Hi
On 4/5/24 20:01, Juliette Reinders Folmer wrote:
In the "It decreases readability" section you make a sweeping statement
about accessibility, but don't back that up with research. Please back
your statement up as based on my understanding, the opposite is true.
For context: This paragraph was added by Larry. It's my name on the RFC of course and I approved of the addition he made. Nevertheless Larry might've additional resources to add here.
Case in point: if written in all caps, screenreaders will spell the
characters out - think HTML -. If written in Mixed case, screenreaders
will try to pronounce the word, making acronyms and other abbreviations
very hard to understand for anyone using a screenreader.
This is something which has repeatedly been pointed out, for instance at
conferences regarding conference acronym hashtags, like #DPC.
So, I'd be very interested to see your statement backed up by actual
research and invite you to look into this a little deeper.
Your remark looks at the accessibility from the PoV of a developer with impaired eyesight that is dependent on assistive technology. However accessibility also affects developers with regular vision. The concerns of these two groups might conflict.
W3C's accessibility guidelines point out that:
> Text in all capital letters is more difficult to read for most people, with and without disabilities.
(Accessibility Requirements for People with Low Vision)
Harvards's Accessibility Guidelines agree:
> Avoid using all caps. Readability is reduced with all caps because all words have a uniform rectangular shape, meaning readers can't identify words by their shape.
(Design for readability | Digital Accessibility)
Of course program identifiers are not regular text and indeed more similar to hashtags, in that they do not permit the inclusion of whitespace to separate words. Underscores would work, but that would just add to the inconsistency.
For Hashtags I found several resources pointing out to capitalize each word separately. For acronyms most resources pointed out that they should be avoided entirely (e.g. Human Verification, https://www.nyu.edu/life/information-technology/web-and-digital-publishing/digital-publishing/accessibility/how-to-guides/social-media.html#acronyms) and that recommendation is already part of the existing guidelines.
I've tested with my Android phone with the 'performHttpRequest' example and both variants where read out equally terrible as 'per-form age tee tee pre-quest' (i.e. merging the p with the request). And your example of HTML was read out as 'age tee em el' (or rather the German pronounciation of the letters) for both HTML and Html. The lack of vowels might've helped with the result, though.
The case of the boundary between two consecutive acronyms being unclear should affect developers with regular vision and impaired vision alike (i.e. PDOODBC and XSLRR).
From my personal experience as a human with regular vision (not even glasses), the variant of treating acronyms as regular words is much easier to parse.
Best regards
Tim Düsterhus
I'll also add