Just copy/pasting my ranting thoughts here:
I honestly feel it would work better the other way around, since images can be compressed as thumbnails or linked in a URL and it keeps each citation from being spaced apart. That, however, does create the problem of capturing other user's (sometimes irrelevant) replies. Maybe we should resort to keeping only the relevant portions of the images. That way, only the important portion of Nitrome's comments can be quoted within the inline citation point, and if the reader wants to see the thread in its entirety, they can click the link to view the image and see the entire thread.
This is a bit of an issue I'm struggling with understanding though: images as citations is exactly like using the wiki itself as a reference, which is self contradictory. Then quoting comments has that same uncertainty. Direct comment linking is nonexistent, but it would have helped a lot in this situation because then users could see the actual thread themselves without having to rely on the wiki to give a replica of it. The whole "using citations as comments" system is just very unreliable at the moment.