Numerical models of corotating solar wind flows have enjoyed considerable success in simulating the evolution of shocks and corotating interaction regions (CIRs) in the region beyond 1 AU, but their performance with respect to stream fronts located nearer the Sun has been somewhat disappointing. In particular, they tend to predict erroneously that corotating shock pairs should occur relatively frequently within 1 AU, given the sort of sharp boundaries between slow and fast flows observed at stream fronts near 0.3 AU by Helios. We use an existing two-dimensional MHD numerical model for corotating flow in the supersonic, superalfvenic solar wind to show that the predictions of premature shock pair formation are due to improper specification of flow conditions on the initial surface (inner boundary) used as the starting point in such models. This faulty initialization leads to the generation of a physically extraneous strong compression along the stream interface just outside the initial surface, which results in the appearance of evolutionary artifacts (like spurious discontinuities) further on in the solution. We describe an initialization scheme incorporating flow conditions more appropriate to stream fronts near the Sun and demonstrate that it produces the smooth initial behavior expected on physical grounds. Thus free of the evolutionary artifacts, we see that the shear flow at the stream interface approximately balances the kinematic steepening near the Sun, which for typical input conditions keeps the corotating shock pair from forming before about 1.5 AU. We describe the criteria for shock formation in terms of the interface dynamics and show that the steepening process cannot be treated even approximately with conventional kinematic techniques. In a subsequent paper we investigate how the three-dimensional geometry of the stream front affects the dynamical evolution and the resulting CIR structure. ¿ American Geophysical Union 1989