In that specific case, the convention dictates that an observation log should be created with a set of quantities that have specific labels (in English). That’s just one example where localization/translation considerations will come up in conventions…
How should language translations / localization be handled more generally in conventions? Let’s use this thread to discuss.
Should convention IDs include a language code? eg: NRCS:pasture-condition-score-en:1.0?
Or, maybe some conventions will implicitly assume the language they are written in?
This question of localization is only a consideration where a convention requires you to enter specific strings. Quantity labels and units will probably be the most common places where this comes up…
Regional ones make sense to me, but language-specific ones seem a bit more fraught.
In the pasture condition score example, it seems like the ideal case would be to capture the scores with some machine-readable identifier, then allow for the possibility of displaying translated versions of the labels in human-facing contexts. That way the data between farms operating in different languages is still easily comparable and able to be part of the same larger datasets.
That said, I could also see an argument for translations of the conventions themselves to exist. They wouldn’t be separate conventions though, just a different description of the same convention. Similar to how we’ve talked about translating the conventions into machine-readable formats as well.
Yea that might be the best approach. The only downside to that right now is that there isn’t a mechanism in place for translating quantity labels into a human-readable format. So it would always show the machine-readable ID in the UI currently.
Agreed! This will definitely be necessary. It would be worth figuring out a strategy for this in the proposals currently developing in GitHub - mstenta/farmOS-conventions