Idea: A check for strings with max-length when the English string is too close to the limit

[emmapeel2]

This is a problem we have with one string. Most translations source strings are in English, and this is a famously short-word language. I do not have numbers, but related to Spanish I can give a ballpark figure of English having 90% of characters of the same string in Spanish. I am sure there is somewhere a list of this, or can be calculated from manuals, novels, etc.

I get that sometimes interfaces have limits, but one thing that gets very frustrating for translators is when for example if you have a string with a max-length of 200 characters, and the English source has 198 characters. It is impossible to fit the contents of English in a Spanish string of the same length. The English string should be shorter than the limit.

In this case, I propose, the DEVELOPERS should get a hint asking for a shorter source string. Maybe a failing check on the English source saying 'Source string is too long, most languages will need more characters than the max-length'.

Trigger a failing check when the English source is more than 90% of the max-length for the string.

This number could maybe be related to the translation languages for the string, or not.

[eighthave]

In my humble opinion, such a fuzzy check would open a can of worms and end up making things more complicated. I think the right approach is for the UI developers to design a UI where the max length limits fit with the languages they want to support. Devs are bad at this, but I can't see how having a fuzzy check will help make devs handle this better.

Then there are special cases like the app store metadata, e.g. summary, description, WhatsNew/changelog. etc. In those cases, the hard limits come from the app stores and can always be changed. So the translator should figure out what to leave out, if they can't fit everything in.

[emmapeel2]

In my experience, many (native English speaker) developers don't realise this limitation, but when you tell them that the string will be longer in other languages they provide a shorter English string.

This helps translators to translate the real meaning of the string. To take meaning out of the string is not easy for translators, they want to add all the content of the source string.

Also I think this check will help native English speakers to spot strings that may not be totally translated, and maybe they want to choose what will stay on the string before it gets to the translators, and not leave that decision to translators.

[eighthave]

I totally agree with you about the problems as you outline them. I think I misunderstood your proposed feature. The idea is just to provide a warning on the source string? E.g is the max limit is 50 and the source string is 49, then give a warning? I think that might work if integrated in the right spot.

[emmapeel2]

Yes, that is exactly what I mean. To give the warning about the SOURCE string. A bit like the 'Multiple failing checks' warning.

[nijel]

This should not be difficult to implement; it's just a matter of finding the actual limit to enforce here. Would it be in chars? Percent?

Once this is clear, please create issue for it: https://github.com/WeblateOrg/weblate/issues/new?template=feature_request.yml