The Strategic Defense Initiative that became the focus of an exchange over "Star Wars" during Sunday's presidential debate differs in important aspects from the way in which it was described by President Reagan and Walter F. Mondale.

Reagan represented the proposal as a "humanitarian" plan that would provide an alternative to mutual destruction in a nuclear war and that holds out the prospect of making nuclear weapons obsolete. Mondale called the idea dangerous and costly "madness" that would escalate the arms race and turn decisions over to computers.

The record shows that the initiative, at least as presently conceived, falls far short of the campaign claims of both candidates.

In its present form, the initiative is the product of a nationally televised speech that Reagan gave on March 23, 1983, in which he called for "a comprehensive and intensive effort to define a long-term research and development program" aimed at developing the technology for a system that could "intercept and destroy ballistic missiles before they reached our own soil."

The speech instantly provoked controversy. Proponents said that a defensive system would provide a humanitarian alternative to a dooms-day system of deterrence based on the premise that the Soviets or the United States would retaliate in kind if either started a nuclear war.

Opponents, echoing the debate that surrounded the ratification of the 1972 U.S.-Soviet treaty that banned anti-ballistic missiles, said that a defensive system was more likely to cause a nuclear war than prevent it. They said a potential aggressor with a successful anti-missile defense might be tempted to begin a nuclear exchange, believing it could survive retaliation.

But despite the rhetoric on both sides, the Reagan administration proposal fell far short of what the president had promised -- and what his opponents feared.

Several months later, the administration collected research programs that had been going on for 12 years and consolidated them in a new Pentagon agency called the Strategic Defense Initiative Office. Administration officials projected spending $25 billion on research from fiscal 1985 through 1989, without emphasizing that $16 billion to $18 billion would have been spent for the same purposes if the president had merely continued existing programs.

At present, the administration is spending $1.3 billion on the initiative, some $250 million less than had been requested. But no new programs have been developed, much to the consternation of industry and government officials involved with the research.

"There have been no new initiatives so far," an industry official familiar with the program said recently. "It's only a collection of disgruntled technologies."

The project is charged with finding out whether a defensive system is technologically feasible and recommending options that would be a basis for decisions by Congress and the next president. Development or deployment of an anti-missile system would violate the 1972 ABM treaty, as Reagan acknowledged in his original speech.

In their debate, Mondale zeroed in on a Reagan statement that he would be willing to share the anti-missile technology with the Soviets, if it is developed, a proposal that the Democratic nominee said involved sharing sophisticated computer information and was "a total non-starter."

Michael Havey, senior policy analyst for presidential science adviser George Keyworth, said that, in any event, the Soviets would obtain such technology within five years of the United States getting it. Reagan repeated several times that he would be willing to share the information but when pressed said he would do so only after a "demonstration" of it and agreement by the Soviets to discuss arms control.

Mondale said "the most dangerous aspect of this proposal is for the first time we would delegate to computers the decision as to whether to start a war." However, those familiar with the concept of ABM defense said there is no inherent reason to rely more on computers with defensive weapons systems than with offensive ones.

In any event, Havey said, the consequence of either a computer or a human error would be far less with a defensive system since the weapons would be aimed at airborne missiles, not targets in other nations.

"You haven't just accepted research, Mr. President," Mondale said. "You've set up a strategic defense initiative, an agency, you're beginning to test, you're talking about deployment, you're asking for a budget of some $30 billion for this purpose. This is an arms escalation."

In talking about testing, Mondale may have been referring to a totally different program, involving tests of anti-satellite weapons (ASAT). This is a program designed to find ways to destroy military satellites rather than incoming missiles. While the technologies are related, there have been no tests of laser or space-based anti-missile systems, which in general pose far more complicated problems than do ASATs.

Neither Reagan nor Mondale seemed to know much about the wide variety of proposed anti-missile technologies, about which Mondale said, "if they are space weapons, I assume they'll be in space." In fact, ground-based lasers are considered one of the most promising possibilities.

Reagan, who had said earlier that he was "not a scientist," offered no rebuttal.