Ponyfills should absolutely be encouraged, I think there's plenty of benefits to using them. However, there's a valid use case of polyfills that gets ignored:
When you have nested dependencies that have less backwards support than you do they won't use ponyfills/polyfills. Now you can't use that particular dependency without using a full polyfill unless you fork the dependency (which could also be several deps deep).
I think package authors should be encouraged to use ponyfills longer than everyone else
I think polyfills should be limitedly encouraged as the correct solution to that particular problem
@ljharb brought this up in TC39 the other day...
Ponyfills should absolutely be encouraged, I think there's plenty of benefits to using them. However, there's a valid use case of polyfills that gets ignored:
When you have nested dependencies that have less backwards support than you do they won't use ponyfills/polyfills. Now you can't use that particular dependency without using a full polyfill unless you fork the dependency (which could also be several deps deep).