This is a toy R package to serve as an example for statistical computing classes.
One purpose is to show how a regression like function (takes a formula and
fits a model) constructs the model matrix from the formula. All of the
code worth looking at is in the file bar.R.
The rest is just boilerplate to make a valid R package.
Another purpose is to show how methods of the generic function anova can
check that models are nested (as is required for validity of most tests of
model comparison).
It does not illustrate other aspects of formula fitting packages. For how the generic functions summary, predict, confint, anova and work, see the source code for summary.lm, predict.lm, confint.lm, anova.lm, or the analogous with glm instead of lm.
However, I have written a gist about some of the the things you need to know to write a regression-like package.