First of all, while a good post with good points, the heading "Pope" guy is wrong
afaik. 737, 747, 727, BAe146, Bizjets, (Collins, Sperry, King etc) and dozens of GA aircraft with HSI's, took the shortest distance to the preset heading bug and turned
towards it when heading select was engaged. Airbus is the only one I ever flew that would do a 270 degree turn the wrong direction since it remembered which way you initially turned the knob on the last approach. Crazy.
My memory, back when we did circles, was that it was up to the PIC to remain within the terps or pan-ops obstacle protection "cloud"
and maintain visual with the threshold. Since say, the vis is reported in Statue miles, and there's no Statue Mile Display in the aircraft, the obvious question becomes, how do you measure it?
Even if you had an ILS DME readout on the landing runway 12 that you cranked in just to get the raw number to the threshold, just to make the calc easy, that readout is in Nautical Miles. So:
2sm = ~1.7nm . Right?
So by 411A's description, you couldn't see 1.8 clicks on the meter on the downwind to base or it was a bust; is that right?
Now part 91 and supplemental, as another poster mentioned, maybe for a time had a loophole out of this tight circuit, whereby the crew could go to 2.2 nm (terps) or 5.x nm (just shy of pan-ops protection) if the pilot's vis was better than reported, however, the flight crew took responsibility for visual separation with terrain, IIRC.
So here's my bitch: Why are op specs and some visability still reported in statue miles? Are we in a car here?
Why the hell are we still using a primitive 1970's Green Screen FMS with awkward command-line interface instead of GUI (graphical user interface) which every kid with a mobile phone has?
Just so Honeywell can have a monopoly and keep employing FMS programers whome no two program the Byzantine crap software the same at any two airports in the world?
Talk about no standardization! There is no standardization between FMS waypoint programers. None!
For a while, the crap would work O.K. and you'd start getting used to it's idiosyncrasies. Then you'd fly up to NRT and on the SID, even though the primary flt plan showed it going to the VOR first,
and then turning onto course, the
Airbus would cut the freaking corner early sending ATC into orbit.
No one could fix it for years, so you just had to be ready to:
disconnect: reconnect
disconnect: reconnect
disconnect: reconnect
disconnect: reconnect
disconnect: reconnect, all night long.
Turb on apprch? Thrust Latch, here comes another acrobatic tail slide!
ding ding ding ding ding ding......
Who the hell designed this freaking mad computer in place of an airplane?
The only way not to get killed was to hand fly it all the time. But then the Glass-Generation F/O's would cry: that damn "hand job" Capt is overloading me!!!!!!! I can't talk on the mike and turn the heading knob at the same time!!!!
If you ask me, the arguments I've read on this thread, that the Captain was an experienced 747 Capt mean absolutely nothing. On the airbus you are no longer really a pilot (if you always fly at the highest level of automation as they wanted you to); you are now an FSS: Flight System Supervisor.
Why do Circle to Lands a poster asked? So you don't run off the runway with a twenty knot tailwind, that's why. If you'll turn off and just quit phucking with all that electric jet chit, and all that performance chit, and just use raw data, and have two guy's heads looking out the window, this sort of silly accident might not happen. (That is, if it is indeed, a CFIT accident, as it appears to be.)
That's what I think anyway.
CC