<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>App on Noetic Nought</title>
    <link>https://punchagan.muse-amuse.in/tags/app/</link>
    <description>Recent content in App on Noetic Nought</description>
    <generator>Hugo</generator>
    <language>en-US</language>
    <copyright>© 2006-2026 CC-BY-SA-4.0</copyright>
    <lastBuildDate>Fri, 05 Aug 2011 00:00:00 +0530</lastBuildDate>
    <atom:link href="https://punchagan.muse-amuse.in/tags/app/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>GitHub Cue: Recommendations for GitHub Repos</title>
      <link>https://punchagan.muse-amuse.in/blog/github-cue-recommendations-for-github-repos/</link>
      <pubDate>Fri, 05 Aug 2011 00:00:00 +0530</pubDate>
      <guid>https://punchagan.muse-amuse.in/blog/github-cue-recommendations-for-github-repos/</guid>
      <description>&lt;p&gt;If you aren&amp;rsquo;t already aware of it, I&amp;rsquo;m one of those people who&#xA;goes around saying, &amp;ldquo;GitHub is my Facebook&amp;rdquo;.  I spend quite a lot&#xA;of time on GitHub, browsing the work of various people, looking at&#xA;loads of interesting stuff that people built.  I keep jumping&#xA;between people pages and projects using the Watchers/Watching &amp;amp;&#xA;Followers/Following pages.  This way, I do come across interesting&#xA;people and projects, but the SNR is too low.  I wanted a better&#xA;way to be able to see stuff, that I find interesting.  That&amp;rsquo;s how&#xA;the idea for this Chromium app &amp;ndash; GitHub Cue &amp;ndash; was born.&lt;/p&gt;&#xA;&lt;p&gt;@baali and I hacked on this, during the last few days and got it&#xA;working.  It works as follows, (from the README) &amp;mdash;&lt;/p&gt;&#xA;&lt;ol&gt;&#xA;&lt;li&gt;Scrapes all the descriptions of the repositories being watched&#xA;by the user.&lt;/li&gt;&#xA;&lt;li&gt;Key terms are extracted from this description text using the&#xA;Yahoo Term Extractor.&lt;/li&gt;&#xA;&lt;li&gt;A list of languages is obtained, based on the languages of the&#xA;repositories, the user if watching.&lt;/li&gt;&#xA;&lt;li&gt;GitHub searches are performed for a combination of 3 randomly&#xA;chosen languages and 5 random key terms.&lt;/li&gt;&#xA;&lt;li&gt;10 random repositories out of all these, are shown.&lt;/li&gt;&#xA;&lt;/ol&gt;&#xA;&lt;p&gt;This is a very simplistic algorithm, but works decently for my&#xA;purposes.  Ideally, I would&amp;rsquo;ve liked to use a Collaborative&#xA;Filtering algorithm, but I found the data to be too sparse, and&#xA;the amount of computation to be too much to be done on the fly.  I&#xA;wasn&amp;rsquo;t really interested in pre-computing stuff and putting it&#xA;onto my server.  I settled down to the next best thing I could&#xA;think of.&lt;/p&gt;&#xA;&lt;p&gt;I would appreciate any further ideas and suggestions.  Thanks!&lt;/p&gt;&#xA;</description>
    </item>
  </channel>
</rss>
