Thanks for the reply! It's nice to hear that a lot of devs are receptive to change and aren't treating the service hostilely. I like that there's categorization with expectations encoded in the category, I can see it helping a lot of things. A huge chunk of my reviews used to be in Code Collab, which... lacked certain desirable things. (But it got right a few others, so using it wasn't a totally bad tradeoff.) One of those I wanted was useful categories -- by default you just have review comments with the only alternative being a bright red Defect comment that would block closing the review until addressed -- but it was rather annoying to all parties and so few people used it. Different conventions arose to try and address the issue but they tended to be team specific. One of my own was prefacing those low priority things with "Nitpick: " or similar like in "Nit: add space between if and (". Another convention came from being able to make file-level comments that aren't targeted at a particular line, the reviewer might preemptively check off the file if they don't expect any action and it's just broader discussion/commentary, or leave that step undone until some sort of response (maybe just clarification) or possibly broad code changes has been done.
I've sometimes had trouble convincing some older programmers about using new (or even not so new but slightly 'advanced') language features like Java lambdas or Optional or generic types (juniors/interns were often aware of and happy to use the new stuff already), to the point that for specific individuals I'd just give up and focus my review efforts on other aspects. However if I ended up touching that code myself later on, or someone else did whom I was reviewing, I'd use those newer features/encourage the other person to use them if they weren't already, and if the original person ever came back to it they'd face an argument of local consistency against trying to change it back to using the old ways. It seems like the approach of longer-term encouraging and supporting a pocket of engineers pushing for better practices would actually work with this service given the ability to make notes for future reviewers. Another flaw in Code Collab was its atrocious search which made it hard to find prior reviews about a set of files.