One can make "regions" of the two deep dimensions, FG & BG colors, into smaller "conceptual kinds/dimensions" such as "cold colors", "hot colors", and maybe "pastel backgrounds". Not all text is very readable with all combinations even if you have all 3..4 of your retinal cones. Another thing to keep in mind is that sub-color-dimension organization is easier with BG colors since many more pixels are lit up.
Someday the world may give us a few more easily noticed and independent text attributes like Serif vs Sans Serif fonts or something. Around a dozen or so is probably at the bounds of perceptual/attention/rapid access memory limits of humans, though. Even just remembering the whole 10-12 is not always easy. So, output dimensions are and always will be a very limited resource..not unlike "rows & columns", if you forgive the segue.
I have a 4K projector now, displayed on a 10 foot by 6 foot image (140 inch
diagonal!). The smallest font I use still limits me to 240x72 character cells
because one only sits so close and angular resolution of eyes in arc seconds
is what it always was. Meanwhile, my smartphone terminal emulator does like
61x19 windows. In my experience lc scales pretty well between these two
extremes separated by a (4x linear, 15x areal) factor. Both real estate and
output dimensions are arguably not getting much better pre-Singularity. So,
whatever tools we have should try to help manage these scarce resources.