BitTorrent Misnomers and The Oscars

February 11, 2008 at 11:38 AM

After reading on about piracy and the 2008 Oscars, I found an interesting article about the limited availability of indie films on BitTorrent.

I tend to have atypical taste in movies, I found this article pretty interesting. However, I don't buy one of the explanatory arguments that it offers:

The very nature of the BitTorrent protocol itself tips the balance towards mass-market movie fare. BitTorrent has been designed to efficiently distribute massive amounts of data within a short time frame. Downloads get faster when there are more users available, and they slow down once people stop sharing. It's essentially a short-tail protocol that quite naturally works better for Oscar smash hits then for unknown Sundance flicks.

I personally don't use BitTorrent for downloading anything but open source software (like Ubuntu GNU/Linux), but my experience along those lines has made me realize that this is a gross oversimplification; it doesn't explain the lack of a long tail in BitTorrent movies, nor does it do justice to what BitTorrent accomplishes.

There is some truth in the statement: BitTorrent is designed to efficiently distribute massive amounts of data. It's purpose is to move some of the bandwidth burden from content distributors to content consumers. For example, If I were to offer a 100 megabyte home movie on this server, users would be able to download it. If 10 people each were downloading the file at 1 megabyte/second, I would need to have a 10 megabyte/second uplink to serve them completely. I would also wind up transferring 1,000 total megabytes of data. However, if each user downloaded 1 unique tenth of the file independently and then shared those portions with all of the other users, I would only wind up transferring the file once (albeit in chunks to different people). Since they would be downloading from eachother simultaneously, none of them would use their full 1 megabyte/second connection to download from me and I could serve them all with an uplink of less speed.

The efficiency from the content provider's point of view is dependent on the number of other users. However, the individual who is downloading my home video doesn't care if there are 6 other users or if there are 6,000; his or her home connection speed will limit the maximum transfer speed. The argument that less popular films don't appear because downloads are too slow isn't really valid. To be sure, there is a ceiling (related to the number of users offering a given file) below which download speeds will be very slow. This is because users' upload speeds are typically slower than their download speeds. However, once there are a handful of others offering/downloading a given file (say, 3-5 people), users will experience perfectly acceptable download speeds. And this process will always offer at least equivalent efficiency for the content provider as a traditional centralized download system, because centralized downloading is simply BitTorrent's worst case scenario (one person downloading the file from only one other person).

So, it's not a question of a bias introduced by the BitTorrent protocol. It seems far more likely that people simply download whatever is popular for a current year — a fact that I attribute to the Oscars being a yearly award (regardless of the quality of that year's films) rather than a Hall-Of-Fame-style honor. People latch on to the "best" of the current year, even if this means exalting mediocrity. To be sure, the same critique can be true of Sundance film awards, although I do think the mean quality tends to be higher in independent film circles.

Personally, I would love to see a system that compared film quality by year, and provided a ranking of films sorted by that comparison. I strongly believe that the power of suggestion affects viewing choices (and therefore downloading choices).