Napster

It was the first file sharing service most people ever used, it lasted about two years in its original form, and the name survives today on a music streaming business with no technical or organizational connection to the program at all. Everything peculiar about Napster, its speed, its reach and its ending, follows from the single design decision to keep the index in one place.

How a Napster search actually worked

A Napster search worked by asking one central index, not by asking the network. The program scanned a folder of audio files on the user's own machine, sent the list of file names and sizes up to Napster's servers when it connected, and that combined list was what a search query ran against. Results came back in a moment because the answer was a database lookup on a single system rather than a question relayed between strangers.

The audio never passed through the company. Once a searcher chose a result, their program opened a direct connection to the machine that held the file and pulled it across, one file from one person at a time. That split, a central index with direct transfers, is the reason Napster is remembered as peer to peer at all, and it is also the reason the design could not be defended: the transfers were between people, but the finding was a service one company ran. The glossary sets out the difference between an index, a tracker and a swarm.

Who wrote Napster and how it launched

Napster was written by Shawn Fanning, then a student at Northeastern University in Boston, who left his course to work on it, and the company around it was founded with Sean Parker in 1999. Parker went on to become the first president of Facebook a few years later, which is why his name recurs in histories of both. The program was free, it did one thing, and it handled only music, effectively MP3 files, which kept transfers small enough to finish over the dial up and early broadband connections of the time.

Adoption was fastest on university networks, where the connections were fast, the machines stayed switched on, and the population was large and concentrated. Several universities blocked the program on their own networks before any court had ruled on anything, because the traffic was crowding out everything else.

Why Napster was shut down

Napster was shut down because it operated the index itself. United States courts found that the company knew what specific files were on its system, since it held the list, and had the ability to police what it hosted, since it ran the servers. Filtering by file name was attempted and leaked badly, because a file renamed slightly slipped through. The original service stopped in 2001 and the company entered bankruptcy the following year. The reasoning of the cases, and why it shaped every design that followed, is set out on the lawsuits page rather than here.

The word illegal is doing loose work in most retellings. The program itself was not contraband, and copyright law differs by country and by case. What the American courts addressed was the company's own liability for a system it ran and could have filtered, which is a narrower question than whether any given person did anything wrong.

What Napster did to the record business

Napster changed the unit the record business sold. Until 1999 the practical unit was the album, because that was the object in the shop, and a listener who wanted one track bought twelve. A network searchable by track name made the single track the unit again, and the industry did not have a legal product in that shape. Litigation came first, licensed download stores followed a few years later, and subscription streaming eventually settled into the shape the market has now, where the catalog is central again but licensed.

The second effect was on expectation. A generation learned that a catalog could be searched instantly and completely, which is a standard no licensed service met for years afterward. The rest of the file sharing era, including eMule and eDonkey in Europe and the BitTorrent protocol everywhere, was built by and for people who had already had that experience.

Does Napster still exist today

No. The 1999 program has not existed since 2001, and nothing running under the name today shares its code, its company or its method. The name itself was sold out of bankruptcy in 2002 to Roxio, which attached it to a licensed subscription service, and it has changed hands repeatedly since, including a purchase by Best Buy in 2008 and a merger with the subscription service Rhapsody, which took the Napster name for itself in 2016. A person who signs up to Napster now is buying a licensed streaming account from a business that happens to own an old trademark.

What the Napster design left behind

What the Napster design left behind was a rule of thumb that every later network obeyed: do not run the part that can be switched off. eDonkey moved the index onto replaceable servers that held no content. BitTorrent went further and kept no index at all, coordinating one temporary swarm per file through a tracker that knows about that file and nothing else, and later making even the tracker optional. Syncthing shows the end of the same line of thinking applied to private use, keeping folders on a person's own machines in step without any shared list existing anywhere. Napster is the reference point for all of it, less because it was first than because its ending was so clearly a property of its architecture.