the right answer is: it is impossible to extrapolate the direction of a turbulent flow....

but why they do not started this experiment one year earlier ?

You are misinformed here : AP but also A/THR could be reengaged under alternate law.
But a RECENT OEB warns pilots not to do so, at least not instinctively ...

Exactly. But they know and have already expressed their regrets. Second interim report, appendix 2, 1.5: "The dropping of drift-measurement buoys by the first aircraft to arrive over the zone would have made it possible to understand the drift better from the earliest hours".

Newspapers in France.
 

Would any of the Francophones like to comment on the relative levels of respect of various French newspapers?

I found two references List of newspapers in France - Wikipedia, the free encyclopedia and French dailies, the daily press in France but as a Brit I am aware that most (all?) papers have agendas determined by editorial policy and politics.

Personally I almost never buy papers, but prefer to read news online or listen to radio or TV news.

I am sure some sources are more sensational than others, and that ownership and major shareholders may also affect any un-biased reporting.

I use this thread as my primary reference followed by the official BEA website.

Hope this is not too far off-topic, this thread has drifted a long way from the original search for wreckage (now accomplished) and has moved on to various other parts of the investigation.

JPI, très certainement! Merci
____________

grity, as I recall they were trying to duplicate conditions at that time of year. During 2009, they were mostly relying on drifters, fishermen's buoys.

As I recall, the drift group did not use the 2010 buoy movement in preparing its June 30 2010 report. But Metron did in its January 2011 analysis.

kit344
Well, among the biggest newspapers, I would recommend "Le Monde" as being the most serious one, and "Libération" is not that bad too. OTOH, I have been very disappointed/upset with "Le Figaro" on several occasions :*, and I prefer not to comment about "France Soir" :mad:.

This is only my personal opinion and YMMV obviously, so take it with a pinch of salt.

They were also the actions of a glider pilot, which might have given him an extra way of thinking about the situation he was in. There was lucky as well - the incident occured in daylight, and the Hudson is a very busy stretch of river.

This recent AD is for wording it more clearly: it is possible to (re)engage AP/FD if 2 ADRs are (re)displaying consistent, but erroneous, data (hence danger). I wrote it many times that it would be possible in such circumstances. But in AF 447 case, it seems very unlikely that they ever had such an opportunity due to the RTL settings still frozen at impact; a PRIM reset is also mandatory before doing it, but it was not made until very late in the sequence. In any case, it would be due to pilot's wrong choice of erroneous data as the computer will not activate AP/FD by itself.

Re 'sûreté' and 'sécurité'....
The explanations in the earlier posts are fine, but there's one more point....

In French secondary schools, you are taught to avoid as much as possible using the same word twice in the same paragraph.
That's OK in literature, but it can be confusing, when authors of technical documentation try to apply the same rule at all cost.

It took me a few weeks (after moving to my first job in France) to understand that "tangage" and "profondeur" can be used interchangeably for "pitch" in an aeronautical context, until a French colleague explained this slightly odd habit.

So, while sûreté and sécurité are not true synonyms, don't be amazed if you discover them being used as such in the same paragraph.....

I see what you mean now.
I have just checked the French version and the English translation seems accurate. I believe there is a possibility the BEA text is inaccurate and should say :
''If none of the three Mach values is valid, a Mach value close to 0.8 is used. For example, it is of the order of 4° at Mach 0.8''
Such wording would better match the Air Caraibes occurrence.

It would have greatly complicated the all issue and probably bring more confusion to an already very confusing situation.
AT THE TIME the procedure was to simultaneously pitch down and apply full thrust. In other words it would have unnecessarily destabilized a situation that was under control.

Those Air Caraibes crews did so well under such confusing environment.

Shame on the BEA and others ... to not have treated adequately this valuable information.

@ kit344 : this webpage (i.e. your second reference) offers an honest & almost comprehensive look. :D

Why "almost" ?
1/ 20 minutes is missing as the 2nd free daily
2/ I don't read the economics press so frequently, but Les Echos deserves to be called a quality daily IMO.

As JPI33600, I've been disappointed/upset with "Le Figaro" on several occasions. But the same is true with "Le Monde", too (though perhaps a bit less often, or on less serious issues).

Hi,

@ kit344 and JPI33600, about French newspapers.About same advice as JPI, but have to said Le Monde seems less "above the others" than some years ago, and Le Figaro has often very interesting wording (verified by myself about domains where I have some knowledge), but is not politcally neutral.
And don't miss "Les Échos", a newspapers about economics, but with very trustable other news: Actualité économique et financière - Information économique et financière - Journal quotidien économique et financier - Les Echos.fr
If you are speaking about French language, Belgian papers have to be read: they are often more open minded than French ones. Try actu - lesoir.be as a "medium" example.

Just a word about "sécurité/sûreté": the meaning associated with "sûreté" is sure ("certain", same word in French). What is "sûr" has to happen, it is predictable, and by the way, safe because not hazardous.

Don't shoot me for asking (B1, B2 Engineer, non Airbus)

Is it at all possible for there to be a huge change in wind direction over a very short distance within this type of cloud, thinking head to tail wind, and if the probes are iced could this go unnoticed? and result in a real stall?

Also how do probes react to icing up, I know if the calibrated bleed hole blocks it will over read but if the opening becomes iced would the pressure inside remain constant due to sudden blockage or would the pressure decay as the opening gets smaller due to ice build up?

Just how large would the ice particles be? could they enter the pitot, build up and block within the hoses after the heated head?

Sorry if this has been discussed previously, taken me 5 hours to catch up with all the latest posts!

Thanks.

In theory (and I presume practice) if you rotate these essentially solid state no moving parts instruments fast enough or accelerate the rotation fast enough you may overrun some counter that is counting interference fringes. Then it becomes inaccurate. I would presume, probably safely, that this is outside the range of motions of even a military jet. The upset is quite different from tumbling a gyro with its reaction to forces perpendicular to its axis due to friction.
I would presume the most likely reason for lost calibration can be as simple as an electric power glitch that went beyond what its internal power supply can handle. As I mentioned above I don't think you can exceed its ability to count beat notes that come from the slightly different effective distances the light traveled clockwise and counter-clockwise (deiseal and widdershins if you will) around the "ring" of laser optics. Bang it hard enough and you might actually bend the optical bench destroying its accuracy and perhaps its ability to work. I'd expect that to require quite a bang, as with a serious attempt to break it with hammer or something like that.

Time will also cut its accuracy down. You measure rotation rate in steps. So if you remain rotating at not quite one full step from not rotating over time your heading will be different from what is reported by a significant margin. It works on the speed of light. So that would be a very small error that would require a very long time to add up to an accuracy loss.

Recovery after it's lost power or whatever requires a very accurate knowledge of the plane's exact geographic position (GPS), attitude, and heading relative to true North. That's easy on the ground, more or less. In the air it could be more than a slight bit difficult, especially when we get some vibration.

edit: auv-ee's description is excellent and in greater detail. Do consider it.

Given that ACARS didn't indicate a power supply loss, that scenario does not seem to fit this case, but ... to better understand recovery or lost capability ... if you had a power loss / or interruption (part of why there are 3 (do I have that right) in a given aircraft ???) ... then, not only do you need to reset location and direction, you have to establish wings level, nose pitch X, (zero, I assume) to re-establish correct attitude sensing from which the gyro can reference.

That is similar to "recaging" a gyroscopic Attitude Indicator if it has "tumbled" in flight, though from your description, a bit more complex.

Newspapers in France
 

kit344;

I would also recommend the website of the weekly Le Nouvel Observateur for factual, informative and timely news whenever there are new developments re: AF447.

Mr. deSitter, please understand IT professionals a little better.

If you are sitting in an airplane, push the button to start engine 1 and nothing happens do you sit and think before you push the next buttons? "Why did it do that? What will go wrong if I do X or Y or Z?" There is probably a procedure to follow. But IF (added) nothing will break if you sit there looking blank while you ponder the situation, check a few dials, and then proceed.

IT professionals are much like that. As a developer I am not strictly "IT person". But I often have what I just wrote do something completely wonkity. I'll sit and review what I did, what might have caused the error, and maybe even try to recreate it one or more times. (If it recreates easy it is easier to find.) But the first thing I do is sit there and exercise by brain instead of my brawn. It helps, before the next action, to have a half a notion what went wrong before making a total mess of things. (One of the REALLY critical questions is, "Was this the disk drive getting flakey?" If it is you want to get a backup before you mess with anything else. I am sure with the engine start scenario there are "co-symptoms" you look for and if they appear you try not to jump out the cockpit window and instead do your best to save the passengers then the plane then yourself from the engine that just caught fire. The disk drive is sort of the IT equivalent.

Think FIRST, then react. Otherwise you might turn something bad into a total disaster. (And for some operator errors on the keyboard all you can do is hit control-c, the command line stop the stupid program command, and pray. With the UNIX derived OSs if you accidentally misuse "rm -rf" you can clean off the entire hard disk if you don't stop it quickly. So the command issued is checked right away. Then you stare blankly looking at things and thinking. And of course that, too, involves thinking on what you did before you stop it. Sometimes stopping a command can do more damage than letting it complete.

Some people paraphrase this as "Don't just do something; sit there."

Edit - added back a word I'd been thinking and didn't type.

deSitter, there is a second issue that has me bemused. It's about the GIMME button. If it is supposed to give you an airplane to fly, and it is basically fly by wire, would you care to define what kind of aircraft it should mimic? It could drop you into a direct law that gave the controls the touchy feel of a P-38 configured for stunts or it could mimic a fully loaded C-130 or something in-between.

I'm thinking of the poor sods who get your demand and have to program it in. I tend to whine like that when asked to read minds by pointy haired bosses.

No, it is not.
The AP not but the FD(s) absolutely.

As well as 3 they should also have pri & sec power supplies, sec normally is an emergency battery. I also remember (training course a few years back) they have a degraded mode of operation, ATT mode, no position input needed and much quicker alignment, however i'm not an Airbus engineer so cannot confirm if they are the same.

JD-EE

Good morning. Thanks for the straightforward description of IT. I get it.
Your passage also describes the challenge facing fbw/pilot cooperation.

You underscore the challenge, for the goal is far from met.

"But nothing will break if you sit there looking blank while you ponder the situation, check a few dials, and then proceed."

I'll fly with gums, no offense.

MM43 - Special Request
 

The graphic that you posted on May 8th, 2010 is a favorite of mine. I love graphical information and that was outstanding, I don't know what program you used to format that but it is a thing of beauty. Would you mind undating that with current information. Would you mind sharing the program that you used to create it. Thanks.

As someone pointed out earlier. Don't hurry but lose no time. This is what training is supposed to inculcate in you. If you look at the Hudson ditching you can't find a boundary between thinking and acting. IMO this is more true of flying (especially in challenging situations) than other human endeavors. If you stop to think, its game over.

bearfoil, I'd meant to type in "But if nothing will break if you sit there looking blank while you ponder the situation, check a few dials, and then proceed." Does the addition of the if make more sense?

Of course, no harm no foul....:D

It does sound a bit like "What's it doing now??"......:ok:

JD-EE

When flying, you are typically looking at a variety of things that inlcudes looking at dials and instruments. It is not an either-or choice, it is "you must do both" and you don't get to choose not to.

The "look at it" is a subset of the required activity, true multitasking, particularly when flying in a condition other than straight and level flight.

Things breaking or behaving in unexptected ways just adds more multi to the tasking. (Apropos dealing with upset, and thus 447, task saturation is a critical point to understand in both training scenario design, and in task prioritization when things go wrong. I'll leave to psychologists and others the deeper details on why the human brain funcitons like that).

The analogy you used probably wasn't well chosen. (EDIT: OK, you updated it, I blundered on anyway ... )

It reminds me of what I used to tell ship driving colleagues about fuel when I flew helicopters from their ships.

"If your ship runs out of fuel, you will still float."

"If my helicopter runs out of fuel, it will first fall to the sea, and then not only not float, but turn upside down and begin to sink."

Their problem was in two dimensions, mine in three.

Maybe your non-moving, non-flying machine analogy is missing a dimension for appliciability. ;) (Oh, dear, back to dimensional analysis and Engineering 101, are we? :eek: )

I would think that the important design point would not be with the computer system as such, rather, relatively high wing-loading in the airplane itself, so that there is built-in stability, particularly in respect to up/down drafts and turbulence.

Vol Rio-Paris : les circonstances de l'accident connues en fin de semaine prochaine - LeMonde.fr

The black boxes of the flight Rio-Paris fished out, the investigators of the Office of investigations and analyses (BEA) try “to make them speak” so that the causes of the accident are known which costed the life of two hundred and twenty-eight people. The BEA announced, Friday, May 20, that the circumstances of the crash landing would be known in nearest end of the week.

BEA to release informations at the end of next week
 

A French TV announced that the BEA shall release at the end of next week some informations about the AF447 crash. These informations should be about the "circumstances" and not "causes" according to what said this journalist (BFM TV)
I think that the BEA is now under strong pressure from the political (and economical) power to move forward. I would not like to be the technicians in charge of the analysis. It is never a good thing to work under pressure, in many duties especially theirs.


EDIT (according to rotor12 post and his link to "Le Monde" article)


(...)

Par ailleurs, selon l'association de proches des victimes Entraide et solidarité AF 447, les enquêteurs du BEA travaillent sous pression dans la perspective du Salon du Bourget, grand-messe de l'industrie aéronautique. "On voit apparaître le Salon du Bourget à l'horizon fin juin et l'on sent un BEA complètement pressurisé pour sortir des informations, qui ne sont pas validées, qui sont contradictoires entre elles", regrette Robert Soulas, vice-président de l'association.


Moreover, according to the association of relatives of victims "Entraides et solidarité AF447", BEA investigators are working under pressure in view of the Salon du Bourget, high mass of the aviation industry. "We see the Paris Air Show at the horizon in late June and one feels a completely pressurized BEA to release out information that are not validated, which are mutually contradictory, regrets Robert Soulas, vice president of the association.
.../...

It is not simplistic, and I am not persuaded that people are not intentionally missing the point. Whatever the "Kind of airplane reverts to the pilots" it is obviously and necessarily one that (needs be) familiar, honest, and responsive. No "take out the book, run memory", and NO 'disregard/don't disregard' critical prompts. Maybe no a/p reselect or not, maybe so; no horizon. These are requirements that are unaddressed, to date. Review this thread and take note of the old guys who still question this format ??

There is a disconnect that perpetuates the discussion, now it may be arcane, picky, or other, but the end user is not well served by ego and pride. Allegiance to "one" format or the other is legend, going way back. If there are concrete reasons that give this argument life, (there are), then something is endemically wrong somewhere.

Well, I can't argue that the pilots did a good job of diagnosing the situation and determining the spurious warnings from the real ones. However, the dangers of such things happening due to a failed pitot/static system have become much more well-known and better understood since the loss of the Aeroperu and Birgenair 757s for the same reasons (and with similar symptoms - stick-shaker and overspeed warnings simultaneously). I'd imagine it takes something of a mental leap for pilots who are conditioned to implicitly trust their instruments to determine which of the instruments cannot be trusted.

The very nature of that kind of failure is hugely dependent upon the failure mode of the sensor concerned, and it is not a matter of being able to prescribe a set of actions that will definitively tell you that there's a failure in the pitot/static system. As such, I think the BEA are right to be cautious. If they had made such a prescription and an incident occurred where the stall had turned out to be real, then they would be in the firing line.

You make some very good points, though I think it's worth making clear at this juncture that the kind of computer/IT issue that you and I encounter daily is a world apart from those encountered in the kind of real-time systems developed for aviation.

That said, some of the better ideas to come out of those disciplines, particularly unit and regression testing, have been making inroads into mainstream software development practice over the last decade.

For non-IT people, the concept is to test each software function across the range of expected inputs, unexpected inputs and edge cases to make sure that the output matches the specification. You then do the same with multiple functions arranged in the manner they will be put together in the final system until you have a comprehensive set of tests that cover the whole application. These tests are then run throughout the software development and maintenance cycle, and if a single one of those tests fails, then you know you have a deviation from spec.

The main advantage of this is that you can prove on paper that the software matches the specification, and the secondary advantage is that if a change in one component causes unexpected behaviour in another, it will be caught in the test harness.

In the case of the A320's software, they went one step further and had two teams working in isolation providing separate implementations - if there was even the slightest match between the two, one team would be told to rewrite the functions from scratch. You then have a "quorum" of machines that can run the data through both implementations and any logical errors can be discounted.

Post #36 (Svarin)

Regarding this message :

2:14:20 FLR/FR0906010213 22833406AFS 1,,,,,,,FMGEC1(1CA1),INTERMITTENT

BEA reported :

Quote:
In any event, the effects of such a message could only be the disengagement of automatic systems, whose associated cockpit effect messages had already been transmitted at 2 h 10
To this should be added, from the same report, that a cockpit effect will only appear once in a given CFR/PFR. What we deal with is a CFR (Current Flight Report). I take it then that a A/P OFF message will not be repeated, since it was sent around 0210.

It is unfortunate that BEA would only give this item of information in that particular, obscure way. They essentially repeat themselves. However, reading between their lines would mean that A/P went OFF again without an ACARS message being transmitted (already done). So A/P had to be turned ON after 0210.

The implication being that erroneous but consistent data had persuaded the crew to re-engage A/P on false data that passed the acceptance criteria for engagement, only for it to disengage again (for unknown reasons) having done who-knows-what.

Was this interpretation discounted ?

Decisions
 

JD-EE, CogSim and bearfoil,

None of you is wrong, and none of you is completely right. Horses for courses?

Complete loss of thrust at a low height is one thing; anomalous indications in the cockpit at FL350 would be another – prior to any upset.

If and while there's time: DODAR, and repeat as often as required. If not, some prior consideration of classic, easily-diagnosed, once-in-a-hundred-careers failures may greatly simplify the thought process on the day. But, if at all possible, you still need to keep your co-pilot in the loop: and there's also a possibility he/she may have to correct your mental model. The latter process didn't happen in the Trident between Heathrow and the Staines reservoir, for reasons blind pew can explain better than I.

In deference to bearfoil's latest post: one hopes equally that the aircraft will keep the flight crew "in the loop".

Hi,

Flashback ..
A second reading of this release of BEA calls several criticisms:
What right does the BEA moves he can judge the morality of newspaper articles?
And in doing so .. take the role of a censor
This release of BEA is not better or worse than the newspaper article to which it refers.
The release of the BEA does not match the mission entrusted by the French government.
BEA's response would have been a denial or approval of the contents of the press articles
It appears nowhere in this release

Wait and see
 

Last news:
First findings on Rio-Paris crash next week

This aspect of the "conversation" has had me thinking also... since, in everything except Mechanical Backup (a realm of last resort if there ever was one), the fact of the matter remains that some interpretation of the crews input the sidestick is translated into a control surface movement - the idea of "Direct" really is a misnomer.

From the prior link describing the A330 control laws:

Am I reading this right - "max elevator deflection is a function of CG", meaning there are still electronically applied limits to what the pilot can demand (in this case, of the elevators)? Sounds like a "law" to me?

Also, my understanding of the side-stick is that the Airbus stick is quite a different design to that of the F-16. The F-16 stick only moves a very small physical deflection and is more of a force sensor (I worked F-16 simulators about 17 years ago!), while the 'bus stick is a position sensor. That in itself though is merely interesting.

As gums has very clearly explained in several posts, FBW control laws are all about constraining the aircraft within an envelope, with the intent of keeping the aircraft from getting into an attitude that is known to be "bad news", however highly agile military fighter jets (and *creative* military pilots) were clearly able to find blank spots outside of the controllable envelope envisaged (the deep stall example for the F-16). It wasn't clear to me whether the control laws were then modified on the Viper to prevent the aircraft getting into this condition...

However the point of the preceding paragraph is to ask a simple question:

Is the Airbus inherently a safer aircraft (not comparing to anything here) because of the FBW system and associated control laws?

Is is hard to ask that without invoking a comparison, but I am not inviting a A vs B discussion here, or wanting to see one evolve. The point is we have limiting systems in cars for example - traction control, limited slip diffs, rev limiters, ABS brake systems. Presumably these are in place because for the greatest majority of time safety is improved because of them - but I guarantee there are a very few cases where evidence indicates the such a system made some accident worse.

My position is that life is inherently dangerous and we are constantly tossing the dice, aside from sitting in the middle of a large field for your entire existence, there is a level of risk that may lead to harm when we do anything, and in particular any human construct can and at some point will fail in some way and possibly cause harm to someone (buildings fall down, dams fail, cars crash, airplanes stop flying). What Airbus has tried to do is prevent most of the obvious bad things from being allowed, through control laws, but as with any control system, once you exceed the design limits, well, bad things may occur.

It's clear that Airbus has done a pretty good job - there are thousands flying. Obviously this is not to say things cannot be improved. Of course improvements are possible.

But, sadly I suspect (yes, this is WHOLLY opinion) this case will result in findings that indicate some sequence of poor human decisions leading to the aircraft being in a very bad place at entirely the wrong time (I see the phrase "holes in Swiss cheese lining up..." applying). Whether it is then appropriate to blame the computers for not saving the day seems unfair. Somewhat akin to driving my car at 100MPH toward a cliff-edge and then blaming the ABS brakes for not stopping the car in time - I'm sure the system would do the best it could until the wheels leave the ground, at which point I become the passenger... Time will tell.

Hi,

Another source:
Google Vertaling

Original source:
Les circonstances du crash de l'AF447 dévoilées dans une semaine - Le Point

Note:
Circumstances but not the causes

Methink all Pprune reader know already the circumstances

But the kind of situation you're describing does not simply apply to aircraft of the same or of a later generation than the A330. Two incidents that spring to mind are the Roselawn ATR-72 crash, and the China Airlines 747 incident over San Francisco. In both cases, the automation tried to manage the flight to the best of it's ability until it reached the limits of it's control regimen and handed control back to an unsuspecting crew. In the case of the latter, there was enough altitude to perform a recovery, but sadly in the case of the former there was not.

Every technological advance, both inside and outside of aviation, has had it's detractors - it's just human nature to be suspicious of change. That doesn't make either side of such "religious" discussions correct.

I believe the only Western transport category jets to retain such a thing are the 737 and the DC-9 derivatives.

Having finally got my hands on a copy of Davies' "Handling The Big Jets", something leapt out at me when I was reading it - and that is the fact that beyond a certain size, transport category jets are simply too large to have effective manual reversion. Having aircraft "feel" transmitted to the pilots via artificial means is something that has been provided since at least the days of the Comet 1.

What you are describing doesn't sound like a "limiting" control law (in fact in Direct Law there are no limitations as such), but an implementation of precisely this kind of "artificial feel", albeit implemented in software as opposed to hydraulic or mechanical devices.

Davies points out that without these artificial feel units, then it would be possible to very easily fall into an upset and overstress the airframe by virtue of the full authority that powered flight surfaces provide.

In many respects, yes. However that does not make the pilot immune to maintaining situational awareness to the best of their ability.


Well, the computers are getting their information from the same sensors as those driving the pilot's instrument display. At this stage a computer is no use as it is outside the parameters within which it was designed to operate - both Airbus and Boeing are well aware of this, and it is for this reason (among others) that there are still two warm bodies in the flight deck of even the most automated airliners, and will continue to be for the foreseeable future.

Lonewolf 50
Regarding in-flight realignment. I'm no expert, but I'd venture to say that on a theoretical level:

As long as you can fly in a reasonably straight line, and can build a GPS trajectory of that line, you can then realign the solid state gyro by applying corrections to whatever the gyro displayed during the straight line.
Additional data needed is airspeed and the airplane's lift polar curve data , which gives the angle of attack for any given airspeed (which affects the indicated pitch angle).
In the roll plane, calibration can be aided by the fact that according to aerodynamics, any bank will induce a turn (assuming no sideslip). The actual turn rate can be calculated from the GPS track. In fact there is a definite relationship between airspeed, turn rate and bank angle, so theoretically it should be possible to calibrate the roll axis even if the plane performs a constant rate turn.

I have no idea of the A330's capabilities in this regard, but e.g. the Garmin 1000 integrated instrument system for GA is capable of in-flight realignment.

jcjeant,

I respectfully disagree with your second reading and criticisms of the BEA and their response to this news article. They made somethings crystal clear:

1. They and they alone have the responsibility to release information concerning the AF447 investigation.

2. The BEA reserves the right to confirm information published from an external source as being accurate and true.

3. The BEA investigation has not reached the stage of determining cause of the accident.

and,

4. Articles, such as this one, does the families of those who perished no good and could cause harm as it is not information either released or concurred by the BEA.

So, by not concurring or denying as you say, the article and its content is null and void, that is to say, not authoritative as no conclusions have yet been reached.

The BEA is doing exactly what the French Government has chartered it to do, investigate the accident, establish factual information, take the factual information derived and establish factual cause/causes, probable cause/causes, and those causes that cannot be determined from all the developed data. Once in a while they get sidetracked by being required to remind the public as to the source of accurate information, that source is the BEA.