It has been argued, for a number of reasons, that the next generation radio
telescope should be a multi-element interferometer with a collecting area
of about 1 km(2). The remaining parameters of such an instrument - frequenc
y range, angular resolution, instantaneous bandwidth, etc. - will be scienc
e driven. The requirements for propagation studies are briefly discussed, a
nd it is pointed out how variable-source confusion may differ from the norm
al variety. Finally, the Dutch project to achieve a large collecting area u
sing adaptive arrays of active antennas is described. A systematic approach
has been adopted, with the construction of arrays of increasing complexity
to test design features at each state. Recently, development of a low freq
uency array (LOFAR) has become an additional option. It would facilitate te
sts of some of the larger instrument's features, and provide real data on t
he influence of the ionosphere and interfering sources.