The following describes some guidelines on how to produce colour figures that, when printed on a B&W printer of viewed
 by a colour-blind person still convey all the intended information.

 This can be achieved by using complementary encodings of information in figures.
 Moody in the "physics" of notations calls this "combining multiple visual variables".

 In practice, this means combining:

 *  colour (by preference dark, saturated colours as rendering them in B&W leads to
    dithering which makes graphical items --in particular text-- in unsaturated colours hard to read). 
     +  only use dark, saturated colours for text.

 *  for lines (including circumference of areas): 
     +  thickness
     +  dashed line style: solid, dotted, dashed, dash dot, ...
        https://matplotlib.org/stable/gallery/lines_bars_and_markers/linestyles.html

 *  markers: point, circle, square, ...
    https://matplotlib.org/stable/api/markers_api.html

 *  for filling of areas: 
     +  use different colours, bearing in mind B&W printing and colour-blind readability. 
        This typically means that the different perceived
        intensities of different area fillings are sufficiently distinct.
     +  use hatching: hatching style
        https://matplotlib.org/stable/gallery/shapes_and_collections/hatch_style_reference.html

 Preview colour images through

 *   printing/rendering in B&W

 *   viewing using https://www.color-blindness.com/coblis-color-blindness-simulator/

 I try to follow the above rules, wherever possible. 
 Plotting (e.g., matplotlib) and diagramming (e.g., drawio) tools provide (partial) support.


 An adaptation of the ACM Accessibility Recommendations for Publishing in Color:

 The most accessible approach would be to ensure that your article is still readable when printed in B&W/greyscale.

 The most notable reasons for reduced readability are:

 *    Most printing is still in Black & White.

 *    The most common type of inherited Colour Vision Deficiency (CVD) is red-green 
      (in which similar-brightness colours that only differ in their amounts of red or green are often confused), 
      and it affects up to 8% of males and 0.5% of females of Northern European descent.

 *    The most common type of acquired Coulor Vision Deficiency (CVD) is blue-yellow (including mild cases for many older adults).

 *    Situational impairments (e.g., bright sunlight shining on a mobile screen) tend to reduce the entire colour gamut,
      reducing colour discriminability.

 NOTE: It is NOT safe to encode information using only variations in colour (i.e., only differences in hue and/or saturation), 
 as there is bound to be someone affected!

 To ensure that you are using the most accessible colors, ACM recommends that you choose sets of colours to help
 ensure suitable variations in Black & White using either of the following tools:

 *  ColourBrewer: http://colorbrewer2.org/

 *  ACE: The Accessible Colour Evaluator: http://daprlab.com/ace/ for designing WCAG 2.0 compliant palettes.