Add localised wordlists to allow passpharase generation in users' native languages.

[FabianVolkers]

Code Contribution Proposal

As a user I want to generate a passphrase in my native language, So I can easily remember and type it.

This user story results out of user feedback at a company in a non English-speaking country. For situations where passwords cannot be copy and pasted from the Bitwarden client but instead have to be transferred manually to a different system, users require words they know.

My proposal is adding localised word lists and adding language to the passphrase generation options. Code changes for the generation seem to be limited to libs/tools/generator and libs/common/src/platform. By keeping English as the default, the localised generation feature is introduced as a non-breaking change. In the client, the option should be presented as a dropdown of available languages.

Considering the dice logic used for passphrase generation, localised word lists need to be selected based on the right criteria. For the time being, I would implement the German word list from https://github.com/ulif/diceware

Are there any caveats I need to consider before implementing this change?

Related discussion and feature requests:

• https://community.bitwarden.com/t/choose-the-word-list-for-passphrases/3372/54
• https://github.com/orgs/bitwarden/discussions/5512

[HydraheadHunter]

I'd like to second this suggestion, though I would ammend the suggestion slightly.

Rather than have the dropdown menu that lets users choose between English, German, Spanish, etc. it should let users choose which language(s) to include in the word-pool: for example, ✓English, ✓German, ✗Spanish would generate phrases that are a blend of english and german words.

The default setting to introduce this as a non breaking change would then be ✓English, ✗German, ✗Spanish, ✗etc rather than English.

And yes to the custom words: I'm from the site of fine shoelaces and there some words in my lexicon that ain't make it into nobody's lib and dicts.

[bwbug]

For the time being, I would implement the German word list from https://github.com/ulif/diceware
Are there any caveats I need to consider before implementing this change?

I'm just a Bitwarden customer, but I would recommend allowing users to use their own word lists. If you are going to hardcode a word list into the Bitwarden clients, then I would presume that Bitwarden would want to vet the word list to ensure it contains no offensive words. The word list you have proposed implementing (wordlist_de.txt from github.com/ulif/diceware) contained the n-word until 2022, and when removing that word, the developer conceded that "there are certainly more words to remove and we must face the problem, that it is not easy to generate proper welcoming wordlists automatically".

By allowing users to supply their own word lists, you step around this problem.