Rainzi said, "Of course not. Livia is proposing a stagedapproach. Before attempting to construct your graph, we’d move towardit in a series of experiments, gradually bridging the gap."

Cass fell silent. Compared to outright rejection this wasa trivial obstacle, but it still stung: she’d worked for thirty yearsto refine her own proposal, and she resented the implication that she’dbeen reckless.

"How many stages?"

"Fifteen," Livia replied. She swept a hand through thevacuum in front of her, and a sequence of target graphs appeared. Cassstudied them, taking her time.

They’d been well chosen. At first one by one, then inpairs, then triples, the features that conspired to render her owntarget stable were introduced. If there was someundiscovered flaw in the rules that would make the final graphdangerous, there could be no more systematic way to detect it inadvance.

"It’s your choice," Rainzi said. "We’ll vote on whicheverproposal you endorse."

Cass met his eyes. The openness of his face was an act ofpuppetry, but that didn’t mean he was insincere. This wasn’t a threat,an attempt to bully her into agreeing. It was a mark of respect thatthey were letting her decide, letting her weigh up her own costs, herown fears, before they voted.

She said, "Fifteen experiments. How long would that take?"

Ilene answered, "Perhaps three years. Perhaps five."Conditions varied, and the Quietener wasn’t perfect. Planning anexperiment in QGT was like waiting for a stretch of ocean to growsufficiently calm that a few flimsy barriers could block the waves andkeep out the wildlife long enough to let you test some subtle principleof fluid dynamics. There was no equivalent of a laboratory water tank;space-time was all ocean, indivisible.