There's reasonable evidence [0] that the director of NASA and ground crews knew the Columbia's wing damage was bad enough to make re-entry impossible. But they had two options, tell the astronauts they were never coming home at the start of their mission, or wait until they discovered it minutes before their vehicle disintegrated.
The director chose the former path, and ended up getting fired as it looked like a cover up. It was really just being compassionate.
No pad abort system would have saved Columbia. But on the flip-side, while everyone is up and arms over the loss, NASA's astronauts are trained to WANT to fly. We can hem and haw and say what a tragedy, but there is a class of people who know the risks, who know the possibility of death always exists and will still pilot or crew the vehicles that we send into space.
As far as I’m aware, NASA knew that there had been a foam / ice strike on the orbiter but had no idea how bad it was. Given those circumstances, why tell the astronauts about a theoretical risk (that turned out to be very real) ?
The reason everyone is up in arms, still to this day, about the Columbia "accident" is because we could have sent a second shuttle to retrieve them very easily.
NASA engineers were fighting the decision for the Columbia to return because they already were assembling a plan to get them home alive.
Seven astronauts died because Bush put Sean O'Keefe in as Administrator of NASA, and Sean O'Keefe decided it'd cost NASA less money to let the astronauts die instead of bringing them back alive.
Sean O'Keefe personally tried to halt any work on sending Atlantis up as a rescue mission, and ended up succeeding, all in the name of budgetary concerns. He didn't give a fuck about those astronauts.
And to him? It was just a revolving door. He's now working some cushy job at Airbus as their US CEO, a military contractor he made friends with when he worked for the DoD and was Secretary of the Navy (under Bush's father, the other President Bush).
So, fuck him. This wasn't an accident, this was murder for profit.
Except for the profit part? It's not like he got the Airbus job because he killed some astronauts.
Every report I've read this, the math doesn't check out here. It takes months at a minimum to get a shuttle ready for launch and there's no proven method for accelerating that time frame or keeping people alive in the shuttle for the extended periods of time required to prep another launch. It's not like they had a second one waiting on the pad ready to go.
This isn't even taking into consideration the required engineering and logistics around moving people between shuttles, which there's no procedure for.
These things aren't minor details.
Source: I've played kerbal space program. :-)
Wikipedia quote time: "The CAIB [Columbia Accident Investigation Board] determined that a rescue mission, though risky, might have been possible provided NASA management had taken action soon enough.[56][57] They stated that, had NASA management acted in time, two possible contingency procedures were available: a rescue mission by shuttle Atlantis, and an emergency spacewalk to attempt repairs to the left wing thermal protection.
Normally, a rescue mission is not possible, due to the time required to prepare a shuttle for launch, and the limited consumables (power, water, air) of an orbiting shuttle. However, Atlantis was well along in processing for a planned March 1 launch on STS-114, and Columbia carried an unusually large quantity of consumables due to an Extended Duration Orbiter package. The CAIB determined that this would have allowed Columbia to stay in orbit until flight day 30 (February 15). NASA investigators determined that Atlantis processing could have been expedited with no skipped safety checks for a February 10 launch."
56: Columbia Accident Investigation Board (2003). "Decision Making at NASA" (PDF). CAIB Report, Volume I, chapter 6.4 "Possibility of Rescue or Repair". p. 173ff. Retrieved March 28, 2013. http://anon.nasa-global.speedera.net/anon.nasa-global/CAIB/C...
57: Columbia Accident Investigation Board (2003). "STS-107 In-Flight Options Assessment" (PDF). CAIB Report, Volume II, appendix D.13. p. 391ff. Retrieved January 30, 2006. http://www.nasa.gov/columbia/caib/PDFS/VOL2/D13.PDF
It would have made for great drama. But either having Atlantis get up there too late to save the Columbia crew, or, even worse, losing Atlantis as well, would have been a PR disaster from which NASA might well not have recovered. Hell, they barely recovered from Columbia as it is.
[1] http://arstechnica.com/science/2016/02/the-audacious-rescue-...
So Atlantis wasn't actually as close to launching as your comments make it sound. Later in the article, they state that it would be a minimum of 21 days to get Atlantis on the pad and ready to launch. And it always took longer than NASA thought; the article author looked for any instances in which a Shuttle moved from OPF to VAB to the pad and launch without errors or faults, and couldn't find a single one.
That's why they planned for this contingency in the remaining Shuttle missions, with the so-called "Launch On Need" planned missions. (In all but one case, they assumed that the Shuttle crew could take refuge at the ISS until a rescue mission could be mounted. For the one mission where they couldn't assume that, STS-125, the Hubble servicing mission, they actually had a second Shuttle on the pad before the first one launched.)
I don't fully understand why it would not have been possible to put a Soyuz in a different orbit but with a near-collision course though.
Regardless, the Soyuz can only hold 110 lbs of extra weight beyond three crew members. So they'd need almost as many skeleton-crew Soyuz modules as there were crew aboard Columbia.
And even with all of that, was the question of, could Atlantis launch without the same insulation damage issue arising? Given that they hadn't prepared any mitigation strategies yet, the answer was "no" and under those conditions, the Shuttle is grounded.
How awful would it have been to lose Columbia AND Atlantis?
1: http://arstechnica.com/science/2016/02/the-audacious-rescue-...
The acceleration will probably knock out the astronauts and cause internal bleeding, but that's still better than immolation.
Failed re-entry is still going to be fatal, but the Dragon can theoretically recover from a failed parachute with its Draco engines.