SC 1.4.3 - Contrast minimum - note on evaluation success criterion using px - misleading px value

[goetsu]

https://www.w3.org/WAI/WCAG22/Understanding/contrast-minimum.html

note in this page say "The ratio between sizes in points and CSS pixels is 1pt = 1.333px, therefore 14pt and 18pt are equivalent to approximately 18.5px and 24px.'

14pt to px using the 1pt = 1.333px = 18.662px not 18.5
so I propose to change the content of the note to "The ratio between sizes in points and CSS pixels is 1pt = 1.333px, therefore 14pt and 18pt are equivalent to approximately 18.662px and 24px."
(or 18.7px if you want to round it)

Note : open after discussion in accessibility-luxembourg/ReferentielAccessibiliteWeb#2

[patrickhlauke]

personally, i really don't think that divergence of 0.16px matters in practice. also note that it's exactly because this was rounded down to the nearest half pixel that it says "approximately" (to indicate that no, it's not "exactly" 18.5px)

[goetsu]

@patrickhlauke
If the idea was to round it then it need to be rounded up (to 18.7 or 19px) not down in my opinion this way nothing below the real value of 18.662px will not be considered as compliant when it's not
for example BITV use 18.7 https://bitvtest.de/pruefschritt/bitv-20-web/bitv-20-web-9-1-4-3-kontraste-von-texten-ausreichend
cc @detlevhfischer

[patrickhlauke]

normatively, the value is given in pt https://www.w3.org/TR/WCAG22/#dfn-large-scale

the informative understanding document simplified the value a bit by rounding it. it's not an actual test procedure, like the BITV one is, just an informative piece of writing. but sure.

[goetsu]

Yes I know the understanding isn't normative but giving simplified information that could result in an invalid result isn't the best idea in my opinion

[mbgower]

The existing text and rounding seems fine, given that:

1. the correct formula is provided immediately before these examples
2. the word "approximately" appears directly before the values, and
3. this is an Understanding document, not a test procedure

That said, you are welcome to generate a PR to alter this. Otherwise, IMO this can be closed.