Better translation system

Feature Requests
1

Hi,

I am one of the translators. The problem about Logseq translation system is:

  1. It’s complicated to translate and submit (for new comers at least).
  2. Very difficult to find changes in existing English strings.

My suggestions are:

  1. Separate English string from other languages, so we can see the every changes easily on git repo
    • Right now I am checking time to time git repo for English changes with these commands:
      git log -p -- src/main/frontend/dicts.cljc
git log -p -- src/main/frontend/modules/shortcut/dicts.cljc
  2. Separate all language string to their individual files.
  3. Change language system so Logseq can use web based translation systems (Weblate, Transifex)
2

CC: @cldwalker any thoughts? or feedback regarding improving the current translation workflow?

1 Like
3

Hi. Thanks for the feedback

Separate English string from other languages, so we can see the every changes easily on git repo

We have a commandline tool for detecting new entries to English e.g. logseq/contributing-to-translations.md at master · logseq/logseq · GitHub. Are you checking logs because existing English entries are being modified? That is an area I would like us, the logseq team, to improve on. Ideally, any change to a modified entry should invalidate the entries of all other languages and cause their deletion

  1. Separate all language string to their individual files.

We have been thinking of doing this refactor and hope to soon

1 Like
4

We have a commandline tool for detecting new entries to English

  • I am aware that I am using it, but the script only informs me if there is a new string.

Are you checking logs because existing English entries are being modified?

  • Yes.

Additionally, I prefer to maintain the order of lines, similar to the English strings, which can also be problematic. I have to identify the new line, compare it to the corresponding string in my language, then add the translation and continue.

From what I can see, it seems that you all have good plans for translators. I will monitor these improvements.

1 Like
5

Dictionary files are separated now.

  1. Although, regarding bb - it is not comfortable enough.
    Great that it shows stats and missed translation.
    It can also export strings which require translation.

But correctly if my wrong, but flow it the following.
We have let’s say 500 lines of English.
Then 20 more lines appear in the middle of this file.

If I would use bb:

  1. I would export untranslated lines
  2. I would paste them in the end of some language file
    With such approach we would have not consistent files:
    English:
lorem
ipsum
foo ;; <- new entry
bar ;; <- new entry
blabla

Another language:

lorem
ipsum
blabla 
foo ;; <- new entry
bar ;; <- new entry

So in the future it would be hard to compare and search between english and other language file.

In my flow I’m trying to put new string to the same place as English to have consistency.

So if bb would sort lines in a file of another language the same way as they ordered in English - it would be great.

  1. Regarding translation please note this issue
    Translate Dates · Issue #5421 · logseq/logseq · GitHub

  2. Also a lot of things just hard coded.
    Hardcoded English text which is missing in dictionary · Issue #9504 · logseq/logseq · GitHub

1 Like