As currently proposed, DOM parts makes it easier to build templating solutions using <template>. However, in reality, most templating solutions are intended to be used both in the client and on the server to produce initial HTML (aka "SSR").

As such, builders of templating libraries will need to build 2 implementations, one that uses DOM parts and one that does not. This makes the proposal a dead-end in my opinion. Most templating libraries will choose to build one solution that works in both contexts.

Given the amount of work needed to build this API, and the limited utility, it needs to be discussed if the work is worth the effort and long-term maintenance requirements.