Extend glossary with forbidden terms

[danielnaber]

Example: One might add folder and its translation to the glossary. Often, this implies that folder should be used, and not a synonym like directory. These forbidden terms could be specified as a regular expression (e.g. director(y|ies)) and there could be a check which makes sure the forbidden words are not used.

[fastestnoob]

... or glossary just gives a message which says "you used another word for existing glossary term. are you sure?"

[comradekingu]

http://www.korrekturavdelingen.no/vanlige-skrivefeil_a.htm has a list of the words that most often end up being wrong, and the corresponding mistakes that pop up. It would be nice to set up a relational search for those.

Our Norwegian Bokmål glossary has entries accompanied by illegal and shunned variants http://i18n.skulelinux.no/nb/Fellesordl.eng-no.html

If you see a strike-through word in the editor, it should serve as a reminder that the word is not to be used :)

[nijel]

I'm not sure using regexp is a good idea. The regexp experience so far is that it's a powerful feature, which many users are afraid and thus won't use.

Scope of this feature:

• Add flag for glossary terms that this is a prohibited term
• Add check looking at translation whether it contains one of these

Possible future extension:

• Once we have rich editor, the forbidden terms could be highlighted there

[comradekingu]

I meant as an automated check. Write a word that doesn't exist, and it pops up asking you to fix it.
Not everyone has a spellchecker it seems.

It could be made impossible to save the string, and it would carry a flag if this check is avoided maybe?

The following is how I see dictionaries:

If the dictionary really is a database of single word strings and the type of word they are, it can be used to auto-translate and ensure consistency, with actually good results. Also it would make sense to work on the dictionary then, because it would be useful for the whole instance, and highlight missing words.
It also takes out most of the grind right away, as most strings are by law of entropy both short and repeated across a lot of components.

Call it the master source translation. Weblate already has a "used elsewhere" for translations.
There just needs to be a way of altering the "key" of duplicated words that are in a different tense or similar.(?)
Right now I have a basic idea of what words are used in what projects, but it doesn't scale, and the info does not populate well between translators. So either you have some consistency with one major translator (and consequently as consistent errors, or you have scaling towards but ultimately away from consistency.

Consequently, the way to start it is to take the entire translation memory of single word strings, sort it by frequency, and start weeding out the errors.

The problem with the regexp is there is no overview of what the available terms are. #4541

Don't mind me obviously, I am just talking out of my head, but then again it makes all the sense in my head.

[nijel]

@comradekingu Spell checker is option as well (see #443 or #977), but this issue is about prohibiting certain words. Typical example could be folder / dictionary mentioned in the initial post.

[github-actions]

Thank you for your report, the issue you have reported has just been fixed.

• In case you see a problem with the fix, please comment on this issue.
• In case you see a similar problem, please open a separate issue.
• If you are happy with the outcome, don’t hesitate to support Weblate by making a donation.