Reaction: Theory, Experimentation, and Application to the Development of Graphical Methods

The paper discusses that even after years of use, the graph design is still unscientific and then proposes a theory for the science of graphs through human graphical perception.

I like that the author first defines the elementary perceptual taks in order of ease of extracting more accurate data and then compares various graph forms against each other to show how they are used to extract the quantitaive information.

I strongly feel that the redesigns which authors proposed after applying their proposed theory, are of much 'perceptual value' and no doubt one can extract more quantitative information from them.For example, I had never seen framed rectangle charts before. When author replaced shaded statistical maps with these, I was able to extract more accurate data for the murder rate across various states in the US.

Author has agreed in the beginning itself that the list of elementary perceptual tasks is not exhaustive and they may not be distinct tasks. Hence, I think that there can be scope of even more better perceptual tasks using which we can discover other useful graph designs which may not have been discovered yet.