Comment by donohoe | Hacker News Reader

Comment by donohoe | Hacker News Reader

Parent Full thread

To be clear, the author is well-intentioned but wrong.

Nothing is so simple to be just black and white (see what I did there!)

I say this so that other HN users that influence the design of their products don't walk away and start making changes that negatively impact their users.

This has been studies and there is a ton of research. Here are a few with selected excerpts- please read the articles in full!

  "white text on black backgrounds creates a visual fuzzing
  effect for people with astigmatism called “halation”."

Source: https://www.essentialaccessibility.com/blog/accessibility-fo...

  "A research study found that “black text on a 
  white background overstimulates the OFF ganglion 
  cells while white text on black background 
  overstimulates the ON ganglion cells.” This finding 
  means that “white text from a black screen could 
  inhibit myopia, while black text on white background 
  may stimulate myopia.” The study advises against 
  reading black text on a white background due to 
  the striking effects of contrast polarity."

Source: https://uxmovement.com/content/why-you-should-never-use-pure...

A collection of further reading:

https://www.esa.org/communication-engagement/2018/08/03/reso...

Read the whole article, though. The author does not advocate for pure black text on a white background on a screen.

Your sources cite the useful study concluding that pure black text on a pure white background on a screen causes eye strain for folks like me with astigmatism. The uxmovement site, however, jumps to the conclusion that black text on a white background is the worst configuration and anything else that gets a good WCAG contrast rating is better, and sufficient for readability. That's unsupported, seemingly false, and the topic of this article. Their asserting the grey-on-white text in their graphic is more readable than the black-on-white is patently absurd.

The appropriate way to decrease contrast is by tinting the background but leaving the text towards the extreme. The author argues that the current WCAG method of determining contrast without considering other factors such as leading, tracking, type size, etc. gives passing scores to inaccessible results, and more recent methodologies are much more accurate, especially in dark mode.

Thank you Drew, exactly.

Your first two links seem to be about contrast polarity, not contrast ratios. (There is a mention of using low-contrast text in the second link, but it is not very good advice in the era of selectable light/dark mode. Those with astigmatism will tend to avoid dark mode, so the "hack" of using grey-on-grey to limit halation is not needed.)

The latter link provides some sources, but not much experimental evidence.

Yes thank you, this particular article was not at all about dark mode vs light mode -in fact it was mostly about light mode, with dark text.

The fact the article is currently displayed in dark mode is that medium really messing things up unless you set the page to #fff, which is visually fatiguing.

Nowhere in the article did I advocate for black and white — and far from that.

This article is part of research I am doing on readability contrast.

Also, "uxmovement" is about the last place to be taking advice on the subject.

Nevertheless, The subject is much more nuanced, but articles like the one you linked lead to misunderstanding.

Spatial frequency is critical, and it is spatial frequency that is the driver of contrast perception for smaller stimuli, particularly for small thin body text, where luminance contrast needs are substantially higher than for large bold headlines.

I said as much in the article.

For those who have this kind of issue you should try out darkreader. Aside from doing a great job of dark mode for any website, the customisation allows you to fix the shitty contrast on most websites so that you dont get headache from reading it.

Reply on news.ycombinator.com