Adding OS support for a new language

Further to Question 1 on language tagging: I am able to create a custom keyboard layout which is coded as Akkadian (Akk), and that generates a new category in Input Sources called Akkadian where this layout is found. Also when using this layout in the Pages app, I'm able to designate the language of a document as Akkadian.

Unfortunately I don't think any Apple apps can tag languages anything below the document level. LibreOffice can do this for text selections and paragraphs. But I think for it to allow Akkadian the app needs modification.

Hello, you can try Feedback...

Feedback - macOS - Apple

blindscholar wrote: if I have a bit of Akkadian text like the name Aššue-muballiṭ, it should be pronounced Ash-shoor Moo-ball-it (with a plosive t at the end). Since the English voice does not know what to do with the š or ṭ, however, it pronounces them as a regular /s/ and nothing respectively. It says "asser moo-bally."

You can get a small improvement by using the Slovak voice, which does know š. But I think no existing voice knows ḫ ṭ ṣ

1. There's no problem creating input sources, either latin or cuneiform, to generate Unicode text in these languages. But if you need such text to be somehow "tagged" as being in one of these languages, that's done (if available, which is not often) by the user selecting a menu in the word processor (or for html probably adding the required codes manually). To have your languages on that menu in a word processor might indeed require they be on the OS "language" list, I don't know for sure.
2. Spell check dictionaries are normally chosen manually by the user, so not problem having whatever language you want in one of those, regardless of OS languages.

Will come back on 3 and 4 later.

Further on the issue of having the Mac automatically change voices to correctly pronounce items in text tagged in different languages, I have come across an app which does this, Ghostreader +:

http://www.convenienceware.com/ghostreader

You can add its own language flags to any text and the app will read it using the right voice.

You would still have the problem of adding a voice that does what you need for Akkadian or Sumerian, but it looks promising.

I'm curious how you think adding these to the list of system languages would be useful. Just about everything on a Mac is independent of your system language, which mainly determines the text display of your OS interface menus and dialogues, which is not possible for Sumerian or Akkadian.

Are there some practical problems that you need help in solving now related to text, input sources, etc?

blindscholar wrote:

3. Tie into text-to -speech and allow for pronunciation with a custom Akkadian or Sumerian TTS voice (this causes a lot of problems right now)

Could you explain that a bit more? Is the idea that you have a text in Latin Akkadian and you need to add a voice to MacOS that would read that according to standard pronunciation system? Or Akkadian embedded in longer text in another language so you want the voice switch automatically from one language to another?

. Tie in to a braille code and present Akkadian and Sumerian text in the proper braille

Could you elaborate on this too? Is the idea that you have a Latin Akkadian text that needs to be sent to a braille display device? Is there a special braille system for these languages?

Thanks much for the further info and examples.

While MacOS auto detection for voice switching can work for Hebrew, I'm not sure it can do the same for short bits of another language that uses Latin script, even when tagged. I need to explore that further.

An alternative might be a bilingual voice, which would include Akkadian vocabulary as well as English and know how the former should be pronounced differently. I have seen such voices somewhere for English and Spanish.

More later.

blindscholar wrote: Similar with Braille. if you have text sent to a Braille code that does not include the characters, it just puts the Hex Code into Braille, which is not helpful.

Is there already a standard way to represent the special characters need for Akkadian/sumerian transliteration into braille patterns?