UUIDs shown in translatable strings

I’m not sure if this is a MusicBrainz or a Transifex issue, but is anyone else getting the following in Transifex?


The string looks normal on the left, but in the editor the source string is full of UUIDs. I seem to be able to write and save a normal translation (without UUIDs), but I have no idea if this will cause problems further down the line.

4 Likes

I have the same issue since yesterday, earlier this week everything was fine. After selecting a string on the left, the normal text is visible in the editor for a short moment, until some words are replaced with these hex-string-identifiers. They seem to appear mainly for longer strings containing glossary terms – but not for all of them.

5 Likes

I hadn’t noticed the link with the glossary yet, thank you for spotting that! I have sent them an e-mail with the issue; I hope they can fix it soon.

3 Likes

Wow - real live translators. Is there any other place you guys talk? I look after the English to English translations and get confused as to the lack of ability to communicate with other translators.

It may seem odd doing English (UK), English (Australia) and English (Canada) but we don’t all follow those odd US spellings. :smiley:

2 Likes

Lol, I only talk about that here, but there simply isn’t much to say. You can leave comments on strings in Transifex, but few people read those, so it’s better to just make a thread here. Don’t forget to add the tag “translation” to your threads.

4 Likes

The main thing I want to do is leave a big note on the front of the English translations in case anyone else walks in to that section. Someone wading in may not realise that the English(UK\Aus\Can) sections have only translated a small sub-set of the English(US) to make it easier to keep track of. Only the phrases that need translation have been translated leaving the majority using the English(US) where the spelling is correct. This means it never says “100% translated”.

Any idea where I could leave a message like that?

I haven’t seen any way of onboarding new translators, I’m afraid. While I’m sure everyone has good intentions, it’s really annoying when someone you never had contact with pops in and start changing and translating things seemingly at random. The translation file for MusicBrainz server alone contains over 20 000 words, so maintaining consistency is not a small task.

1 Like

Sorry @mfmeulenbelt but I’m not in marketing so no idea what “onboarding” is. I don’t think it is relevant to what I was trying to ask. I’m thinking about how to contact the dozen editors already listed as translators in the English(non-US) sections.

All I wanted to do was leave a readable note on the specific language. I realise notes can be left on each translated phrase, which is why I was hoping there is a system where a note can be left at a higher level.

Here you can find all current members of the team for British English: https://www.transifex.com/musicbrainz/teams/13846/en_GB/ You can find links to their profiles and send members a message from there, but you’d have to do that manually and regularly check for new members. That’s why I wrote that having some kind of onboarding mechanism would be nice.

The translations are a little bizarre. When Picard first included the word “Color” I dived in and brought the English languages up to date. Most of that being deleting all those phrases that didn’t need translation making it easier to spot Colour, Catalogue and similar words. I then poked around a bit more and tried to fix related areas. It isn’t even really clear which of the projects are actually live.

It feels to me like wandering around a large library. Not quite sure if anyone is in charge. :smiley:

In the coming weeks I’ll try contacting a few of the names on that list. See who is still around. Mainly so I can get the list thinned down really. Make it easier to co-ordinate.

I was hoping that there would be a little bit more “collaboration” in this software.

I need to be able to stick a pin in the English (UK\Aus\Can) and set them as “completed” even though they only show a tiny percentage as done. I am worried someone will come through and just punch a bulk copy of all those phrases that don’t need translating just to get a 100% score :(.

That bug is very fancy and problematic. It also failed to save shorter strings that do not contain glossary terms. Might just be a temporary load issue?

As coordinator for the Dutch translation you should be able to, see Transifex Help Center.

:100:

For now, all we have are language-specific pages in MusicBrainz documentation, wiki-editable.

I don’t see how this would be a problem, quite the opposite. Source strings that localize the same should be copied to improve localization completion score.

I just set you as coordinator for the English UK localization so you should be able to.

2 Likes

I’ve sent Transifex an e-mail with the issue, so hopefully they are working on it.

I can see who is in the team, add team members and change roles. But I don’t see an option to even add a welcome message.

2 Likes

The English translations are a bit of a special case. So few words need changing that it is MUCH easier to keep track when only those phrases are translated. Then it is quicker to spot colour, catalogue, categorise, etc. When all the other phrases are also copied across it makes it hard to check.

The only reason I can see for copying “all” phrases it just to score a 100% score. This isn’t about scores, it is about accuracy. :slight_smile:

Thanks. When I get some time in the week I’ll see if I get any responses.

By following this approach you might actually miss other words that would need to be localized.

This list of words is not complete, and it might be impossible to make a complete list. There seems to be no way to reliably distinguish strings that do not need localization from strings that may need localization.

When a localization is near 100% complete, it is MUCH easier to check new strings for localization.

1 Like

I was writing a quick list off of the top of my head. Of course that is not complete. :slight_smile: Everything has been correctly localised. I was trying not to repeat the previous conversation on this from the thread from last year.

When I first approached this translation I downloaded the whole English (US) file as text and pulled it to my PC. And the “translated” versions for (GB\Aus\Can). This allowed me to start combing through it for differences in language. Slapping it in to a spell checker to also flag up words I may have missed. As well as doing a lot of other research online for dictionaries showing the Canadian and Australian differences to my native UK English.

There is a thread or two about this as the work was happening last year.

And I can assure you that my OCD kicked in and I didn’t miss phrases. :smiley:

When EVERYTHING was just sitting there in a “100% complete” format on the server it was full of errors. Many of the key words were not actually translated. A previous editor had just done a large copy and paste of all phrases on the “translated” side. It was clear that there were many errors. It was very inconsistent.

This is why I went the minimal route and stripped out those phrases that did not need translation.

That old saying “couldn’t see the wood for the trees” was relevant here.

Edited to add: I’ve now tagged the original English to English thread better so i don’t have to repeat that conversation again. :slight_smile: Translations to English variants

Sorry if I often come across as annoying on here, but I have run development projects in the past. Used to write network code for many of the big printer manufacturers. This included translations of the GUI into multiple languages. European, Japanese, English(GB\US\Aus\Can). In those cases we also went with the minimal differences in the English resources files as this also saved on the memory requirements.

This has genuinely been thought through. I’m not just an annoying noob :wink:

3 Likes

The issue has been fixed.

5 Likes