Broken for iTerm2?

[JesseAldridge]

The page for this repo: http://www.yasiv.com/github/#/costars?q=gnachman%2FiTerm2

Looks different from the pages for these repos: http://www.yasiv.com/github/#/costars?q=edebill%2Fmodulecounts http://www.yasiv.com/github/#/costars?q=tomislater%2FRandomWords http://www.yasiv.com/github/#/costars?q=anvaka%2Fghindex

Not sure why.

[anvaka]

It is expected. The similarity calculation has high CPU complexity, and it is a function of number of stargazers. For popular projects, like iTerm2 to save time, I'm computing similarity offline, store it into json, and then display on a page. For less popular projects, that didn't get more than 200 stars in certain period of time the calculation happens when you open a page. That's why you see the difference. Also note, that for smaller number of stargazers the jaccard similarity may not give nice results...

[JesseAldridge]

Ah, ok. Thanks for the explanation.