1 of 14

Developing an internationalization plan �for WCAG 3’s Clear Language Guidelines

Updated in November 2023

2 of 14

Agenda

  1. COGA’s internationalization proposal (5-10 minutes)
  2. i18n feedback and other concerns (15-30 minutes)
  3. Next steps — how we can work together (Last 20 minutes)

3 of 14

COGA’s internationalization proposal

4 of 14

The challenge

It’s essential to internationalize WCAG 3’s Clear Language guidelines — as well as Making Content Usable 2.0.

We plan to develop conditional tests by language, but there are too many languages for us to try to do this for all languages before initial publication.

We need buy-in from the internationalization (i18n) team on a feasible work plan.

5 of 14

Proposing a feasible work plan (based on Rachael’s Github suggestion)

Before initial publication in WCAG 3

  • Make guidelines as universal as possible
    • Example: Use a simple tense and voice to aid understanding
  • Develop conditional tests for initial set of agreed-upon languages
    • See next two slides for our preferred initial set, plus other sets we considered
  • Create guidance for translating the guidelines into additional languages

Ongoing work

  • Continue to expand local plain language resources as requested
    • When WCAG is translated, include new conditional tests for Clear Language when needed
  • Can update conditional tests in WCAG 3 as we go
    • Outcomes are normative, only edited with full publication
    • Tests are informative, allow ongoing updates

6 of 14

Clear Language’s recommended initial set of test languages

5 initial languages for WCAG 3:

  • Arabic
  • Chinese
  • English
  • Hindi*
  • Russian

���

*Hindi ranks high for most spoken and low for most used online and is included in this list is to avoid exacerbating the digital divide.

How we selected this initial set:

We started with the United Nations' 6 official languages as our starting point.

To create a robust model, we removed languages that were from a similar language family. So French and Spanish are removed because they are similar to English.

The next default was commonly spoken languages. We added Hindi because it is the most commonly spoken language that is not one of the 6 official United Nations languages.

The resulting list is a diverse group of languages that includes:

  • Right-to-left layout
  • Tonal language
  • Vertical layout

This set covers major alphabets used by ~86% of people:

  • Latin alphabet: ~2.6 billion people (36%)
  • Chinese script: ~1.3 billion people (18%)
  • Arabic alphabet: ~1 billion people (14%)
  • Devanagari script: ~1 billion people (14%)
  • Cyrillic alphabet: ~0.3 billion people (4%)

Source: The EU’s WorldStandards

7 of 14

Other sets of languages we considered

4 official W3C �languages

5 most-spoken languages

5 most-used languages online

6 official United Nations languages

  • English
  • Chinese
  • French
  • Japanese

  • Arabic
  • Chinese
  • English
  • Hindi
  • Spanish

�John Foliot’s source: https://lingua.edu/the-most-spoken-languages-in-the-world/

  • English
  • French
  • German
  • Russian
  • Spanish

�Source:

https://www.statista.com/statistics/262946/most-common-languages-on-the-internet/

  • Arabic
  • Chinese
  • English
  • French
  • Russian
  • Spanish

8 of 14

i18n feedback and other concerns

9 of 14

Addison’s initial feedback (via email on October 26th)

Addison on generalized guidance vs tests in specific languages:

I like the care you’ve used in choosing representative languages, but I would want to understand better if such selection is strictly necessary. There may be some value in starting from generalized guidelines and providing a framework that any language or culture could use to address localized needs.

COGA thoughts:�Our plan has always been to provide generalized guidance. This summer we received feedback from i18n on our exploratory draft that made us think we also need conditional tests in specific languages to prove our guidance is robust. Did we misinterpret?

Addison on large- vs small-population languages:

We might also want to explore ways that other languages could contribute or, where appropriate, be used as examples when one or another language or script better illustrates or illuminates a clear language concept. I’m also wary of choosing only large population/high status languages. Smaller population languages often face struggles that illuminate problems more clearly.

COGA thoughts:

This sounds good in theory, but the challenge remains developing a feasible work plan. Can the i18n team help provide examples from smaller-population languages?

Addison on variations of Arabic and Chinese:

Two of the languages on your list (Arabic and Chinese) are not as well-defined as they might first appear. Both “languages” actually encompass a variety of writing styles or scripts and regional variations (some of which are languages in their own right and some of which are considered dialects).

COGA thoughts:

Any concerns about using the variations that the U.N. uses—Mandarin Chinese and Modern Standard Arabic (MSA)?

10 of 14

Other concerns

(Julie to add new concerns raised during the meeting)

11 of 14

Next steps

12 of 14

Next steps — How COGA and i18n can work together

  1. Agree on an approach (5 guardrail languages + i18n review at regular gates)
  2. Look at early drafts (use horizontal labels in GitHub)
  3. Help us recruit language experts?
  4. How to work on other localization issues (color, non-native persona, etc) that are related to COGA and internationalization

13 of 14

Engaging international groups/resources for this work

14 of 14

Current core working group