My understanding of the weird OS/2 marriage was that one early plan (that never happened) was that IBM OS/2 and MS OS/2 would be different SKUs. Microsoft's wou

My understanding of the weird OS/2 marriage was that one early plan (that never happened) was that IBM OS/2 and MS OS/2 would be different SKUs. Microsoft's would be "the cheap SKU" and probably the one most bundled by OEMs because of existing Microsoft relationships with OEMs. (Keep in mind there was a mutual antagonism between IBM and the other OEMs and it would have be weirdly unlikely for IBM to sell OEMs on bundling their OS directly; part of the "marriage of convenience" was directly Microsoft's sway and relationships with the OEMs.) Some IBM accountant would declare some features as too expensive or too "real business" for the MS SKU and Microsoft would have to cut them. In theory Microsoft could have said some Windows things couldn't be in the IBM SKU because IBM OS/2 would be for "real business" and Windows was a "silly toy" for "consumers" (until it wasn't, when 3.1 started to win big). (Though that never seemed to happen either, Microsoft wound up letting IBM OS/2 Warp include the Windows compatibility layers.) It was expected that despite the shared OS/2 brand name (IBM's dumb insistence) the Microsoft and IBM OS/2's might have wound up feeling surprisingly different. Microsoft never actually sold an OS/2 SKU so there's all kinds of speculation of what that would have been like and how it would have differed from the IBM OS/2 SKUs we did get. The hard-to-find preview versions like the article talks about are our only glimpses into whatever crazy world that might have been had the weird OS/2 marriage "succeeded". (It's also a weird glimpse, as the article points out, into the insight that Microsoft probably was ahead of IBM on their versions of OS/2 and likely were intentionally sandbagging how much wound up getting sent over to IBM's side of the fence and their SKUs.) It's definitely a strange fork in the world of computing history, if the "divorce" hadn't happened, if Windows 3.1 hadn't won over everybody, if Microsoft hadn't attracted Dave Cutler to start on the NT Kernel, and so forth: the operating systems we use would have been hugely different (and probably not for the better, given IBM's IBM-ness, the design-by-committee parts of OS/2, and some of the still underappreciated technical marvels of the NT Kernel that we did get in this timeline).