That's all true, I just don't think it's as huge a factor as a lot of people think it is. If one OS market share is so small that the result is very, very few good apps being written for it, and many big-name apps are not available for it (like Twitter or Facebook or Candy Crush or whatever), that's a big problem. But up until that point, it doesn't really matter.

For example, IMO Windows Phone has a much smaller app store than iOS or Android, but it's big enough that the size of the app store is probably not going to be the #1 factor (or even in the top 5) when someone is deciding what kind of phone to get.

Consumers might say "I don't want Windows Phone because Windows Phone isn't appealing to me", but it's probably past the point where very many people would say, "I really like Windows Phone but won't get one because there aren't enough apps." Point being, Windows Phone is a very distant third behind Android and iOS in terms of market share. It's in the low single digits in market share, but that's still enough to get over the adequate-app-store hump (IMO at least).

So, I'm not denying that market share is essential to drawing developers, I just don't know if the market share extremes are ever going to get big enough to have an impact on iOS vs. Android.

My main objection is when people (and I'm not accusing you of this) slip into there-can-only-be-one-ishness, where there is an implicit notion that the market being discussed is a natural monopoly, and that the ultimate outcome is inevitably a market share horse race that leaves only one company standing. The argument is always that whoever has the most market share gets more apps, and whoever gets more apps gets more market share, resulting in a natural tendency towards a last man standing situation.

This was such a huge theme in the '90s in the computer industry- so many people just assumed that market share alone would cause Windows to completely take over to a much greater degree than it has. Mac OS would just vanish, and Linux would turn out to be just a fad amongst CS undergrads. Of course, Microsoft still has a monopoly, in the sense that it has 90%+ market share pretty much any way you look at it, but that's a lot different than 100.0% market share that people were predicting it would have. That leaves plenty of room for OS X and Linux and others to find their niche, and those niches are frequently more profitable and/or interesting than the mass market that Microsoft has covered so thoroughly. I wouldn't be surprised if the mobile world shook out in a similar way, long-term.

Of course market share mattered in the 90's. It was a huge problem for the Mac and Linux (and all those operating systems that died). They found niches but Photoshop and most other software , for example, was never ported to Linux because it never had enough market share to port to it. Photoshop was originally a Mac only app but the dominance of Windows meant that it needed to be ported.