Glossary unexpected non-wordforming characters

Question

A project I’m working on wants to have both singular and plural forms of the word for some glossary entries, separated by a slash. So the keyword will look something like this:
\k araba/arabaat\k*
But the markers check comes up with this error:

image.png917×31 15.9 KB

Yes, it is a non-wordforming character, and I’m OK with that, so how can I make the error go away? A couple of keywords contain commas, and it doesn’t complain about those, so why isn’t it accepting this other punctuation? I’ve verified that it has been validated as a character, validated in the punctuation inventory. I even tried adding it to the “Word break characters” list in the Language Settings.

What small thing am I missing?

Ancienne publication - affichée dans sa langue d'origine

Answer 1

So now I have a configuration problem. Before I made this change the only character in the “Other Characters” tab of the Language Settings was a “-” in the Word-medial punctuation field. In spite of this character being there, in the word list, words that had a hyphen were broken, e.g. “hukum”, “al” and “l” were in the word list, but not their combined forms. As soon as I added the “/” to the word-medial punctuation, I guess Paratext recalculated the word list, and it seems to have discovered that there are a lot more words in this language, including “al-hukum” and “l-hukum”, all of which are marked as unknown spelling state, a big problem for this project nearing complete Bible publication! To fix this problem, I had to move the “-” to the Word break characters field. But once I did that, the “-” is now marked as an error in the glossary keywords:

image.png731×24 10.6 KB

So is the underlying problem that the keyword SFM expects only wordforming characters? But it’s fine with spaces and commas, so that it can have phrases? Sounds kind of contradictory to me…

Ancienne publication - affichée dans sa langue d'origine

Comments

Do not move - to the Word break characters When you so that - will disappear when you view it in Print preview mode. Keep it where it is supposed to be, in Word medial punctuation. (Word break characters is intended for Thai and some other South-East Asian scripts that do not have spaces between words.)

Now that Wordlist recognizes - as a word forming character, you really should complete the Wordlist checks again. In particular you should use the Incorrectly joined or split words check to find inconsistencies in the use of the hyphen to join words.

However if you want to quickly approve all hyphenated words you can do that easily by filtering using hyphen and approving all of the words that appear. (If putting just a hyphen in the Filter box does not give you what you need to see, use regex:-)

Ancienne publication - affichée dans sa langue d'origine

Answer 2

Have you tried adding the / as a “Word-medial” punctuation in the Language Settings?

Ancienne publication - affichée dans sa langue d'origine

Answer 3

That does seem to solve the problem. What I don’t understand is why it doesn’t complain about the comma as part of the keyword, since it’s not in the Word-medial punctuation list. Is it a different kind of character somehow?

Ancienne publication - affichée dans sa langue d'origine

Comments

I didn’t look into the history as to why, but the check currently skips commas in the keyword of the glossary entries.

From the best I can tell, the checks were added for DBL.

John+Wickberg

Ancienne publication - affichée dans sa langue d'origine