Pavel N. Krivitsky

Formulas come with an environment in which their terms should be evaluated, e.g., `~nodefactor(x)`, where `x` is a variable present in the formula's environment. If the environment changes, the formula...

Looks like `d*sp()` and `d*sp()` terms are supersets of the non-directed versions, with an extra argument at the end. Is there any reason to keep their implementations and their documentation...

Organise help bibliography using Rdpack?

6

I did it experimentally with https://github.com/statnet/ergm.multi, and it's really nice: you just put a `bib` file in `inst/`, replace parenthetical and text citation with `\insertCite{key}{package}` macro, put `\insertAllCited{}` into the...

b*starmix() terms should autodetect when focus node does not have nonmissing values.

1

I.e., if we want homophily among the alter nodes only, and only nodes of the alter's mode have non-NA values for that attribute.

b1starmix(k, attr, base=NULL, diff=TRUE) and b2starmix(k, attr, base=NULL, diff=TRUE) terms are misnamed?

2

They count the number of k-stars with focus node in the specified mode (b1 or b2) for which the alter nodes are homophilous according to `attr`. This suggests that they...

Currently, they have the deprecated `base=` argument but not the new `levels=` argument.

Make proposals internal by default?

5

At the moment, their documentation is a part of the manual and the index displayed in HTML help. Since they are meant to operate "under the hood", we might want...

Currently, selection is based on class, reference, constraints/hints, priority, and shorthand weight. It may make sense to also allow a proposal to specify a function that takes as its arguments...

In particular, the developer needs to remember to check if the number of toggleable dyads is 0, in which case failure to initialise must be signalled.

It appears that the R side of the implementation doesn't currently take any arguments, and the C implementation only exists for bipartite networks. It's a potentially useful constraint, so we...