The strategy elements that I'm suggesting that Microsoft might use with .NET has lots of precedents. Not just with other vendors, but it is what Microsoft has done with core markets (eg office) once it hits market saturation.
Furthermore if Microsoft starts facing a sea change where they begin losing control of core markets (office, desktop) in a big way, then you are likely to see big shifts in their behaviour. The fact that this has never happened in the past isn't good proof that it will never happen in the future. In fact Open Office already is putting serious pricing pressure on them. If it starts taking significant market share, Microsoft is going to have to adapt or die. And adaptation means a lot of transitions.
One of those transitions is that they have to become more selective in choosing loss-leaders whose purpose is to drive people to buy into core markets. Will .NET make a more selective cut? I dunno, but it is a significant risk that I won't accept without good reason.
Cheers,
Ben