My trade is computers and during my degree I have studied a thing called HCI, Human Computer Interface.Computers can do things very fast and accurately, humans can't. With humans there is usually a trade off balance between, speed and accuracy, dependant on experience and human performance (tiredness, environment, spirit, energy levels etc) All of these things I am certain aviation authorities are aware of.However, computers need to be commanded in what it is you want them to do. It ends up after analysis that this is the time when a computer is wasted, waiting on a slow human. HCI is the study of the interface point between man and machine.Something that has struct me in my days in a virtual cockpit is that some systems seem to comply badly to the basics of the HCI findings. In short the HCI interface should provide, clear purpose, affordance and feedback. Purpose is usually ok in cockpits, things are clearly labeled. Affordance is also usually ok in cockpits, buttons are usually raised and "look like they should be pushed", levers, dials and sliders provide good indication as to how to use them. However I have found that feedback can be lapse.A perfect example and a previous post reminded me of it, is the APP light going out on approach. This is BAD, it provides completely the oppisite feedback to the user that it is meant to provide. It is saying that it's happy and that it is locked in all Go OK. However the feedback that it provides the brain, without training is that it has failed, or disengaged.OK, so I have mentioned the word I will now counter it. Training. Pilots are trained well, so they know from that training that the APP light going out means all is well.This is probably my bottom line point about cockpit HCI in that it relys very heavily on the training and professional competance of the pilots. In my opinion, given that I am only a undergraduate computer programmer, I believe that this approach is followed to much and a lot of basic HCI (or human machine interface) principles are pushed aside with the gem, "We will train the pilots to understand that" instead of think it through and find the intuative interface for it.The FMC is an example of a perfectly unintuative system, it's interface is terrible. It's complex, fiddily and quite often difficult to work with. I am good with using computers and I understand the way the FMC works, but I still have to refer to the manual to get certain things to work. Thus I believe again it relys on the training of the pilots, rather than provide an intuative interface for it's use. There are somethings where on one page you use it this way but on another you use it a different way, this is bad consistency which adds again to the operators concentration level and work and memory load.If a computer is designed to remove the work load from the pilot, then the command interface should not present almost as much work load as doing the task manually. The FMC comes close. To load a new approach and setting for alternate into the FMC is actually more time and workload consuming than actually tuning the nav radios and flying the bearings and fixes by hand!It's just some thoughts. Nothing more, maybe someone can comment on it.

I think you have to take into account most pilots flying these things have flowen other glass cockpit airliners (I assume) which work in pretty similar ways (the boeings anyway) - so it's a very good interface for them because it's intuitive to them.I could well be wrong with the assumptions I made. I'm going to be doing undergraduate computing for a few years, assuming I can find my exam certificates (GCSE for chirst's sake - what's the point in them?) because my university needs them for matriculation purposes. Can't think why when I got in on A levels and interview, it's probably the government's fault in some way - any way, I'm sure I'll learn about HCI on that and we'll see.

Only one comment - and you provided it:"without training...."Cheers,Paulhttp://www.strontiumdog.plus.com/sbird.jpgwww.BAVirtual.co.uk[p]Officially licenced by British Airways plc for use of name and logo[/p]

Interesting stuff Paul,You may well be right. I make one or two observations;Your comments go to the core of the debate between the Airbus/Boeing strategies for Control Column implementation. The Airbus sidestick/FBW systesm is criticised by many pilots citing the failure of such a system to impart appropriate feedback. Boeing, on the other hand, retained the central column with a conventional style of feedback. Even when under autopilot control, the control changes that are made can be felt (and seen) by the pilot.Thinking outside the box, I would be surprised if the design of the MCP hasn't been given some considerable thought and research. How about this for an angle - in that the APP light extinguishes, it DOES give feedback as to progress BUT it also forces the pilot to change his focus to the primary displays (and out of the window!) to actively MONITOR that the approach is being executed directly (a primary function). If the APP light stayed illuminated, it may be easier for the pilot to "switch off" to that primary function by assuming that the autopilot is correctly flying the approach (and potentially into a hill). The brain at this point would be saying "the light isn't on - I must actively monitor what's happening here"In the same way, many aviation accidents are attributable to overfamiliarity. Could it be argued (oh god what am I saying) that in that the FMC (and any other system) must be operated in a specific, unique and training specific way, it encourages a greater level of concentration and forces the mind to operate in a methodical way.So, for mission critical systems in any sphere, is there advantage in making that system too similar to any other system? - to make the logic too similar to another logic?I don't know BUT I bet there's been some research.We're not talking CLARITY here - there have been numerous examples of the design of indicators/instruments that have proved to be less than satisfactory (Flight BD092 B737 8th Jan 1989 G-OBME ... design of engine vibration indicators for example) - we're talking about systems that appears to be CLEAR but are UNIQUE in the method of their operation.In some way you allude to what I'm saying here when you quote "...adds to the operators concentration level." Isn't that desirable?..... and I'm not sure the bit about programming the FMC for alternates stacks up. Given a similar level of concentration and input (OK you have it there), the approach is far more likely to be flown correctly, safely and accurately by the FMC than by the crew having to concentrate right down to the tarmac on approach plates, nailing needles etc. ? I'll stop now before the pills wear off :)Respecto dude :-luke

>Thinking outside the box,...that's a central tenet of my everyday concious world. It's what I'm paid to do! :-)>How about this for an angle - in that the APP light>extinguishes, it DOES give feedback as to progress BUT it also>forces the pilot to change his focus to the primary displays>(and out of the window!) to actively MONITOR that the approach>is being executed directly (a primary function). If the APP>light stayed illuminated, it may be easier for the pilot to>"switch off" to that primary function by assuming that the>autopilot is correctly flying the approach (and potentially>into a hill). The brain at this point would be saying "the>light isn't on - I must actively monitor what's happening>here"As a one-per-month long-haul passenger, and considering that my wife - and mother to my children - is a BA long-haul stewdie, I would hope to God that the guy driving the bird pays more to the "whole" than what one, single, little light shows.I agree that perhaps the logic of when the APP light is lit/not lit is not exactly intuitive, but...c'mon.. there are surely more cues available to the pilot to indicate a go/no-go situation.Do you really believe the pilots stare at the APP light on approach for guidance?EDIT: "you might be right". Of course I am. I'm no professional pilot, but I have enough respect for commercial aviation to appreciate that training is everything.Cheers,Paulhttp://www.strontiumdog.plus.com/sbird.jpgwww.BAVirtual.co.uk[p]Officially licenced by British Airways plc for use of name and logo[/p]

>>Do you really believe the pilots stare at the APP light on approach for guidance?Of course not Paul ...... and that's exactly my point. Perhaps, though, the subconscious message of having this light illuminated would be "everything is OK". Just a thought.The APP lamp illuminates to show that acquire mode is ARMED. Once capture has occurred the lamp extinguishes. Entire focus should then be on ensuring the approach is proceeding correctly by monitoring the approach - not monitoring whether the autopilot is set to carry out the approach.Gotta go - left something on the bunsen ........ :)

You left out an important attribute that humans have that computers lack is flexability and the ability to take an unusual situation, analyize and adapt faster and better than a computer. One human with proper education and training can fly a Piper Cub one day and next day step into a triple seven and cross the Atlantic. All this with the same hardware they were born with. What is clear to you may not be clear to me. It is another of our human foibles. My aviation experience is as a private pilot instrument rated and the very occasional airline passenger otherwise known as self loading cargo. I am one of those possibly brain damaged types which find NDB approaches to be quite easy and fun. Even flying holding patterns I can "see" in my mind where the plane is in relation to the needle and can navigate without spinning the compass card. Quite a number of my fellow flying club members think of the ADF as a black art. One downright refused to do an NDB approach on an IPC ride and failed because of it. Go figure. The matter of lights on, lights off. It depends on how you design something. For your typical layman, blinking lights and beeps are BAD BAD BAD in an airplane. Ever see a passenger in a light aircraft freak when taking off on an ILS equipted runway and you forgot to turn off the marker beacon? Lights blink the speaker goes beep, beep, beep and their blood pressure and heart rate go through the roof. On the other hand for the pilot during an approach , NOT hearing those sounds on an instrument approach has a similar effect. With the PMDG 737 and the overhead panel, it didn't take very long to get used to lights going off as being normal, but a light comming on as woops. When doing a panel scan, a lot of lights and illuminated numbers slows down your scan. A blinking light, moving needle, tilting attitude indicator catch your attention. I wonder how often line pilots are brought into the loop when designing a cockpit and the layout and function of the instruments. Unfortunatly engineers tend to go with what THEY are familar with and what works rather than what the pilot can use. No offense intended, but with the difficulty of learning your typical new software product be it database or graphics, the folks designing software don't have a clue as to what is user friendly. It seems that we can't afford to scrap the whole lousy mess so we throw band-aids on the problems and shoehorn in fixes. We adapt, overcome and prevail. I don't think you will ever get a universal cockpit and FMC that the newbie can just step into without a manual and use. It would be cool if the pilot could plop in the left seat and say "computer, configure aircraft for passenger load, plot a course to Heathrow, and takeoff at 14:45UTC." Starting Trouble and right engine Bob K.

All good points, good to debate with you all. When I was mentioning "concentration level" what I meant was that the pilots work load can be increased by fiddily systems like the FMC and directed away from the basic approach.NDB approaches... I just read a tutorial on air navigation and NDB approaches appeared new and ugly to me and I was afraid of them, it didn't surprise me when I read the instructor who wrote the tutorial mirror your views and mine that pilots hate them, and as you put it, believe them to be a black art, do it once in the rating and avoid it wherever possible.However, when I started learning NDB inbounds, outbounds intercepts and approaches I was suddenly awed to realise I had done this all before and it was easy.Anyone ever play an old 8bit Specturm or Amstrad flight sim? Hell even the 16bit Atari and Amiga ones and some modern PC ones, where the ONLY navigation you have is "bearing to airport". I was completely shocked that when I was 10 years old, trying to land in a sim with no scenery, and the runway only visible under 3 miles as a rectangle with no other visual cues at all, bar instruments and the bearing, I was able to fly a pattern, based on that bearing, and as you say, "see" with my minds eye where I was. Distance, when available was a great help, but ... that is where I learnt to fly sims.Today I take every opertunity to fly NDBs. Next thing from there is to learn dual DME positioning, but, noone is that mad, are they?

Guys,This is a highly interesting topic and though I don't have a lot of knowledge about it I thought I'd provide you with a link to a thesis made by a friend of mine at the Royal Institute of Technology (KTH) in Stockholm, Sweden. She has done research on the warning sounds on the flight deck and came to some pretty interesting conclusions. Here's a link to an abstract of the thesis. On that page you can also download a 98 page summary of the thesis.http://media.lib.kth.se/dissengrefhit.asp?dissnr=3679It's pretty heavy stuff which only shows that this subject is more complex than many of us believe.One interesting conlusion in the paper is the low fuel warning sound she investigated. She found that using a "slurping" sound like when you are emptying a glass with a straw is way more effective than the sounds used today because that particular sound triggers subconsious perceptions of something getting emptied by the listener. Happy reading, :-)

Nice one Mats - something to read in the bath. I'm sadder than I thought !:-)