I haven't seen the explicit resistance, but I'd assume it exists, at least it just doesn't feel right to use non-nostr protocol for a nostr client. And all the points you mentioned are valid too, plus having to support yet-another protocol/API is just costlier for app devs.

I think it would be great to have many features as NIPs, but then there wasn't much demand for them before, and so investing in complex non-obviously useful NIP made no sense to me. Now that some features are starting to get traction my hope is that 'someone' comes up with a NIP and I'd just convert some of API endpoints to the new format.

'Trending' or other algo-generated lists seem like a simple enough concept to fit into a NIP, probably a top one on my todo list.