Isolating polylang?

I've already complained about polysynthesis before. I still think it's a stupid term. Anyway, here's one (very strict) definition I've seen for it:

1) polypersonal agreement
2) noun incorporation
3) extensive derivational synthesis
4) pervasive head-marking
5) verb-marking more than noun-marking

There's nothing about the phonological coherence in this one. Which, if I understand correctly, allows for the mythical "isolating polysynthetic language". Now this is a concept I've heard about before, in the back corners of internet forums and the like. I never understood how it was possible.

Then I met Abui. The author describes it as polysynthetic. Yet it sure doesn't look it. I think the most morphemes I've in a (phonological?) word is 5 and most of 2-3. Yet it's serial verbs allow for very complex verb phrases. While it's not isolating by any means (it's squarely in the agglutinating camp), it does show the diversity of "polysynthetic" languages and how the term really doesn't do justice. I'm sure if the average amateur (con)-linguist looked at it (even with glosses) they'd probably not label it as polysynthetic. Yet the author of this grammar was confident in doing so and I haven't seen anything disputing this. (Another fun one that I keep seeing brought up as polysynthetic (including by experts in the field like Michael Fortescue) with no discussion as to why it is classified as such. From my skimmings of the grammar, it sure doesn't look it).

In other news, I'm looking forward to the release of the Oxford Handbook of Polysynthesis which comes out in a few months.