[Rd] deprecate "freq" argument to hist

[Tim Hesterberg]

I request that the "freq" argument to hist be deprecated.

Reason:
One nice thing about SAS is is the consistency of certain inputs
across many PROCs.  In particular, they have "weight" and "freq"
statements.

We have added "weights" and "freq" to a number of functions in S-PLUS,
listed below.  I would like to see the same in R.

In porting the glmpath package to S-PLUS, we found one compatibility
issue - hist.default in R has an argument "freq", interpreted
to be !probability.  S-PLUS uses "freq" for frequencies in hist.

I request that the use of "freq" for !probability be deprecated,
to allow the argument to be used for frequencies in a future release.

Of the functions in R packages that are loaded by default, hist.default
is the only one I found with a "freq" argument.

--------------------------------------------------
These functions have both "weights" and "freq" arguments in S-PLUS:
hist
rowMeans rowSums rowVars
colMeans colSums colVars colQuantiles
cor, var, stdev
density
skewness, kurtosis
qqline
qqnorm.default
quantile

Some others have "weights" but not (yet) "freq", including
mean median ppoints tabulate.  Other functions like lm
have always had a weights argument.

--------------------------------------------------

Tim Hesterberg
Disclaimer - my own opinions, not Insightful's.