Why was this a "fiasco"? The result of the process outlined in that memo was, not to put to fine a point on it, one of the most profitable business decisions in world history.
I admire Gordon Letwin as much as the next guy, but OS/2 was never going to be the commercial success that Windows became, for basically the reasons the memo lays out: Windows 3.0/3.1 absolutely dominated the exploding PC market because it was targeted at hardware that anyone could afford. OS/2 was dependent on a level of hardware that wouldn't be commonplace until another generation or two of Moore's Law cycled through.
Microsoft simply didn't have that time to waste, and neither did their customers. The choice was between abandoning a phenomenally successful OS and API for the "next generation", or slowly building a bridge from that OS into the future carrying their customers with them. They made the right call, even if it did take roughly 11 years until Windows XP for the transition to truly be complete.
Yes, but OS/2 2.0 was not going to be released in 1990 either so it would be stupid to abandon Windows 3.x immediately anyway. And as I mentioned in the comments, nothing prevents the two from coexisting for a few years with OS/2 2.x being compatible with Win3.x.
Win3.x (OS/2 2.x would be able to run Win3.x apps) or both, but this problem would only last a few years at most. Eventually 286s and <4MB of RAM would become obsolete. (And BTW OS/2 2.x not running well in 4MB of RAM had more to do with the Workplace Shell, MS had similar difficulties with the Explorer shell too)
I admire Gordon Letwin as much as the next guy, but OS/2 was never going to be the commercial success that Windows became, for basically the reasons the memo lays out: Windows 3.0/3.1 absolutely dominated the exploding PC market because it was targeted at hardware that anyone could afford. OS/2 was dependent on a level of hardware that wouldn't be commonplace until another generation or two of Moore's Law cycled through.
Microsoft simply didn't have that time to waste, and neither did their customers. The choice was between abandoning a phenomenally successful OS and API for the "next generation", or slowly building a bridge from that OS into the future carrying their customers with them. They made the right call, even if it did take roughly 11 years until Windows XP for the transition to truly be complete.