If a user doesn't select any tags they are interested in, one of the buckets we draw from for follows are recently active users, which is in theory an indication of likelihood that they will produce interesting things to say more often.
It's likely you occasionally fall into this bucket.
I don't know if this is the right way to do things, and honestly we've just let this confusion linger too long. But it is the explanation.
It checks that they have at least a few comments, and that they have commented lately. This whole functionality needs to be rejiggered, but this path is not the primary path so anyone in this category gets a dribble of occasional followers. I think by making it a little more clear in the UI the difference between followers on onboarding and followers derived through the app would help the problem regardless of the algorithm.
👋 Hey there, I am Waylon Walker
I am a Husband, Father of two beautiful children, Senior Python Developer currently working in the Data Engineering platform space. I am a continuous learner, and sha
This is the file in the codebase where we determine who is suggested as a possible follow during onboarding.
github.com/thepracticaldev/dev.to/...
If a user doesn't select any tags they are interested in, one of the buckets we draw from for follows are recently active users, which is in theory an indication of likelihood that they will produce interesting things to say more often.
It's likely you occasionally fall into this bucket.
I don't know if this is the right way to do things, and honestly we've just let this confusion linger too long. But it is the explanation.
Does it at least check that the user to follow has a few posts?
It checks that they have at least a few comments, and that they have commented lately. This whole functionality needs to be rejiggered, but this path is not the primary path so anyone in this category gets a dribble of occasional followers. I think by making it a little more clear in the UI the difference between followers on onboarding and followers derived through the app would help the problem regardless of the algorithm.
Well, at least found the culprit :D
yes, good to know
Thanks for clarifying Ben!