google / google/codeworld

Predictive model for code

Open
#878 1 comment 0 reactions 0 assignees View on GitHub
discussion
Dominant language
Haskell
Stars
1.3k
Forks
201
PR merge metrics
No merged PRs in 30d

Description

This might be a bit outlandish, but one interesting idea is to train a predictive model, similar to what gmail uses here:

![80b29d50-a118-42fa-958f-039c99b657db](https://user-images.githubusercontent.com/544744/55049743-d4749b80-5023-11e9-925e-3a61859f97ff.gif)

This model would be for code, though. There are probably a few tricks we can play given that we know the grammar and vocabulary -- so, for example, we would only suggest completions that use defined symbols.

The argument in favor of this is that knowing what to type next is a big blocker for many students, and sometimes the answer is pretty formulaic. So when it's easy, let's make the suggestion for them, and let them just accept it.

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.