Almost a classic! But not an easy one.
As a non-scientist (I studied design), I found it especially the introduction to modes of perception difficult to read. Although Ware writes well and brings so many examples, I’ve never read a “data vis” book that challenged me as much as this one. I’ve had trouble wrapping my head around all the V1’s, V2’s, neurons, etc. – while I appreciate that Ware went that deep into explaining all of that, my brain seems to refuse to understand anything about how it works 😞
I wish Ware had made even more a connection between the theoretical basics and its implication for information design. One or two sentences in the book let me believe that he assumed a reader who could make more of the connections themself than I was able to. Like when he writes “…and that’s why we are bad at xxx”, I sometimes thought “I wouldn’t have been able to draw that conclusion myself”. (I felt Munzner’s book was way closer to the pracitical side of data vis. Less “this is why this works”, more “this is what works” (about color, interactivity, 3D charts, etc.).)
I reeeally liked that point about “to become really good at something, you need to do it again and again and again to automate it” in the last two chapters. He writes on page 179: “It is so long since we first learned to do it that we have completely forgotten that stage at which the task required intense concentration. This is why the expert is often intolerant of the novice. He simply does not understand how an apparently obvious design flaw can be overlooked”. True words.
He also connects that to the idea of “we’re really bad at critiquing our own visualizations because we know them so well”. This is something I’ve been thinking about a while now, but I’ve never have it seen described as eloquently as Ware does. He talks a bit about how to learn to judge an own visualization well (e.g. not looking at it for two days; going through the visual queries that the audience might want to go through). I wonder if there are more tricks like that. I feel most data vis would be better if people knew how to see their own visualizations with fresh eyes.