This also assumes that the software must constantly evolve and `pull ahead'. In reality, a lot of the software in use usually evolves to a point of being useful

This also assumes that the software must constantly evolve and `pull ahead'. In reality, a lot of the software in use usually evolves to a point of being useful and does not need anything but routine incremental adjustments (I did not say improvements because one man's improvement is another man's useless feature), like compilers being retargeted, or typesetting software using new image formats, or a new language. Software becoming free (as in everything, beer and speech) is a natural progression because our patterns of using software are not that flexible so when a new use case appears, in a few years the free software catches up and it is no longer possible to abuse the customers as before (like Autodesk is doing now), just as it should be. Not that this model is bullet proof, unfortunately.