ACARS data origins
Publicly known ACARS messages originate from a June 4th (evening?) "French television station France2 produced the ACARS messages from the last minutes of the flight of the A330. The data was clear on screen and has since been reproduced here."
Flight 447 was lost on June 1st. So we have:
* 3 to 4 days interval between loss of Flt 447, and appearance of the public ACARS data.
* The data 'release' involved both error-prone visual/manual transcription from a screen capture, and a mainstream media source.
* At least one potential factual error has been identified in the data.
* The 'known' ACARS data is very likely incomplete - it omits non-maintenance streams, and possible records prior and subsequent to the available ones.
* There is no formal provenance to even the 'known' ACARS messages. I do not have a transcript of the TV station broadcast, so don't know the actual wording used. I must consider that there might have been commentary to the effect of 'typical', 'representative', etc in conjunction to the on-screen text. In any case, TV stations are under no legal obligation to be truthful, and we know nothing of their source.
One thing I find a little odd. All the copies of the 'ACARS text' I can find are images, as opposed to plain text files. Also, someone has gone to the trouble of creating those images complete with faint alternate line shading, as if they are direct scans of line printer output on typical fan-fold tractor feed paper. Yet these are supposedly manual transcriptions, from the screen of a TV news item?
I found two sources of this image:
http://img197.imageshack.us/img197/3410/acarsaf447e.png
http://www.eurocockpit.com/images/acars447.png
They are substantially the same image, but they are not identical. Examining the images closely in photoshop, via a red/blue overlay (see http://everist.org/flt_447/a-b_compare_flat.png) determines:
* They are exactly the same x-y size.
* They are identical in almost all pixels, except for three points of difference, where characters have been altered, added, or moved. (See yellow shaded items in my image.)
* Both versions very likely created by the same source, due to exact pixel correspondence of the non-aliased font used. One presumes that one or other version is an update, with corrections as errors from the visual transcription are noticed.
* They are most definitely electronic image constructions, rather than processed scans of paper documents. There are zero image/colour defects of the type resulting from scans, also the shifting of the "CTL SEC 1 FAULT" item could not result from a scan error.
* Both are identical in the "NAV TCAS FAULT" and the associated '3443' JASC code, which is suggested should read '3445'. (Shaded blue in my image.) We do not know if this is a transcription error, or if the original code really is '3443'.
From the above, one can conclude that the known images really are the product of a manual transcription from a TV frame-grab, as opposed to actual printout scans, merely *said* to be transcriptions to protect someone from charges of leaking confidential documents. Thus the TV station broadcast was the original source.
Now, I'm not suggesting that the known ACARS messages are a fabrication. However one must logically bear in mind that lacking a formal release origin, they easily _could_ be a fabrication.
Worth remembering, given the comments from several official sources that the flight recorders are unlikely to ever be found.
---Some previous relevent posts---
http://www.pprune.org/rumours-news/376433-af447-87.html#post5001546
Being curious about the source of the two pages of ACARS messages, which are labelled 28/256 and 29/256, I looked around to see where they originated. Amazingly, this seems to be it:
http://www.flightglobal.com/blogs/unusual-attitude/2009/06/af447-accident---icing-pitot-t.html
So the crucial text on which everyone is basing analysis, was hand copied from a screen visible during a TV interview! Furthermore I gather from other comments here, that this is just one ACARS stream (maintenance messages?) of several. Also that it isn't publicly known for certain that these two pages are even all the relevent messages from that stream.
http://www.pprune.org/rumours-news/376433-af447-89.html#post5002147
"I'm new to this forum, so hello to all gentlemen.
Some tiny details in the ACARS message list presented by Air France suggest that at least part of it may have been manually edited, a process prone to human error. Looking closely at it, I found a discrepancy between the warning labelled "NAV TCAS FAULT" and the associated JASC code. This code, 3443, stands for NAV/DOPPLER SYSTEM in the JASC classification. However, the a/c was not Doppler equipped. TCAS should be coded 3445.
All other codes appear correctly related to their labels.
Whatever explains this discrepancy, including a possible use of a slightly different classification by Airbus, I believe it should be considered doubtful that a TCAS FAULT warning ever happened."
http://www.pprune.org/rumours-news/376433-af447-89.html#post5002195
"There have been contradictory, vague and confusing descriptions of this 0214z warning and it would be helpful if one of the experts would clarify exactly what it means."
http://www.pprune.org/rumours-news/376433-af447-89.html#post5002268
"What is shown on page 27?"