Why didn't BP try the seemingly successful custom-built cap method at the outset of the disaster?
The reason is a blend of factors, from underestimating the severity of the situation to a fear of making a bad situation worse. The 75-tonne cap has never been tried in such an extreme situation, almost a kilometre below the surface.
After the April 20 explosion, BP consistently and vastly underestimated the size of the spill - and its actions reflected that. The well's failed blowout preventer couldn't be made to work and it wasn't until May 7 that the first attempt to stop the oil was made, with a four-storey containment dome over the head of a length of broken pipe connected to the blowout preventer.
The dome was supposed to connect to a surface vessel, but its opening became jammed when high pressure and cold water crystallized the natural gas mixed in with the oil.
Soon after, the first small success came when a tube was directly inserted into the broken pipe, allowing a paltry average of 2,400 barrels a day to be brought to a collection ship at the surface.
On May 2 and May 16, relief wells were started, which in August will ultimately stop the oil.
In late May, moves that have become part of the lexicon of the blowout - the "top kill" and the "junk shot" - were failed attempts to attack the well itself, to plug it.
Under ever more pressure, BP severed the broken pipe that bent out of and away from the blowout preventer. From the start the fear was that this move could unleash even more oil into the Gulf with perhaps no way to stop it - but given the options by late May there was little choice.
"They were afraid to take the thing off," James Williams, a veteran oil industry consultant in Arkansas, said.
A cap, smaller than the current cap, was installed, which stemmed the tide and the cap was linked to the collection ship at the surface.
The current 75-tonne cap replaces the previous caps and can connect to more containment ships, if that is the strategy chosen.