Speed up downloading

[bartekch]

Right now list of results grows incrementally. This could slow up thigs when bigger sets of judgments are involved, should be rebuilt to avoid incremental growth. Take advantage of prior knowledge of number of results.

[bartekch]

It looks like it wouldn't be necessary, because R doesn't copy all elements of list when it is extended. Tests showed that when a list of apropiate length is prellocated an increase in speed is insignificant compared to time needed solely to communicate in API. For instance when we download 20k judgments the time has decreased from 720s to 690s, so a little. Therefore this issue will be put aside.

[bartekch]

Reopening, just to remember