The article identifies novelty and avoidance as separate drivers, but they're one mechanism viewed from two angles. The productive side tasks aren't just novel. They're safe. They generate evidence of competence without risking a verdict on the thing that actually matters.
The temporal discounting data is the most interesting part because it reveals something the article doesn't quite name: the older a task gets, the more it shifts from "work I'm choosing to do" to "work I should have already done." That transforms its emotional signature from opportunity to obligation, and obligations trigger avoidance regardless of whether you enjoy the underlying work.
The Zeigarnik point works against its own framing too. Multiple productive side projects create multiple open loops, all competing for the same working memory as the main task. The "productive" procrastination isn't just avoiding the main task. It's actively fragmenting the attention budget that the main task needs.
You're right, in my initial draft, I also took a side-step into figuring out the obligation aspect, but this ended up scattering the article too much. It's definitely a related part of it, the same tasks I love will become a chore as soon as I am obligated to do them. I love how you phrased it in your comment, transforming its emotional signature - very well put. That's what it is!