Agreed. Thiago's introductory PARA writings do seem to imply a strict deadline, but like the parent, I use a concrete definition of completion ("completability") as the defining characteristic of a project. If I can't do this, then it's a major red flag and there's some deeper problem and the thing either isn't a project (e.g. it's an "Area" - something you to want to maintain a standard on over time) or it's not well defined enough yet and there's prior work needed.
Case study: Frederico Vittici of MacStories recently wrote about revising his own system to eliminate deadline times from his system as unnecessary overhead that he'd picked up long ago. It was a case of an item from "someone else's system" which was adding stress for no benefit.
All organizational systems are really custom-fit jobs. Look at others' systems and understand the individual techniques and how they fit together. Then apply those to your situation. This is a bit like the advice to not just take someone's complicated dotfile setup (vim, shell, etc.) and add it to your own wholesale. Instead to learn and apply piece-by-piece, understanding that sometimes several pieces must "go in together" to make everything work as intended. (this also applies 100% to every single software team's process in my entire career, btw.)