Some hereditary diseases, such as retinitis pigmentosa, lead to blindness due to the death of photoreceptors, though the rest of the visual system might be only slightly affected. Optogenetics is a promising tool for restoring vision after retinal degeneration. In optogenetics, light-sensitive ion channels ("channelrhodopsins") are expressed in neurons so that the neurons can be activated by light. Currently existing variants of channelrhodopsin--engineered for use in neurophysiological research--do not necessarily support the goal of vision restoration optimally, due to two factors: First, the nature of the light stimulus is fundamentally different in "optogenetic vision" compared to "optogenetic neuroscience". Second, the retinal target neurons have specific properties that need to be accounted for, e.g. most retinal neurons are non-spiking. In this study, by using a computational model, we investigate properties of channelrhodopsin that might improve successful vision restoration. We pay particular attention to the operational brightness range and suggest strategies that would allow optogenetic vision over a wider intensity range than currently possible, spanning the brightest 5 orders of naturally occurring luminance. We also discuss the biophysical limitations of channelrhodopsin, and of the expressing cells, that prevent further expansion of this operational range, and we suggest design strategies for optogenetic tools which might help overcoming these limitations. Furthermore, the computational model used for this study is provided as an interactive tool for the research community.