Hacker News new | comments | show | ask | jobs | submit login

I will definitely create an auto-refreshing version eventually. For now, though, I'm trying to stay focused on my original goal. As I mentioned at the top of the article, I'm trying to create an entire directory of Clojure libraries. That directory will tell you about the relationships between libraries and help you to find specific libraries that could help your project. There could also be a list of featured projects, which will deliberately be spread all over the network of projects.

For now, I need to focused on improving my calculations, fetching all the data I need (what I showed in the article is just the tip of the iceberg), and finding better ways to infer relationships between projects.

Right, I didn't mean to distract you from the original goal :) The directory would be very, very cool to have as well. Sticking with the watcher/forker theme, perhaps the overlap between two projects' watchers/forkers (either in raw numbers or as a fraction of the total) might be a way to gauge their relationship?

Can't wait to see what you come up with. Best of luck!

I was imagining using the number of watchers and forkers as a measure of popularity. Looking at overlap based on WHO is watching/forking sounds like a better idea. Thanks for the suggestion.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact