You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
During our last community call, we talked about rolling releases (#419) but also got into ongoing tests, and especially A/B testing. Those are usually done in production with the experimental features behind some configuration and where the default is to have them off.
We could turn this good practice into advice to developers in the criterion Use continuous integration to be clear that any reuser shouldn't get an A/B test running by just replicating the codebase.
The text was updated successfully, but these errors were encountered:
The criterion this most relates to, is creating portable and reusable code. I'd like to see an implementation guide example where it is shown that A/B testing and deployment specific experimental features are put behind feature flags where individual deployments enable them.
ericherman
transferred this issue from standard-for-public-code/standard-for-public-code
Jan 16, 2023
During our last community call, we talked about rolling releases (#419) but also got into ongoing tests, and especially A/B testing. Those are usually done in production with the experimental features behind some configuration and where the default is to have them off.
We could turn this good practice into advice to developers in the criterion Use continuous integration to be clear that any reuser shouldn't get an A/B test running by just replicating the codebase.
The text was updated successfully, but these errors were encountered: