WHAT is the Internet? To most people, it鈥檚 the Web and maybe e-mail. But to a
large extent, these are to the Net what filmed radio shows were to TV: new media
imitating older media until people figure out what the new ones are really good
at.
The peer-to-peer (p2p) and Web services conference last month in Washington
DC was a strong reminder that the Net is still a work in progress. Peer-to-peer
came to fame because hooking people鈥檚 computers together without any central
server enabled them to 鈥渟hare鈥� music files (netspeak) or 鈥渟teal鈥� them
(recording-industry speak). Napster, p2p鈥檚 first incarnation, took off when
music files began disappearing from websites in response to threats from the
industry.
Cue John Gilmore, co-founder of the Electronic Frontier Foundation: 鈥淭he
Internet perceives censorship as damage, and routes around it.鈥� Gilmore鈥檚 famous
line alludes to the Net鈥檚 origins as a network designed to survive bomb damage
at any point.
Advertisement
But over the past 10 years, the Net has become increasingly centralised. The
Web underlined this by becoming a medium that media and corporate interests
could understand as publishing, broadcasting or retailing. As Net analyst Clay
Shirky pointed out at the conference, in many ways Web users are hostile to one
another because quick downloads are only possible when the number of people
logging on is limited.
Many of the Net wars of the past five years have focused on issues of
centralisation. The most obvious is the revamping of the domain name system (the
system that assigns newscientist.com to 快猫短视频, and which is, in
my view, the single greatest contribution to making the Web and e-mail
user-friendly). There鈥檝e been endless squabbles over the formation of the
Internet Corporation for Assigned Names and Numbers: its constitution,
authority, agenda, relationship with corporate interests and so on. But the
joke鈥檚 on those of us who cared about all this: none of it matters in what the
folks at the p2p conference called the 鈥減ost-Web world鈥�.
From instant messaging programs such as ICQ and Jabber, and file-sharing
programs that use the Gnutella network, to resource-sharing programs such as
SETI@home, none of the earliest harbingers of p2p care about domain names. ICQ,
for instance, identifies you by a user name and password. Gnutella doesn鈥檛 care
who you are until you download a file, and even then all it does is hook your
machine, identified by its IP number, to another machine, also by IP number.
The vision we took away from the conference is very different from the Net we
all know. For one thing, it assigns a new value to scattered bits of resources,
just as eBay turns attic junk into cash. And those resources could be anything
from the single file you need to make an outdated software program work to
elaborate, collaborative work.
Emerging technologies that make unstructured data and the Web
machine-readable play right into this: new services will marry scattered bits of
information and resources into something wholly new.
What will it all be good for? I鈥檓 not sure鈥攂ut it鈥檒l be fun finding
out.