Functions to deprecate/remove

[droglenc]

We should consider removing the following functions as they are only peripherally related to the purpose of FSA and cause some maintenance or performance issues. For some of these we could create a vignette (or the like) that demonstrates how this functionality is better performed with functions from other packages.

• bootcase et al. ... was added because bootCase() was replaced in car with Boot(). Possibly keep confint.bootCase() etc but changed to confint.Boot(), etc.
• chooseColors() and palletteChoices() ... these should be updated ... there is much more info on properly choosing colors now than I was aware of when I wrote these. These should probably not be exported (i.e., should be treated as internal functions) as there are much better packages for handling colors.
• compIntercepts() ... this can be replaced with emmeans() in emmeans package. This does NOT appear in IFAR.
• compSlopes() ... this can be replaced with emtrends() in emmeans package. This does NOT appear in IFAR.
• diag() ... this is not used by any other functions ... can just be removed (i.e., moved to FSAmisc).
• filterD() ... I used this for teaching (automatically drops unused factor levels after filtering) ... probably not something we ultimately want to maintain.
• fitPlot() ... this was largely added for teaching purposes ... might be better to teach students how to make these plots manually so that they can transfer that skill to methods not supported by fitPlot(). This does NOT appear in IFAR.
• hoCoef() ... I added this only for the purpose of testing that b=3 in a W-L regression ... not sure it has any greater purpose.
• mapvalues() ... could just teach students to use plyr::mapvalues() rather than re-exporting it here ... this is not used in any function in FSA ... will need to write an errata for IFAR ... this may allow removing plyr Imports
• plotBinResp() ... I added this for a logistic regression example in my stats class ... pretty hacky ... probably better handled in a more manual way ... this is used in fitPlot() for logistic regressions.
• residPlot() ... same idea as fitPlot() ... this WILL require an errata for IFAR.
• Subset() ... has been deprecated for years ... should finally remove
• Summarize() ... another one I made for teaching ... probably better to teach group_by() with summarize()???

[droglenc]

Summarize() is used in several other functions in FSA, is used in several tests, may generally be a useful quick summary to users (similar to peek()), and provides some functionality not readily available in base R (e.g., number of valid observations). Mostly to minimize possible breaking changes to other functions I suggest we do NOT deprecate Summarize().