Re: heterogeneous DBssowa <email@example.com>
Date: Tue, 14 Sep 93 00:18:56 EDT
From: sowa <firstname.lastname@example.org>
To: email@example.com, firstname.lastname@example.org, interlingua@ISI.EDU,
Subject: Re: heterogeneous DBs
> How curious. I send a message saying roughly:
> Has KIF been used in any applications? If not, why is it being proposed
> as a standard?
> and the response is:
> AI is stupid and should go away.
The answer was not that AI is stupid nor that AI should go away.
Instead, the observation was that AI is dying. Not that it should
die, but simply that it is dying.
What is killing AI is a lack of solid, industrial strength software
that real businesses can use to build robust, reliable applications.
Multibillion-dollar companies do not want to trust their businesses
to software coming from tiny little software companies that might go
out of business at any time. Nor do they want to trust their business
to software coming from other billion-dollar companies that might change
their product line at any moment. Instead, they want software based on
open standards, where they can pick and choose between compatible
products from multiple vendors.
Unix is not the best operating system in the world. C is one of the
lousiest programming languages in the world. SQL is a decidedly
mediocre database query language. And C++ is the world's worst
object-oriented programming language. Yet that bunch of lousy
products has become a world standard for one reason only -- you
can get compatible versions from every vendor.
If AI is going to survive -- and I certainly hope it will --
then people in the AI field will have to get some solid, reliable
open systems out that users can mix and match with other open
Bottom line: Without standards, no customers. Without customers,
no products. Without products, no applications. Without applications,
no funding. Without funding, no research.