Even in scenario B, for most of the '90s and early '00s, a Windows-based PC was the presumed default platform. If you said "I want to buy a computer", it would

Even in scenario B, for most of the '90s and early '00s, a Windows-based PC was the presumed default platform. If you said "I want to buy a computer", it would normally be assumed that you meant Wintel. Scenario A, was, until relatively recently, only meaningful for very specific application categories: if you did DTP, you might buy a Mac, if you did video editing, you'd use an Amiga, and if you did research, you'd get a Unix workstation. But almost all non-specialist applications, and even most specialist software, was targeted to the DOS/Windows platform, leaving most developers just as much in control of their destiny as in the earlier iteration of scenario A, if not moreso, since they didn't have to expend capital porting their products to a half dozen different platforms and gamble on which platforms would survive. Platform vendors successfully 'commoditizing their complements' is a very new phenomenon, and one that's attributable to the vendors integrating their own distribution channels into the platform. Scenario C may be helpful for developers' independence, but desktop/native applications aren't going anywhere in the short term, and the threat of Scenario D - the web itself becoming a more closed, regulated platform due to the influence of Google/Facebook/etc. - still looms on the horizon. What might be helpful is an open, distributed software sales channel that can compete with the various proprietary app-stores across multiple platforms. Content retail needs to become a protocol, not a place. But this is probably a relatively easy technical project and a tremendous marketing challenge.