Geneva is a city long accustomed to international negotiations. One of the grander parks on its chilly lake houses the Palais des Nations, a pillared concrete monolith built to house the League of Nations after World War I. The League, faltering predecessor to the United Nations, didn't fulfill the ambitions of its founders, as World War II proved. But the Palais remains—an imposing reminder of how difficult it is to resolve conflicts between nations, but how important it is to try.

The agreement in Geneva on Thursday between Iran and a coalition of nations (the five veto-bearing Security Council members, plus Germany) affirm those lessons—and with them another principle of international affairs: nothing can be resolved without talking. First reports from Geneva must be read skeptically. The U.S. and European diplomats who briefed the assembled Western media had a vested interest in positive spin. (Briefings by the others there—Iranian, Russian, and Chinese diplomats—will supply correctives in the days ahead.) The truly important fact about this daylong session at the elegant Villa de Saugny is that Tehran and Washington were meeting openly at the highest level in 30 years. And the 45-minute lunchtime one-on-one between the Iranian emissary, Saeed Jalili, and William Burns—the most-senior diplomat in the State Department—does appear to have produced at least a way out of a confrontation that has been on course toward war.

Iran has agreed to send perhaps as much as three quarters of its known stock of low-enriched uranium to Russia for further processing. It says it will allow inspections of its clandestine nuclear facilities within two weeks by the International Atomic Energy Agency, the U.N. body with global authority to monitor nuclear activities. In return, the U.S. and its allies have reportedly agreed to put  sanctions against Iran on hold—and to postpone the year-end  deadline for Iran to agree to a more comprehensive settlement.

But champagne toasts are premature. What the Villa de Saugny session has done, it is hoped, is buy time. Without highly enriched uranium, Iran cannot fabricate a bomb. The deal, if that's what it proves to be, is in line with another important lesson taught by generations of negotiations in Geneva. First, stave off the immediate crisis. Then tackle the deeper issues. On that front, there is still a lot to tackle.

Even as an interim deal—a holding pattern—the accord leaves important questions unanswered: what about Bushehr? The world has known about Iran's Bushehr-1 nuclear reactor for years, but it doesn't appear to have been part of this deal at all. Why? Because it's not an immediate concern. Even when finally operating at full speed, Bushehr couldn't produce bomb-useful quantities of plutonium for a year or more. Western intelligence doesn't even think Iran possesses the reprocessing facilities needed to extract plutonium from Bushehr's fuel rods. And there are credible reports that Russia, which built Bushehr, already has a private agreement with Iran that puts Russia in charge of those fuel rods.

A more justifiable question is whether Iran will allow the IAEA to inspect all its nuclear facilities, not just one. At Geneva, Jalili apparently agreed to allow inspection of the newly revealed facility in the mountains outside Qum. (The CIA has come in for much battering of late. It deserves plaudits for its detection and monitoring of the Qum site.)

But how many undisclosed facilities does Iran have? In its letter last week to the IAEA, fessing up to Qum once it learned the CIA had detected it, Iran talked of undisclosed sites—plural. Western intelligence agencies reckon that Iran has "more than a dozen" construction sites whose purpose is "suspect," according to the 2007 National Intelligence Estimate sent by the CIA to President Bush. Daily satellite runs over Iran have enabled the CIA to track work at those sites. But satellite images can't prove their purpose. Will Iran really agree to give the IAEA nationwide access?

Even worse, Tehran could balk at what little its envoy has already agreed to. During the 1990s, Iraq constantly blessed inspections, than stymied inspectors; if Tehran follows suit, it could land the country on the same path to war Iraq induced.

Hawks in Washington will dismiss even negotiations with Iran as evidence of craven surrender by the Obama administration. Hawks in the leadership in Tehran will condemn any compromise by Iran. But as Churchill, belying his reputation as an incomparable war leader, observed: "Jaw-jaw is better than war-war."

Realistically, the best outcome will be that Iran settles for a deal renouncing possession of the bomb while retaining the capability to swiftly construct one. In return, the U.S. will almost certainly have to agree to a nonaggression pact, reassuring the regime in Tehran that Washington has abandoned notions of its overthrow. That would require politically controversial decisions on both sides—tougher, probably, in Washington than in Tehran.

That's the best hope. To fulfill it, another generation of diplomats will have to relearn the lessons of their predecessors: the Genevoises are notoriously standoffish toward visiting delegations, and the decent restaurants there are priced way beyond their per diems.