Distributed referral, semantic markup, and transport / exchange /
dissemination of bits of information. It stands here and now in 1997 as
the world's only truly working, global distributed object and document
information exchange system.
I do admit that it is the worst generic system in the world for doing
global distributed object and document applications, except for all the
others. Unless you are writing a custom application where you pick your
own protocols, un/pickling strategies, and naming conventions, nothing
else has even close to the performance properties of the Web, when
scaled to billions of communicating objects. Worse, nothing else has
even close to the installed infrastructure in place as the Web has
presently in Web servers and client browsers. History has taught us
that people are slow to migrate to a completely new platform; they will
simply latch onto whatever emerges from existing, already-installed
platforms. This is why Beans are coming to the Web, not the other way
around.
Furthermore, I think that anything new and improved that will emerge
from the dissatisfaction that people have with the Web, is not going to
turn its back on the culture and principles that made the Web itself a
resounding success in a precious few years.
The Web is end-all only in the sense that the philosophy of the Web, for
better or worse, was realistic enough to work on such a global scale.
These principles include the decentralizing the trust management and
information storage, using tags to say precisely what you mean, linking
naming to ownership, and allowing the system to be robust enough that
brittle local failures don't bring the global whole system down.
The Web is now an integral part of our everyday lives, and that is no
small feat. We'd be fools to think that the reason the Web succeeded is
completely random. It has some important features that have brought
great benefits to the domain of distributed objects and documents.
Could it be better? Certainly. As part of the next step in the
evolution of the Web, we'll find more things automatically extracted
from distributed sources, markup for ratings and content selection,
more security and reliability including watermarking and digital
signature standards, and flexibility through extensible protocols,
markups, and names.
But ask me if the Web is good as it presently is, and I'll unrepentantly
say, yes. I remember life before the Web. I much prefer the chance of
finding something in a search engine or referral to slogging through a
library and/or making an array of telephone, email, and fax calls and/or
wading through loads of television, usenet, and radio broadcasts.
----
adam@cs.caltech.edu
When you have to kill a man, it costs nothing to be polite.
-- Winston Churchill