To the extent that there are basically no security implications (i.e. mapping larger public keys to smaller adresses), this reduction in combinatorial space is not dangerous. If it is not dangerous (and I probably agree) we could have simply used smaller keys (why then use secp256k1?). It does complicate matters for a potential attacker but it is certainly not advertisable as a security feature (suppose NSA can easily generate SHA256 collisions, then inspect the "public keys" for weak public keys, I dont think the signature check algorithms feature a check on how safe the key is).

In my original post I described how I feel that the original developer should have posted more design considerations, more documentation.

I can read a part of the code and concoct a reason why something is done a certain way. I can ask on the forums, but Satoshi (if it even is a single individual) is not around.

Simply coming up with an explanation does not mean it was the motivation.

At first I could only think up: an added layer of security, to only publish the public key once after which you never intend to use it again.
This made some but not enough sense to me (it is basically a cryptographic claim that it must be easier to crack secp256k1 public key, than it is to collide sha256/RIPEMD combo to generate compatible public keys, of which one might be much easier to crack)

Then I get a reply which makes a lot more sense: adresses become shorter for usability...

Then somebody elaborates my first interpretation and advertises it as a security feature...

This shows my prediction that I would like to see more DESIGN CONSIDERATIONS (think S-boxes) instead of having to resort to speculation.

The person/group that created bitcoin originally, could still elaborate their design considerations, perhaps write a book about it.

Now we have 2 reasons. Perhaps if we can find enough different explanations/speculations for why the public keys are hashed into adresses, Satoshi might feel the need to explain some of his decisions?