by Eric Wagner (with Mariva
Tuesday, May 15, 2001
Giant newspaper headlines declare, "The Web is
Dead." As the dot-com industry implodes now as quickly as it had
exploded in 1995, it would seem that the Web is, indeed, dying,
if not already dead. But before we lower the coffin into the ground,
we might heed a slight stirring, the quiet sound of something unostentatious,
yet imperative, which may be about to resurrect our hope in networked
computer technology. The stirring, a phoenix whispering from the
ashes of the Web, may be the first sound of the next revolution.
IT development veteran Eric Wagner dubs this revolution the Paranet.
Spiders: merging metaphors
There's a spider
on my wall, meandering slowly toward my computer. It seems to be
lost. Perhaps it can't figure out where to build its web. I'm not
sure there is much for it to catch around here, anyway.
As I watch the wandering fuzzy arachnid, I realize
that the spider is like a dot-com. Just like the spider, the dot-com
can't figure out how or where to make its web or how to make a living
(generating revenue or ensnaring bugs) even when a web is woven.
But I won't continue this analogy. (For now.) It's just a bit too
Before I noticed the spider, I was thinking about
how the Web isn't really a web. It may seem like a web of connections,
but it's really just a collection
of links presented in different ways. Some color here, some
flash there, and voilį, you are, to use that (already) mixed
metaphor, surfing the Web.
What else is there to the Web? Sure, there are
"revolutionary" technologies that streamline proactive
channels, and envisioneer scalable infomediaries, as well as mesh
collaborative schemas (see the Bullshit
Generator for other useful and innovative technologies). But
is there anything really new?
Here are some ideas that could be new, or at least
unusual. Of course, even good ideas can be destroyed by bad
companies. In any case, these ideas could make the Web, well,
more like a web.
eXtensible Resource Name System: 100
Degrees of Solitude
According to its corporate home page, XDegrees
is "peering into the next wave of Internet apps." (Read
the entire InfoWorld article
from which this quote is borrowed.) This quote could in fact be
true; however, one couldn't be faulted if one assumed this to be
simply standard dot-com hyperbole.
I am not sure about the future of XDegrees, but
I am sure that its engineers are on to something: capitalizing on
the unused capacity of everyday, ordinary client workstations. Most
companies have a vast amount of unused file storage and CPU time,
not on the servers, but on all of the client systems in toto. In
fact, in many companies, this unused storage and processing power
far exceeds the amount of resources on their servers.
way to put to use the potential resources is with peer-to-peer networking.
Of course, this is hardly a new idea: peer-to-peer networking has
been around for years (remember Microsoft's gem, circa 1992, Windows
for Workgroups?) The new twist on peer-to-peer has been the addition
of a location repository (ą la Napster),
or a distributed-location repository (such as Gnutella).
But the uses of these systems are limited, because they are simplistic
The XDegrees system is based on something its developers
call eXtensible Resource Name System, or XRNS. XRNS extends DNS
to encompass all network information including machines, devices,
applications, and data. What is most interesting is that the system
is not location dependent; that is, a specific document or application
need not be stored in a single location. XRNS keeps track of the
multiple locations-but access requires only one address. This is
like combining DNS, load balancing, file sharing, and even a disk-array-based
Why is this so interesting? Because, with XRNS,
for the first time the power of client-server can be melded, synergistically,
with peer-to-peer. XRNS can be controlled centrally, although, unlike
client-server systems that rely solely on expensive central servers
for storing applications, files, and data, XRNS can use the power
of all network-attached systems. This type of network would add
an astounding amount of computer capacity-capacity that already
exists-to corporate, and perhaps worldwide, networks.
Project JXTA: The
Joy of Juxtaposition
future does not need us, as Sun Microsystems chief scientist
Bill Joy says. The future, however, does need Project
JXTA (or something like it). JXTA (short for Juxtapose) is an
open source Sun Microsystems project whose purpose is to create
a peer-to-peer network protocol that allows for fully controllable
distributed applications and data.
What is intriguing about Project JXTA is that it
represents the first attempt to create an open source (as well as
open API and protocol) peer-to-peer communication system. Although
peer-to-peer has existed before, it has been mostly limited to file
sharing and distributed computing, such as SETI@home,
which is a screen-saver, running on millions of computers worldwide,
that analyzes radio telescope data for signs of extraterrestrial
Project JXTA has the potential to completely change
the structures of the Internet and the Web. In many ways, JXTA combines,
even blurs, the Internet and the Web as never before. (Perhaps the
surfing spider isn't such a mixed metaphor after all.) Essentially,
JXTA allows for a fully utilized and multi-connected web of information,
data, connections, and ideas on a global scale.
Exactly what will come out of Project JXTA isn't
known yet. But it is not a stretch to imagine publicly-accessible,
large-scale distributed-processing networks whose computational
power up to this point has only been possible with supercomputers
or single-purpose distributed-processing systems like SETI@home.
The Semantic Web: What
Director of the World Wide Web Consortium Tim
Berners-Lee, in an article
in Scientific American,
introduces an idea that seems to have come straight out of a William
Gibson or Neal
Stephenson novel. Berners-Lee's proposal for what he calls "The
Semantic Web" is, much like his original concept of the World
Wide Web, a science fiction idea.
The last time you used a search engine to find
specific information, how long did it take? Most searches return
an exasperating amount of useless, unrelated, and irrelevant information.
For example, if I am looking for information about spiders (the
live, creepy-crawly type), a quick search will probably return a
mix of links about automated Web spiders along with information
about biological spiders. Try it for yourself: go to Alltheweb.com
and search for "spider".
A key feature of The Semantic Web is the supercharged
functionality it adds to XML.
The Semantic Web extends XML to include not only document data types
but also relationships between types. An example in Berners-Lee's
Scientific American article is the representation of a zip
code. A zip code could be called "zip code" in one document type,
but a "postal code" in another. The Semantic Web adds functionality
to XML that can link multiple names for the same type of information.
It also uses XML to create document types (meta-document
types, actually), that assist what Berners-Lee refers to as agents
to read multiple and disparate pieces of information and documents
in a way that dynamically pulls desired information from the multiple
In addition, The Semantic Web not only uses automated
agents, it almost humanizes them. The Semantic Web's agents, exploiting
the supercharged XML, are personalized search and organizational
systems that individuals can tailor to their own uses. Of course,
many tech development companies have, for years, claimed and boasted
that their products can tailor computerized personal assistants.
In this case, however, a fully specified protocol allows (requires,
really) the agents to communicate amongst themselves.
Much of what Berners-Lee describes exists today
on an ad hoc basis. The Semantic Web implements a common standard
for storing descriptive information-what Berners-Lee calls ontologies-that
creates links and relationships between all types of information.
Such a standard makes possible a fully connected web of information-information
connected in a meaningful way that could, for example, tell us how
These three ideas-XDegrees's XRNS, Sun's JXTA,
and Berners-Lee's The Semantic Web-are related in the sense that
they all venture beyond simple
extensions to the Web. But are they really new? One test of true
novelty is to see if there exists a buzzword for these ideas. If
none exists, then a concept is either novel or useless. No buzzword
for this type of network innovation exists yet ("The Semantic Web"
and "XRNS" aren't all that catchy). My suggestion is paranet,
prefixing "net" with "para-," which means "beside," "beyond,"
or "similar to." A paranet (lower-case for an insulated network),
or the Paranet (capitalized to represent the global, inclusive network),
could be the technology of not-too-distant future networks, based
on the networking technology we currently use, with similar
interfaces that run beside current networks and exceed far
beyond their current capabilities.
The spider is no longer on my wall. Perhaps it
has been inspired to build its web. The dot-goners,
I mean, dot-commers (including me), are no longer clogging the Silicon
Valley freeways. Perhaps they are about to create their paranets.
I have already begun weaving my own.
© 2001 Eric Wagner and Mariva H. Aviram. All rights reserved.