I suspect many of the people reading HN these days aren't old enough to remember the "golden age" of Microsoft (or simply of "of age" at the time). Back in the 80s and 90s Microsoft was an unstoppable behemoth that people would describe as a battleship that could turn on a dime (IIRC this referred to Microsoft reinventing itself in the mid-90s so that everything was about the Internet).

Three things happened to Microsoft (IMHO):

1. The DoJ suit in 1997 was a serious blow to Microsoft inertia. Speaking as nothing more than an observer, it seemed that Gates lost his way. He simply didn't know what to do. From this I get the distinct impression he really believed Microsoft wasn't doing anything wrong and things like bundling a browser with the OS were just "innovations";

2. Microsoft became too hierarchical combined with more B and C managers. Once you don't have A managers in a software (or any tech) business it seems like it's all over. Nothing rots from within so deeply and completely as poor management; and

3. Ballmer had and has no business being CEO of a software business. This isn't simply due to him being a manager rather than an engineer but that certainly doesn't help. Gates was a programmer and actively participated in tech reviews in his heyday. Ballmer is the Carol Bartz of Microsoft. It's simply taking longer to kill Microsoft from within than it took Yahoo.

As far as stack ranking and these other things go: on the face of it I consider them symptoms rather than the root problem. The root problems are:

1. Microsoft, as an organization, is deathly afraid something will kill the Windows/Office golden goose;

2. Everything Microsoft does is done from the perspective of furthering the Windows/Office cash cow. It's why you hear Ballmer describe tablets as "just another PC form factor". He sees them as simply a means to an end and that end is selling licenses;

3. Microsoft is at the mercy of hardware OEMs that are shortsighted enough to run themselves into the ground taking Microsoft with them;

4. There doesn't seem to be any central strategy (beyond sell more Windows/Office licenses) or innovation; and

5. The "backwards compatibility" crowd lost out to the "breaking changes" crowd, which hastened the switch to Web-based applications.