Change name of delta argument on mlEstimationTruncDist()?

[wleoncio]

According to the function documentation, the delta argument provides, "indirectly, the difference between consecutive iterations to compare with the error tolerance". In contrast, if the user calls the function with print.iter = TRUE, they will see a "delta" number. Internally, that number is actually coming from a variable called delta.L2, which is calculated as follows (see lines 95 and 101):

TruncExpFam/R/mlEstimationTruncDist.R

Lines 95 to 101 in af75716

	delta.eta.j.plus.1 <- delta * grad.E.T.inv %*% T.minus.E.T
	new_eta <- eta.j + delta.eta.j.plus.1
	if (any(is.nan(suppressWarnings(natural2parameters(new_eta))))) {
	stop("Failed to converge. Try a smaller delta")
	}
	eta.j <- new_eta
	delta.L2 <- sum(delta.eta.j.plus.1^2)

I find this potentially confusing, since a user could set their own delta and wonder why the "delta" values in the output are completely different.

Therefore, I wonder if we should change the name of the argument (or, alternatively, the name of the output) to something more meaningful or clear.

One suggestion would be to rename the argument to delta.step.size, which is much more descriptive (though I don't really like how long that name is).

[wleoncio]

Solution:

• Replace "delta" only on printed output label to "delta.L2"
• Explain delta.L2 in function documentation (@note, maybe) as "sum of squares of delta.eta.j"