Yep, the main problem is that whatever.com could be a completely different page for user A and B, and if they both commented on it, they'd be commenting about d

Yep, the main problem is that whatever.com could be a completely different page for user A and B, and if they both commented on it, they'd be commenting about different things, then user C goes to view it, and is confused. Which could be helped by storing the page source or something (which still has a ton of issues associated with it) but then you have a problem of discovery, and how do users tell the pages apart with the same URL? Simple version numbers? Another problem is you might have 2 URLs that are basically the same page, but with some query parameters that subtlety distinguish them - they are different URLs though. Should there really be two discussion sections for a page that's only differentiated by a referral link or a stylistic preference? Probably not, but there is not an objective metric here we can use to make the call. What if we use just the base path, e.g. facebook.com? It could be useful for small sides but for large sides you'll run into issues with scoping topics. it's a really hard problem IMO