SOAP, the Simple Object Access Protocol, is an evolving W3C
standard developed by representatives of IBM, Microsoft,
DevelopMentor, and UserLand Software for the exchange of
information over a network. As more SOAP servers become
publicly available on the Web, SOAP is doing for programs
written in nearly any language -- even short little programs
written in popular, simple languages like Visual Basic,
JavaScript, and perl -- what HTML does for Web browsers: It
gives them a simple way to take advantage of an increasing
number of information sources becoming available over the Web.

Like HTML, SOAP provides a set of tags to indicate the roles
of different pieces of information being sent over the Web
using the HTTP transport protocol (and since SOAP 1.1, SMTP
as well). SOAP, however, gives you much more power than
HTML. With SOAP, your program sends a "SOAP request" (a
short XML document that describes a method to invoke on a
remote machine and any parameters to pass to it) to a SOAP
server. The SOAP server will try to execute that method with
those parameters and send a SOAP response back to your
program. The response is either the result of the execution
or the appropriate error message. Public SOAP servers are
available to provide stock prices, the latest currency
conversion rates, FedEx package tracking information,
solutions to algebraic expressions, and all kinds of
information to any SOAP client that asks.

Before SOAP existed, programs trying to use this kind of
information had to pull down Web pages and "scrape" the HTML
to look for the appropriate text. A visual redesign of those
Web pages (for example, putting the current stock price in a
table's third column instead of its second column) was all
it took to render these programs useless. The SOAP spec,
along with its brief accompanying schemas for SOAP requests
and responses, provides the framework for a contract between
clients and servers that creates a foundation for much more
robust information-gathering tools.