library(BiocInstaller) ?biocLite will bring up a help page, which gives the argument list as biocLite (pkgs=c("Biobase","IRanges","AnnotationDbi"), suppressUpdates=FALSE, suppressAutoUpdate=FALSE, siteRepos=character(), ask=TRUE, ...) Note the final argument is the ellipsis (...). and there is this statement under the 'Arguments' section. ...: Additional arguments. Which is code for 'arguments that can be passed down to lower level functions', one of which is install.packages(), as noted in the 'See Also' section. So if we look at ?install.packages there is dependencies: logical indicating to also install uninstalled packages which these packages depend on/link to/import/suggest (and so on recursively). Not used if repos = NULL. Can also be a character vector, a subset of c("Depends", "Imports", "LinkingTo", "Suggests", "Enhances"). So all you need to do is biocLite("whateverpackageitisyouwant", dependencies = c("Depends","Imports","Suggests")) and you will get all the suggested packages. Best, Jim > > -- output of sessionInfo(): > > . > > -- > Sent via the guest posting facility at bioconductor.org. > > _______________________________________________ > Bioconductor mailing list > Bioconductor at r-project.org > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor -- James W. MacDonald, M.S. Biostatistician University of Washington Environmental and Occupational Health Sciences 4225 Roosevelt Way NE, # 100 Seattle WA 98105-6099