Friday, May 11, 2018

Visual Studio IntelliCode

Microsoft:

IntelliCode generates recommendations by using a machine-learning model that is trained on thousands of public codebases – today it uses over 2000 GitHub repos that each have more than 100 stars to ensure that you’re benefiting from best practices. The model is used in your IDE along with your local code context to provide .NET related APIs that are likely to be the most relevant for you given the line of code you’re writing. We’ll be growing and improving the model over time so the recommendations will get better as we progress.

1 Comment RSS · Twitter

Martin Wierschin

This is potentially a very interesting idea for improving code completion. Xcode already does a great job with autocompletions, but sometimes it can be a nuisance wading through auto-completion chaff when it suggests a host of infrequently used symbols or methods. It would be cool to see if Xcode's suggestions could be improved using this technique.

That said, the predictability of an alphabetically sorted list is useful too. If this AI ends up often sorting oddball suggestions to the top it could waste more time than it saves. Well, this is definitely something to keep an eye on and see if it can increase productivity.

Leave a Comment