RE: Unicode characters used as images

This has always been the case in the publishing world. Centuries ago, the Greeks borrowed letter-like forms from the Greek language to represent math and science functions and operators. Examples: Pi, omega, Esh. My 1918 Underwood typewriter doesn’t have a key for number 1 and instead I learned to type a lowercase l instead. And I have a short lecture about the various dashes used in English grammar: from an em dash to a minus sign, they all look alike. 

 

It’s a very messy and unorganized part of our languages!

 

We use the same character to represent many things:

*        X = The actual letter x in Latin languages,

*        X = Multiplication in Maths,

*        X = Multiple doses in medicine,

*        X = A ballot box check.

*        X = A choice that represents “no.”

*        X = Twitter.

 

In my publishing classes, I call them lookalike characters: they look the same or similar, but are interpreted differently by sighted readers based on their related context.

 

This century’s old practice doesn’t work for accessibility. And thank goodness Unicode addresses this problem. Math symbols, for example, have their own glyph to represent their function. Capital Latin letter Esh Ʃ is Unicode #0149, but N-Ary Summation for maths is Unicode #2211 ∑.

 

The solutions that we find work best:

*        Use Unicode fonts and the correct Unicode glyphs for these special characters. In Word, Insert / Symbol shows the Unicode glyphs available on each of the fonts on your particular computer. Different fonts will have a different set of available glyphs.

*        If the character isn’t available, or isn’t voiced correctly by screen readers, use Actual Text (in PDFs) and the appropriate Aria in HTML (I no longer develop websites and can’t recall the Aria for this). 

 

As an industry, we should work to improve this for everyone involved: better authoring tools for those who create content, and better recognition of the glyphs by assistive technologies. Note that not all Unicode glyphs are recognized by JAWS and NVDA, which is a damn shame.

 

— — —

Bevi Chagnon  | Designer, Accessibility Technician |   <mailto:Chagnon@PubCom.com> Chagnon@PubCom.com 

— — —

PubCom: Technologists for Accessible Design + Publishing

consulting • training • development • design • sec. 508 services

Upcoming classes at  <http://www.pubcom.com/classes> www.PubCom.com/classes 

— — —

 <https://mailchi.mp/e694edcdfadd/class-discount-3266574> Latest blog-newsletter –  <https://www.pubcom.com/blog/2020_07-20/alt-text_part-1.shtml> Simple Guide to Writing Alt-Text 

 

From: Pyatt, Elizabeth J <ejp10@psu.edu> 
Sent: Tuesday, February 6, 2024 9:51 AM
To: Léonie Watson <lwatson@tetralogical.com>
Cc: S <Starry_sky@live.com>; Ms J <ms.jflz.woop@gmail.com>; w3c-wai-ig <w3c-wai-ig@w3.org>
Subject: Re: Unicode characters used as images

 

Having monitored this thread, I would observe that some characters just have multiple uses. These include x (and also times/close/delete), the . (period/decimal point), * (star/times), and the - (hyphen/minus).  

 

Sighted users are able to unconsciously distinguish the meaning based on the surrounding text. I recognize that this could be more difficult on a screen reader for a number of reasons.  Encoding the same shape to have different meanings is one way to account for this, but it probably won't account for every case such as the >  (greater than/forward arrow variant). 





I'm not sure what the solution is, but I do miss having an ARIA Alt mechanism sometimes. Maybe some sort of shape key?





Thinking out loud.





Elizabeth







On Feb 6, 2024, at 2:38 AM, Léonie Watson <lwatson@tetralogical.com <mailto:lwatson@tetralogical.com> > wrote:

 

 
You don't often get email from lwatson@tetralogical.com <mailto:lwatson@tetralogical.com> . Learn why this is important <https://aka.ms/LearnAboutSenderIdentification> 

 

It might be argued that "heavy multiplication X" is not the most user-friendly way to indicate that an item has been checked off a list, but I don't believe it fails any WCAG SC. The way Unicode characters are announced by screen readers is not always as useful as we might like, but they are announced and the names they're given are usually understandable if less than perfectly so.



On 05/02/2024 21:44, S wrote:

But that "heavy x" is not an image or ascii art and will not be interpreted that way regardless of the intent. And, it would be confusing to mark up text as if it was an image.  Suggestion is to use text "Y" and "N" as indicators for better cognitive recognition.  But if it has to appear as "x" for visual effect, then they should use an actual image with valid alt text so it is recognized accordingly.  

On 2/5/2024 7:53 AM, Ms J wrote:

Hello

 

If I had a shopping list and each item had a 'cross' next to it to indicate it was completed but the unicode 'heavy multiplication x' character was used for the cross, is this a failure of any sc?

 

They're basically using a text character like an image because they're using it for its physical characteristics, but they're not marking it up as an image (for example with an aria img role) and giving it an alt. It's read with JAWS as 'heavy multiplication x'. 

 

My thoughts are - could it fail name, role, value because it's used as an image but doesnt have that role?

Could it fail info and relationships because it coveys information visually but not programmatically? (But then thats like saying images with unclear alts should fail 1.3.1)

I don't think it fails sensory characteristics because there's no corresponding instructions that refer to it by its appearance 

 

Thanks

 

Sarah

 

Sent from Outlook for iOS <https://aka.ms/o0ukef> 

 





-- 
Léonie Watson (she/her)
Director
https://tetralogical.com <https://tetralogical.com/> 

 

=-=-=-=-=-=-=-=-=-=-=-=-=
Elizabeth J. Pyatt, Ph.D.
Accessibility IT Consultant

IT Accessibility
Penn State University
ejp10@psu.edu <mailto:ejp10@psu.edu> , (814) 865-0805 or 
accessibility@psu.edu <mailto:accessibility@psu.edu>  (main office)