This kind of package management causes problems that are easily avoided...
Same Hmisc package version (3.16-0), but now that it's built on R 3.2.2, a wrench gets thrown in the spokes for allowing backwards compatibility to R 3.2.1... and this adds extra work of updating to the newest R, then they begin their workflow using their personal and shared library... and in the process, they install this package to the common library and now everyone who works with that common library has the error that Hmisc only works with 3.2.2
Backwards compatibility with 50 or so packages requires archaic workarounds, uninstalling, reinstalling, checking dependences, checking compatibility with current code and processes... This prevents use of R in a consistent manner within groups with shared libraries on networked drives in a windows environment as well as utilities like KNIME which run R snippets but don't access shared libraries... incompatibilities like this end up snowballing into multiple broken processes from a simple rebuild.
So how does one easily fix this? How do I get back the Hmisc package version of 3.16-0 that worked on R 3.2.1? Cause now that I've attempted an update to 3.2.2, there's the issue that the newest version of ggplot2, a dependency of Hmisc, couldn't be loaded!
library(Hmisc)
" Execution of R script failed: Error: package 'ggplot2' could not be loaded"
So this is more of a "please verify backwards compatibility" comment than a plea for a fix...
This kind of package management causes problems that are easily avoided...
Same Hmisc package version (3.16-0), but now that it's built on R 3.2.2, a wrench gets thrown in the spokes for allowing backwards compatibility to R 3.2.1... and this adds extra work of updating to the newest R, then they begin their workflow using their personal and shared library... and in the process, they install this package to the common library and now everyone who works with that common library has the error that Hmisc only works with 3.2.2
Backwards compatibility with 50 or so packages requires archaic workarounds, uninstalling, reinstalling, checking dependences, checking compatibility with current code and processes... This prevents use of R in a consistent manner within groups with shared libraries on networked drives in a windows environment as well as utilities like KNIME which run R snippets but don't access shared libraries... incompatibilities like this end up snowballing into multiple broken processes from a simple rebuild.
So how does one easily fix this? How do I get back the Hmisc package version of 3.16-0 that worked on R 3.2.1? Cause now that I've attempted an update to 3.2.2, there's the issue that the newest version of ggplot2, a dependency of Hmisc, couldn't be loaded!
So this is more of a "please verify backwards compatibility" comment than a plea for a fix...