The major advantage of HTML5 is that you only have to write one app. If I launch a start-up that requires mobile client apps I can write a half-dozen clients (one for each platform) or basically just one with HTML5. I might write a native client for my preferred platform and hit the rest with HTML5 even if my long-term strategy is native everywhere. That is simple expediency and management of scarce resources. The broad reach of HTML5 is a major advantage so stop saying it does not have any. For non-mobile apps, the whole fact that you do not have to deploy anything to reach a customer is another simply gigantic win.
I agree that the tools to develop HTML5 are not as mature. They are improving though and you can even get integration local storage, devices like GPS or camera, and other niceties. The whole point of things like B2G and Gaia is to blur the lines between what is possible with HTML5 and native. Not everybody wants to be an innovator but there are some rewards for doing so.
Now, I am not claiming HTML5 is a panacea. I further your point by admitting that, today in mobile, I prefer native apps both as a developer and a user. If I wrote them all, every app might follow this strategy:
I also think that, in general, the best apps will always be native. I think that is possibly part of your point when you say "no good positive selling points" for HTML5. The thing is, there are other ways to differentiate your apps other than using every platform specific feature. Sometimes the differentiation has nothing to do with the app at all. The Kindle app for iPad is HTML5. Not only is it quite good but it would have to be truly terrible not to get used. People use it because they want to use Amazon not because of some exhaustive evaluation of similar apps.
I actually think it is pretty likely that MOST apps might eventually be written in HTML5. Certainly this has already happened on the desktop to a large extent. It will be the one platform pretty much every developer is guaranteed to know and that every device is certain to host. Unless you are doing something special, the question will be why giving up the benefits of HTML5 will be worth going native. The question is not "why HTML5" but rather "why not".