Thanks for sharing your workflow. Really helpful for figuring out what works for me, especially regarding the division of labour between DT and Zotero. I like your futureproofing mindset of not locking yourself to MacOS. But if I understand correctly, for that workflow to work, it is crucial that you let Zotero manage your pdfs and that means you can’t have you pdfs in the linked attachments folder but you need to let Zotero put it into its own folders, right? More specifically, I think this is what is required for the Zotero backlinks to work.
I don’t like this approach because I want to have all my pdfs in a single folder, which makes them easily accessible on any device, regardless of what software is on it. So I’m currently using Devonthink backlinks, basically trading your future-proofing against flexibility in how/where I store my pdfs - in a sense future-proofing in case I stop using Zotero 
Anyway, the next thing I need to figure out is how to best take notes while reading and highlighting the pdf. I know there are a number of good workflows out there, but they all seem to assume a linear work flow: once you’re done reading and highlighting, you do some extracting and converting and voilà, you have your highlights and notes wherever you want to have them (even the excellent script by @ryanjamurphy seems to assume linearity). But that is not how I work. I often go back and forth between multiple pdfs, reading and annotating a bit here and there and then re-reading and annotating somewhere else, so I’m looking for a solution where the annotation file is readily accessible in both DT and Obsidian simultaneously so that I can do my stuff in either DT or obsidian as it suits me.
Why can’t I just work in one of them? I’d love to, but Obsidian lacks pdf annotation and highlighting and DT is not as good for linking your thinking and building a second brain, so any thoughts that are not direct comments on the text in the pdf go into Obsidian, but in the annotation file for that pdf (so that they stays close to the text that triggered them).
So, I’m finally getting to my point/question: I just discovered that devonthink not only supports the usual pdf annotations (that are stored in the pdf) but also a separate annotation file for each pdf (in fact: for any file) located here:
So whatever you type into the Annotations field will be stored in a separate file. Markdown is supported (I read somewhere that it is MultiMarkdown v6, hope that is compatible with Obsidian…) But from what I understand, I cannot choose where the annotation file is stored (indeed, I think it resides in the database), so it appears that I can’t put it into the obsidian vault. Is this correct, or is there a solution?