JS20

It'll get updated soon :)

Two of my favorites: http://www.youtube.com/watch?v=eU3vqkPFJi8

<oops>

"I love you Mr. Bukkake. I love you and I love your ways."

Somebody was stomping glass bottles in the WROOOONG type of shoe!

Hundreds of thousands of these being created now. I have two computers running the compute-node VM to help render them :)

http://soundcloud.com/user18081971/18-ssba

The red plate is like the download button for inthesky

Did you see my previous post about aggregating track positions from imported playlists to obtain global rank? Voting on user playlists themselves seems like it'd be rather overwhelming, especially if it went into the thousands of lists. When you first mentioned SC playlist import, I envisioned it as being a way to count votes for each track based on overall "favoriteness", not so much just voting on the best playlist(s). Hypothetical: 10,000 users' ranked SC playlists are imported, each with an arbitrary number of AFX tracks. "28 Organ" is found in the #1 slot in 60% the lists, and #2 slot in the other 40% of the lists. 28 Organ therefore has 6000 votes for "1 of 228", 4000 votes for "2 of 228", and so on... That way, users could rank as many or as few tracks as they like so that every track they rank (even if they sort all 228) influences the total score for each track. I think the end result would be a highly un-biased list of tracks sorted by overall popularity, compared to the current vote results where bias has surely had an effect by now. The only way this hypothetical result would be identical to the current result would be if all users created their ranking lists to mirror the current voting tool results. All you would have to code here is the formula, then schedule periodic culling of all linked SC playlists doing a re-calc against the formula afterward. Assuming A means voting on lists and not ranking tracks, I like B more. If B, I think it should be random lists rather than human-influenced lists. Eras or styles might end up with huge imbalances in track counts, or create unpredictable bias. If random list population is appealing, then how about: 6 playlists filled randomly with 38 tracks each, where users may spend between 0 and 5 votes within each list, but no more than 5 votes may be used in any one of the 6 lists. Could also do 3/76/0-10 but that would have less diversity. In any case, 30 of 228 is 13.15%, but it might be interesting to see what happens when a given Top 30 (from the current voting tool) is broken into 3 or 6 smaller lists and the votes must then be evenly spread across the lists. If 10 of the tracks making up a particular Top 30 list happened to all end up in "List 2 of 6" after randomization, then they couldn't all be voted for in this new system because of the MaxListVotes=5 requirement. My $0.02

Apples and Oranges

Classic NSFW fireworks hilarity

So lucky only two fuses lit. I thought I was watching a trailer for Idiocracy 2 for a few seconds. http://www.youtube.com/watch?v=_OzwEUGu0hk

Pretty cool, albeit having a misleading title and tagline. Reminds me of the 2004 rat neuron flight simulator experiment, but they used actual rat neurons instead of software algorithms: http://www.sciencedaily.com/releases/2004/10/041022104658.htm

Updated MP3 and JPG files accordingly:

The last one has a few fan-mixes, which might be useful for testing if you want to add a func() to filter out ineligible tracks. http://soundcloud.com/jsmcpn/sets/afx-top-100 http://soundcloud.com/jsmcpn/sets/afx-playlist-bubbable http://soundcloud.com/jsmcpn/sets/afx-playlist-donkeyscare http://soundcloud.com/jsmcpn/sets/afx-playlist-ventolin-1-5 http://soundcloud.com/jsmcpn/sets/afx-playlist-ctrdj

The actual "World's #1 Dad"

XF, how tough would it be to make a separate page for a "Rate as many as you like from 1 to 228" system where a user could pick one SC playlist to be used as a ranking tool? Suppose you have many users' "My Favorite SC Trax" playlists imported to your platform, where the playlists are sorted in descending order from favorite to least-favorite - users could rank as many or as few tracks as they like (up to SC's 200-displayable limit). Could you grab the specific position of each track from each playlist and tally a global rank from that? Maybe it would be even easier to implement a track-sorting UI natively on your platform rather than rely on SC playlists. Here is my playlist - I ranked my top 100 favorites in descending order. If I were to rate all 228 on a scale of 1-to-5, this top 100 playlist would represent all the tracks I consider to be at least "three stars" or better. https://soundcloud.com/jsmcpn/sets/afx-top-100