Paratext 9.5 is breaking words where internally letters are separated with narrow no break spaces.

Question

Our project uses Arabic Script and is an agglutinating language. Because of our language's orthographic requirements, many words need some kind of non-breaking space within the word. Until the release of Paratext 9.5, we used a zero width non joiner, but with the advent of white space controls, we are trying to replace these with a more satisfactory narrow no break space. However, when I attempt to correct the spelling of a word in the word list by replacing the zero width non joiner with the narrow no break space, the word is treated as two words and no longer shows up in the word list. What would be the best way to move forward with this?

I really appreciate the Paratext Development Team's desire to incorporate white space controls into the program. I foresee this being a really big help! Thank you!

Answer 1

Currently, it looks like only the zwj and zwnj characters will work to keep the words together in the wordlist. When you press the Paragraph symbol in 9.5 it should display the various whitespace characters and allow you to see that you have a zwnj (even though there is no visible space).

Answer 2

I think that it's because the Zero Width Non-joiner (U+202F) is categorized as a Space Separator in Unicode. This makes Paratext treat it as a normal space. It's probably worth using Help > Give feedback to report this problem so it can be fixed by specifically treating joining-type and no-break spaces as word-forming characters instead of spacing characters.

Comments

Actually, the more I think about this, I think you might just be using the wrong character for what you are trying to do. I think the no-break spaces are only for layout purposes (i.e. the words on both sides of the spaces shouldn't be on different lines in the text). I think the joining-type spaces are actually what you should be using if you expect two letters separated by a spacing character to be considered one word.

---

Further research and it doesn't look like there is any character in Unicode that does what you are wanting. Any spacing character is designed to separate words. Any joining character is zero-width. You might have to use some other non-whitespace character inside Paratext (e.g. a dash) and replace them during publication.