Code Characterisation

What a rigorous error-correcting-code characterisation should measure before deployment or comparison.

Code characterisation is the process of building a technical profile of a code and, where relevant, its decoder. The purpose is to replace vague labels such as 'strong code' with measurements that can be compared against an application requirement.

Core idea

The measurements depend on the code family. They can include length, dimension, rate, rank, parity overhead, minimum-distance evidence, decoder success curves, frame error rate, failure modes, convergence behaviour and sensitivity to implementation choices. For iterative codes, error-floor behaviour and rare structural failures can be as important as the waterfall region.

Engineering relevance

A useful characterisation preserves the test conditions and separates theoretical properties from decoder-dependent results. Software, simulation and hardware results should be labelled clearly. When hardware is involved, latency, throughput, memory, area, clock frequency and power-related constraints may need to be included.

Key points

  • Characterise the code and the decoder separately where possible.
  • Record the channel or error model and all important decoder settings.
  • Report recognised failures separately from silently wrong outputs.
  • Preserve enough provenance for an independent party to reproduce the result.
Public technical resource

This page explains established concepts and engineering principles. Proprietary JobMettle methods and internal ThinkQuick technology are not disclosed.