Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
Users corpus #738
This pull request includes feature that allows users to mark sentences as correct, incorrect or unsure. By doing so, each user will be building their own corpus within Tatoeba. The corpus of a user will basically represent their "linguistic world".
This is the first step for a more long term goal: once we have collected enough data about who thinks what is correct or incorrect, we can try to calculate a "trust" score for each sentence on which users can rely to decide how much they can trust that a sentence is correct.
What the feature looks like so far
Things to consider
Note: I don't expect to have an answer to all the things listed below before the feature actually gets released and used in real conditions.
Things to do next
@PaulPeer, you can simply report issues on Tatoeba's wall. We will do the necessary to log the issue in GitHub at the right place.
I'm aware of Horus not merging the reviews. This will be implemented in a later update.
Regarding the icons, this is another issue that I think you will also see on a Mac, not just on iPhone. It is not related to the new feature. You will probably notice other icons that are cut of (for instance I think the user icon next to the owner name is cut off).
I think the word "collection" is not a good choice.
One of the following words would be better. Perhaps others can suggest other possibilities.
rating / rated
I think I like either the word "rating" or "quality" best, since what we are actually doing is rating the quality of an item.
"Proofreading" is close, but usually also implies making corrections and not just reading.
I'd still like to see some way to indicate that a sentence has been read and but not rated.
I also think it would be very useful for people studying languages to know the difference between the following.