TODO list for the MuMIn package:

** WAIC

** fit.args - additional arguments for updated models (underwrite)
**  handling random effects in 'dredge' (with 'lme', 'lmer', etc). Currently
        'varying' may be used with lme's "random" argument for that purpose.
*   predictions from 'multinom'
*   untransforming predictions from 'unmarkedFit'
*   model.names: return a textual representation of the model
*   model.avg: checking for consistency of contrasts used in the models
*   print 'model.selection' output in a shorthand form, with model formulas as
        row names and no columns for terms.
*   pdredge: checking if 'nobs' for all generated models in are equal (already 
        done in 'dredge')
		
*   effects for "averaging" object?
