Question: do docs contain misinformation about the CRC-16 polynomial? about libacars HOT 2 CLOSED

The document you are referring to, is very old and is not even a standard, but just a report from some tests done by FAA. It says: "To detect messages [did they mean "errors"?], undetected by ACARS, (...) the ground station had the capability to compute and append CRC codes to each message". So i guess that they added an extra CRC to each message as a double-check, and it used x^16+x^15+x^l3+x^5+x^3+x^1+1, because: "The probability of an undetected error is greater if the polynomial applied at the transport layer is the same used at the link layer (CCITT).". So they wanted to minimize that probability.

ACARS decoder in libacars uses the la_crc16_ccitt function, while la_crc16_arinc is used only by the MIAM Core v2 decoder. MIAM v2 uses CRC16 with 0x1021 polynomial as well, but with a slightly different bit ordering, so it was easier for me to implement this as a separate table rather than rearrange the input bits to match the order expected by la_crc16_ccitt. The name of this function comes from the fact that this CRC is defined in the ARINC Report 665. la_crc16_arinc665 would probably be a better name for it.

from libacars.

Thanks for the info.

la_crc16_arinc665 would probably be a better name for it.

The most widespread (used in different heades and catalogues) names for that one are CRC-16/IBM-3740 and CRC-16/CCITT-FALSE.

from libacars.