But if you are reasonably sure that they will be viewed in colors, it would be a shame not use such a great design option. (Of course you can’t antecipate how far the charts will travel and how far from the context they were designed for.)

And if you use a good blend of hue/luminance/saturation you will get good results in BW printing too.

I usually agree with you, but not in this one, I am sorry to say. As a dog, you can’t see colors as we humans do. You see some pale colors, a range of greys and you are color blind in the green/red axis. So, the way you see the world is not exactly similar to ours. We can’t provide best pratices…

Of course using colors is difficult. But some chart types are more difficult to understand than others (scatter plots, for example). Shouldn’t we use them? Or should we try to show how better they are for some tasks and, yes, educate the audience? After all, the world is in color…

We do have to know some basics on the use of color. Not always is necessary to use it. For example, in
this chart (http://bizviz.jorgecamoes.com/portugal-um-grafico-social-parte-2/en/) I don’t need color to show the data. But in this one
(http://bizviz.jorgecamoes.com/portugal-um-grafico-social/en/), with the same data, is interesting to use color to provide a focus+context view.

When there is a strong color code, like in the example you provide, how can you not use colors? For example, if you compare football results of Chelsea with Liverpool, would you use green for Chelsea and yellow for Liverpool? No. You use blue for Chelsea and red to Liverpool and you don’t need a legend.

So, it all depends. Or dog-pends?