@screwtape
Still listening to this week's LispyGopher show. Now you're talking about whether CLIM is "about graphics" or "about present/accept".
Personally, I'd say these operate at different levels of abstraction so aren't in competition with one another.
One of the cool innovations of what was originally the Lisp Machine's "Dynamic Windows" and later became CLIM is this notion that present/accept are a very abstract way of saying that your program has stuff to give or receive without getting too into the weeds on how that will be done.
And yet when you render a specific presentation, it can't just be abstractly, you have to make concrete choices.
Graphics is one way to do it. Graphics offer a nice way to densely pack information in a way most of us have visual hardware to unpack efficiently.
Then again, one of my best friends is blind, and she'd likely dispute the efficiency of graphics as a way to communicate, and that's what's important about the presentation system--that it has rich enough understanding of what it's done that you could ask it, even after-the-fact, to re-present something in a different way that might be better accessible. And yet to do so in a way that doesn't say "different information needed to be presented", but merely "the information that got presented would be more helpful if it could be re-rendered based on different assumptions".
(Of course, it's been decades since I used CLIM. But I assume those kinds of things haven't changed much.)