[R-sig-Geo] raster: use extract to get data from a categorical/discrete/factor raster

[Andy Bunn]

I have foo which is a SpatialPointsDataFrame and bar which is a raster containing land-cover data (i.e., bar is categorical).

The easiest way to extract the landcover for each point in foo is: 
    extract(bar, foo, method='simple')

However, I'd like to do get the majority value from the values of the four nearest raster cells - the categorical equivalent of: 
    extract(bar, foo, method='bilinear')
which is obviously the wrong move with a discrete raster.

Advice appreciated!