ORLANDO, Fla. (AP) _ NASA went ahead with the Challenger's launch despite an engineer's warning that it was playing Russian roulette because of the cold and icicles on the shuttle, according to newly released NASA transcripts.

The Orlando Sentinel, which obtained the transcripts under the Freedom of Information Act, also reported in today's editions that space agency officials rejected the recommendation of the ice team chief, who wanted the twice- postponed launch delayed again because of cold and ice.

The transcripts cover conversations on 10 channels used by NASA, Lockheed Space Operations Co. and Rockwell International from 7 a.m. Jan. 27 to noon Jan. 28, 21 minutes after the shuttle exploded.

Transcripts indicate that sheets of icicles, some of them 2 feet long and and an inch thick, hung from the launch pad of the shuttle, whose seven astronauts died when it broke up 74 seconds after liftoff.

''Ice does look bad, yeah,'' said John Tribe, Rockwell's engineering director at Kennedy Space Center. ''... Some of the of the close-ups of the stairwells looks like something out of Dr. Zhivago. There's sheets of icicles hanging everywhere.''

Ice also was spotted on the attachment rings between the shuttle's external fuel tank and its boosters, the newspaper said.

John Peller, vice president of engineering for Rockwell, said the ice was too unpredictable to approve a launch.

''We still are of the position that it's a bit of Russian roulette; you'll probably make it,'' he said. ''Five out of six times you do playing Russian roulette.''

In the transcripts, engineers for Rockwell, the shuttle's manufacturer, appear frightened that ice on the launch tower would fall during liftoff and seriously damage Challenger.

''The big concern is nobody knows what ... is going to happen when that thing lights off and all that ice gets shook loose and comes tumbling down, and what does it do then?'' Tribe said in a conversation to Rockwell headquarters in Downey. ''Does it ricochet, does it get into some turbulent condition that throws it against the vehicle?''

Ice team chief Charlie Stevenson, who said he later approved the launch, at first suggested it be postponed, the transcripts show.

''I'd say the only choice you got today is not to go,'' Stevenson says in the transcripts.

On Jan. 28, it was 38 degrees, colder than for any of the previous 24 shuttle launches.

The transcripts show other cold-weather problems. Several video cameras used to monitor activity on the pad stopped working. One radio channel went out. Other failed equipment included one of three sensors on the shuttle that check oxygen levels in the crew cabin, automated gas valve controls on the pad, a part in a guidance system and a device to detect hazardous gas at the pad.

All but the oxygen sensor were repaired by launch time.

William Rogers, who chairs the presidential commission probing the tragedy, has said NASA's decision to launch was flawed. The commission has determined that design flaws, freezing temperatures and other factors caused joints on the shuttle's right booster to fail, triggering the deadly explosion.