E-mail List Archives
Re: Multilingual Web Accessibility Testing Approach
From: jp Jamous
Date: Aug 10, 2023 6:58PM
- Next message: Praful Sojitra: "Re: Multilingual Web Accessibility Testing Approach"
- Previous message: Brian Ward: "Canvas Studio and PaperCut MF breaking due to incompatibility with NDVA screen reader"
- Next message in Thread: Praful Sojitra: "Re: Multilingual Web Accessibility Testing Approach"
- Previous message in Thread: glen walker: "Re: Multilingual Web Accessibility Testing Approach"
- View all messages in this Thread
Hi Praful,
It is usually difficult to test accessibility with foreign languages. Even if you use Google Translate, there will always be things that would not translate properly.
I speak 3 languages. English, Arabic and French. Here is an example where Google Translate would fail.
Arabic: "Shoe Fee Ma Fee?"
English: "What's up?"
Google Translation: "What is in not in"
Yes, you read the Google Translation correct, because that what was said verbatim in Arabic. Another example would be:
English: "I need to run to the store"
Arabic: "we cannot use the words Need or Run, because they would sound weird in Arabic. We would have to say, "I want to jump to the store."
That is the reason why interpreters never translate verbatim rather they provide the point behind each phrase.
The concern with A11Y testing is you can get a link label incorrect if Google Translate converts it to English verbatim. Yet, in the native language, it is a clear link label. That is the biggest difficulty you are going to find with this.
As far as markup, that is no problem. Wave and Axe along with other testing methods would be just fine.
Videos is another impossible way of testing. You must convert the foreign language into something Google Translate can read. Honestly, I don't know of anything that can handle videos. Even closed captioning of movies are still being done by human beings as far as I know. AI might be something promising for the future.
Lastly, Arabic, Spanish and many of those old languages can have various dialects. While there is a common official Arabic dialect that can be understood by almost all Arabic speaking countries, there is a list of dialects that can differ completely from one another. Here is an example,
I speak Lebanese Arabic. I can understand Syrian Arabic, Egyptian Arabic, Jordanian Arabic. As far as Iraqi, UAK and Saudi Arabic, I would have to stop people often or ask them to speak slower. Some words and even alphabet letters are completely different from what I know. As we look at Arabic speaking countries in Africa, I have a very hard time understanding Libyans, Moroccans, and Algerians. That's how complicated it can be.
Sorry if I made it more difficult than you thought, but that is the nature of human languages. That is why Music and Math are universal languages. We all know them and can comprehend them. ð
I would encourage you to find someone that knows English and the foreign language you are testing. You might find some Ais that might help, but be very careful of dialects. They can change the meaning of a word completely.
- Next message: Praful Sojitra: "Re: Multilingual Web Accessibility Testing Approach"
- Previous message: Brian Ward: "Canvas Studio and PaperCut MF breaking due to incompatibility with NDVA screen reader"
- Next message in Thread: Praful Sojitra: "Re: Multilingual Web Accessibility Testing Approach"
- Previous message in Thread: glen walker: "Re: Multilingual Web Accessibility Testing Approach"
- View all messages in this Thread