This is an excellent proposition. But if you come to think of it, at some point, it becomes outrageously uneasy to manage.
You made mention of the popular dapps being hardcoded. what happens when more dapps begin to originate, and then they become more popular, and consequently they feel they deserve to be top on that list instead of, say Busy.org. and let's assume they actually earned that spot better than busy. It becomes a battlefield of grief and uneasy rivalry.
Also, given that SMTs would be launched soon on the Steem blockchain, the number of these Steem frontends would increase drastically because more people would want to dive into the arena of building for this blockchain.
Consequently, you start seeing a massive influx of these dapps and any attempt to handpick the popular ones, no matter how objective and accurate they might be, there would always be an issue of supposed favouritism and inequality.
I think that is one of the main problems that might arise later, and it needs to be addressed before plunging right into this solution because trust me, then, it will become harder to manage and remain transparent.
If it has to be included at all, then it should be available for all dapps to make use of the canonical link. On the other hand however, this has adverse effects for th steem blockchain itself. All the attention and spotlight is taken from the default front-end, which I know for sure would hamper it from growing as far up the ranks as it should have.
I think the hard-code solution is fine for now when there are a handful, but it gets thrown out the window when the added size gets too large.
I think that what you talk about in the last sentence is actually healthy for steem and fine for all of us, if steemit is crap and no one wants to use it then let the market and the users decide which apps and sites become successful. I think it's actually an inevitability that steemit itself pretty much dies, considering how slow steemit's growth has been, how few users actually stay on the platform aside from the boom times, and the fact that historically blog and long-form platforms have all either ceased to exist or evolved into different models than the one steemit tries to emulate poorly.