I guess it comes down to 'it would be a good idea to release their OS like everyone else does'. The question is in what sense and for whom? It could be good for consumers (maybe, though it would diminish even further the chances they would try out an alternative platform that would be better for them than Windows), maybe good from 'the way the world *should* work'. It could be good for companies like Lenovo, Dell, HP, etc as their costs go down and therefore they can price lower. It's not a good idea for Microsoft as a business though. It doesn't have much potential financial upside for them and represents sacrificing a boatload of their revenue. They already are doing precisely what you prescribe in mobile/tablets because they felt they had to, but taking it further in a general just wouldn't make a lot of financial sense. The exceptional case may be upgrades, I suspect MS suffers pretty high costs supporting older editions to try to get people to spend money on upgrades. That said, 8->8.1 was a free upgrade and 8 was considered horrible, and even so about 30% of the 8 users never bothered to do the 8.1 update, so there may be no good answer for how they could mitigate their costs without a risk of decreasing likelihood of upgrade-by-buying-new-device as an unintended consequence..
The point about share is that it continues to be the case in *SPITE* of being the more expensive option (though admittedly most of their users don't get direct control over giving them money). I don't say the secret to their share is charging, it's just that they manage to get away with charging even when free alternatives exist. It only becomes a 'good' idea from a business perspective if they are under some competitive threat that would force their hand to do what everyone else has been forced to do. That threat just doesn't exist, so they'd be trading current revenue stream for a relatively meaningless small bump in mindshare. I'm not going to enjoy this crappy OS more just because it's free, I'm still going to prefer Linux (not much chance of them making inroads into *THAT* 3.4% of the market no matter what). For consumers the cost of their OS license is already hidden in the system purchase, so they don't even give a single thought to all of this.
The business side is interesting and highly dependent on circumstances. While it makes little sense for MS to go free-as-in-beer for Windows, it makes a ton of sense for RedHat to more prominently promote free-as-in-beer use of their platform as they rapidly lost hearts and minds to Canonical. While they are being better about it (they used to go out of their way to make life tough for projects like CentOS) they are still being peculiar about it (CentOS v. RHEL instead of just 'RHEL for free'), but they are under pretty serious threat to their credibility as people selecting Ubuntu dilutes their image as 'the' authoritative Linux company. Here RH gives up the fantasy that they'll extract revenue from large chunks of the market, but protects their image. MS just doesn't have this problem. On the desktop, they have a lock. On the server their fortunes aren't so nice, but their platform is *SO* different it's unlikely to convince people to change one way or another.
I'm not saying MS is best or knows best or makes the best products. I'm just ignoring the technology and focusing on the chances MS would go along with such a plan based on current business realities and assessing how 'foolish' or not they are being by charging for their offering.