Closed krivit closed 3 years ago
This follows up from https://github.com/statnet/ergm-private/issues/261 .
What about edgenow
? It's shorter than status...
edgestate
is another slightly shorter option.
that seems like a good choice
It's trivial to change (since it's a search-and-replace of a unique string), so I'll go with it for now.
@drh20drh20 suggested edgeval
.
In the end, I am going with edgestate
: edgeval
is shorter, but it feels ambiguous in that it's not necessarily the current value; edgehint
is also shorter, but it's vague.
@smjenness , @ifellows , both EpiModel
and lolog
link to ergm
. If those packages don't define any new-style (c_
and u_
) statistics, they should not be affected---and as far as I can tell, they don't---but I want to flag this to avert surprises.
A c and a u function in the new API currently gets the following arguments:
Vertex tail
-- tail of focus dyadVertex head
-- head of focus dyaddouble weight
(valued only) -- new weight for the focus dyad(Wt)ModelTerm *mtp
-- model term(Wt)Network *nwp
-- network state (pre-change)Rboolean edgeflag
(binary) /double edgeweight
(valued) -- presence/weight of the focus dyad pre-change (i.e.,nwp[tail,head]
)For consistency, it would be helpful to have the last argument---the "hint" about the current presence indicator / weight---had the same argument name. There has been some discussion, which was tabled until the code was merged, so the time is now to make the change. Some possible replacements for
edgeflag
/edgeweight
:Any other ideas or preferences? @mbojan @drh20drh20 @chad-klumb @handcock @sgoodreau @CarterButts