Jini
by David S. Bennahum
Bill Joy, Sun Microsystems's President, co-founder, and arch-enemy of the
"other Bill" is riffing on a favorite theme: the evil of Bill Gates and
Microsoft. "Microsoft maliciously hijacked Java," Joy says, referring to
Sun's foiled attempt at creating a way for the same program to run on any
computer platform without modification. According to Joy, Microsoft so
modified the Windows version of Java that its promise of universality was
destroyed. "Microsoft is getting sued because they are not obeying the
compatibility requirements," Joys explains, referring to a Sun lawsuit
against Microsoft. "Microsoft was malicious," Joy repeats. Then he takes
the high-ground, a favorite spot for Sun, a company whose corporate
identity vacillates between the mantra "the network is the computer" and as
last-defense against Microsoft's plans for world domination. "Microsoft
has scared the bejesus out of everybody, and we are the only real
alternative," Joy tells me. "Steve Jobs sold out Apple to resuscitate the
company"-- an allusion to Microsoft's 1997 investment in Apple -- "We
are a total free agent. It gives us an opportunity to do something better."
The something better in question is Jini,
the cutely-named follow-up to Sun's 1995 launch of Java. Scheduled for
release this fall, Jini, as Joy puts it, is supposed to be "the first software
architecture for the coming wave of high-connectivity environments."
What this means in plain language is that Jini will let different kinds of
devices-- cell phones, laser printers, thermostats, desktop computers,
automobiles-- share information by communicating to one another. For
instance, a Jini-enabled air-conditioner could be connected to a home
computer, and from there to the Internet and your office computer. While
away at the office on a torrid summer day, you could control your air-
conditioner, telling it to switch on 30 minutes before you head back for
home, so you arrive in a cool house without having had to run the unit all
day long. Were the air-conditioner to break, it could send a self-diagnostic
message to the service company, listing what parts it needs, allowing for
quicker, more efficient repair.
The beauty of Jini-- in principle-- is that these devices don't need to be
programmed to speak to every kind of possible device on the planet.
Instead, much as a fax machine "handshakes" with another machine,
figuring out speed settings and image quality, Jini-enabled devices will
figure out how they can communicate and what they can communicate. So
by plugging in a Jini-enabled cell phone into a Jini-enabled printer, the
phone could send your stored address list onto a piece of paper without
needing any special one-of-a-kind interface. Jini acts as a match-making
service for anything with a computer chip and communications port. Two
appliances that want to speak use Jini to look up the right protocol, a
translation-table of sorts. Those protocols are written by whoever creates
the appliance and can be stored for easy retrieval on the Internet by Sun.
Since the protocol is in Jini, the second appliance can understand how to
communicate with the first. It's a system not unlike the way Internet
Domain Name Servers turn a string of letters, say www.sun.com into a
string of Internet Protocol numbers (192.9.49.33 in this case), numbers
that are actually what's identifying the destination of packets of data going
in and out of your computer as you surf Sun's Web site. According to Joy,
what's made Jini possible is the pervasiveness of the Internet, combined
with the supposed propagation of Java. The former creates a universal
communication network for appliances to hook into, the latter is the
programming language that Jini is written in. The payoff, Joy argues, is
potentially huge.
"We're talking about something that can be put on a stapler, a tennis shoe,
any device in the office," Joy explains. It's a strategy similar to the "Intel
inside" campaign that Intel successfully launched in 1995. Every personal
computer with an Intel chip bears somewhere the Intel logo of a swirl.
Similarly, Sun imagines a future where myriad consumer products, from
stereos to refrigerators and cars, bear a Jini logo (scheduled for unveiling
in January). Like Intel, Sun will not dictate the kind of software that can
be created with Jini, nor will it claim a commission on sales (although it
will collect small royalty payments earmarked for trademark protection).
Instead, Sun's revenue will come from building new Jini-based "network
services," according to Mike Clary who is responsible for managing the
Jini project.
Jini's launch is nearly a perfect copy of the company's 1995 hugely
succesful launch of Java. Back then, Sun released Java without fanfare,
simply putting up the source code, documentation, and some sample
programs on their Web site. Within weeks the buzz online was that Java
would become the de-facto lingua franca of the computer industry, finally
ending the division between PCs, Macintoshes, mainframes, workstations,
and anything else with a computer chip in it. Instead of Balkanized
systems, Java promised "write once, read everywhere"-- meaning a
programmer could write one set of instructions and have them run on any
computer with a Java "engine," or programming environment, something
close to an operating system. While Java first worked as cybernetic
voodoo in 1995-- within a few months of launch over 100,000 people had
download versions of the Java development kit, Sun's share price had risen
sharply, and Netscape's had plummeted while Microsoft's stagnated-- the
Java reality-distortion field quickly weakened.
Java programs were slow to run, and relegated to animated sections of Web
browsers, a form of CD-ROM multimedia lite for the Net. Java's fatal
flaw was that it required significant customization from system to system.
Instead of "write once, read everywhere" programmers often found they
were back to "write once, read once." Crucial to this failure was the
complicated nature of linking Java programs to the "user interface"-- the
windows, icons, and desktop environment-- of different systems. Netscape
wound up creating a "flavored" version of the Java environment in its
browser, thus forcing Java programmers to make their code a little less
universal. Microsoft went all the way, dousing Java with so many
"flavors" that it became an entirely different brew. This led to Sun's suing
Microsoft for violating the terms of the Java developers license, which, in
part, requires developers to stay true to universal Java-compatibility.
Instead of Java becoming a "paradigm shift" as guru-marketing types like
to say, it became yet another example of bold, visionary computer
dreaming interrupted by reality, what others call vaporware. Jini could be
Java 2.0, another iteration of a big, bold idea that evaporates once it's
actually taken out of the lab.
When I ask Joy whether Jini's headed down the same path Java took, he
points out that this time around Sun is not dealing with rival computer or
software manufacturers. The audience for Jini comes from another
universe: consumer goods. "The companies that are doing the pagers and
the personal digital assistants do not have the weird political intentions that
companies in the computer industry have. They are more willing to pick
up a solution and just use it," Joy says. "This is like Java but breaking out
of the computer industry. We are talking to car companies. It is Fortune
500 corporations we're talking to, not the top 20 computer companies."
To date, Sun has confirmed publicly that Epson, Canon, Seagate, Quantum,
and Federal Express have agreed to embed Jini in some of their systems.
Equally important to Jini's survival, and far more subtle, is the fact that
disk drive makers, car companies, and pager manufacturers tend to be
skilled at user-interface systems that have nothing to do with mice,
windows, or icons. Once you're out of the realm of standard PC interfaces
dealing with things like LCD screens on telephones, buttons in a car's
dashboard or on a printer, you are in the territory of age-old ergonomics.
It's the difference between the genius behind the buttons on a Sony
Walkman and the idiocy of a "dialog box" in Windows 98.
If Jini can let people easily and intuitively access powerful computer
networks, and myriad arrays of appliances-- say by going to the touch-
screen on your fridge you could somehow tell your car to start the engine
and warm up on a freezing morning-- as easily as one picks up the
telephone and places a call today, then Sun will have product worthy of the
moniker "breakthrough." Tactically speaking, however, it doesn't matter
if Jini fulfills on its high-tech promise. On the Internet "mind share" and
"market share" of "attention" are the benchmarks by which companies are
deemed successful, allowing Amazon.Com to be worth more than Barnes &
Noble, even though it has a small fraction of latter's sales. In that world,
Jini is poised to do some Java-like mind-warping. With a strong release
this fall, measured in thousands, perhaps millions of Jini development kit
downloads, coupled with bold press releases from consumer companies
announcing their intention to embrace Jini, Sun's stock should get a nice
lift. Whether GM can ever use Jini to sell cars hardly matters. In a
universe where hype begets billions, Jini looks like a winner.
Subscirbe to MEME via e-mail.
Visit the MEME archive with issues 1.01 forward.
Search the MEME archive for anything you want.
MEME is published by David S. Bennahum. Duplication for non-commerical
use is permitted. Contact me if you have questions. Direct comments, bugs and so on to
me at davidsol@panix.com.In this issue of MEME:
