Open serkandurusoy opened 10 years ago
Hmmm. Interesting.
Is this what sites like reddit and hackernews do? I always had the impression that they used a simple 'one person / one vote' rule.
I wonder if maybe we should surface these "prominent authors" in the site -- then users can just go an eyeball what packages such a person has starred as a great source of social proof.
Will ponder.
Reddit and hackernews don't account for user reputation. They often break news from anonymous (and very recent) users and they both implement a time decay algorithm that's very similar to that of telescope.
On the other hand, quora and stackoverflow do impose reputation based affect on their calculations.
And that totally makes sense becuase both platforms allow and thrive on their users' need for recognition. That's very logical for a Q&A type site where each good answer from a user gets to indicate another good answer to be.
I think packages do resemble answers more than they do news. Since they are actually answers (solutions) to common questions (problems/tasks).
Thousands of devs are flocking into the Meteor ecosystem with little to no prior experience with any of the atmosphere packages, only to discover relevant/good packages through lengthy re/search, trial and of course the scoring systems already provided within atmosphere.
Having spared a fair share of time on Meteor over the past year, I'm overwhelmed by the number of famous, join, modal, model, select, testing etc packages out there.
The latest starring system, having gained enough traction, should become a great indicator of quality for packages.
But a star that I grant to a package should most certainly be less relevant/important than that given by prominent package authors and contributors.
So, why not include this signal within the rating system? A package that is starred by one of those devs whose packages already have high number of stars/downloads is most probably a very good candidate among other similar packages to end up in my app.