####What:

A ‘broad-catching’ client that will parse a variety of RSS feeds (or scraped HTML pages) for links to BitTorrents of video files and automate their download. In addition to providing a simpler interface than browsing a multitude of websites, the application will:

* Find popular shows and recommend them to the user (perhaps based on the number of users that are seeding a particular torrent)
* Find shows based on what your peers are also downloading
* Remember shows that the user likes
* Filter out unnecessary content (i.e., ignore filetypes like applications, and compensate for dead links)
* Parse titles and perform a search engine lookup to find descriptions for that title.
* Download torrents that are behind a walled-garden (many torrent sites require you a cookie to be set to enable a file to be downloaded)

####Why:

BitTorrent promises to revolutionize the video content delivery. But it is still relatively difficult to find good content. Legal issues aside, my broadcatching application will provide a more streamlined way to find quality video content, be it shared broadcast TV shows or independent VODcasts. By finding popular shows, it will allow people to discover the content that people are already showing interest in. This will let people take advantage of BitTorrent’s most unique quality: the more popular a file is, the more distributed the bandwidth becomes. This will also provide some interesting data to judge a show’s ‘quality’ by.

I don’t intend to limit it to the shows that are already popular. Ideally by providing a streamlined interface to find programming of all types, as well as links to their descriptions, lesser known shows can also grow in popularity, and likewise benefit from BitTorrent’s bandwidth efficiency.

####How:

Building upon my ‘Feeder’ application from last week, I will also need to

* collect a set of several test RSS feeds to parse and determine the different patterns used to link to the actual torrent file
* Parse data from either the torrent itself or the torrent’s tracker website to determine the number of users seeding a file (perhaps use proHTML library)
* Save user preferences; remember a history of shows that have already been downloaded
* Attempt to strip titles of extraneous tags (e.g., XVID.DVDRIP.SUPerIOR)
* perform google searches of torrent’s titles.
* either display title search result within the applet or link to site in browser
* perhaps implement a server-side script to defeat browser and cookie-sniffing
* design GUI to navigate the torrents that are found