The author describes a novel activity-based model which uses a simple competitive mechanism to form a topographic map and ocular dominance stripes simultaneously. The model does not attempt to take into account all that is known about the development of the visual system. Rather, its aim is to show how simple Hebbian learning combined with a suitable weight normalization rule can lead to the development of these mappings. The model forms an appropriate mapping when presented with distributed patterns of activity in the retina. This is much closer to biological reality than the activity regime used in previous models. The author considers input consisting of random noise filtered through a Gaussian as a first step towards investigating the effects of retinal correlations on receptive field and map development in this model. Simulation results on the monocular and binocular cases are presented.