It was reported that the Land Transport Authority (LTA) has reviewed its rail disruption counting methodology to omit disruptions due to "external factors" from the overall measurement ("LTA using new way to count train delays"; last Saturday).

However, to a commuter, a disruption is a disruption, and its causal factors, whether external or internal, do not matter.

I propose that the LTA adopt reliability engineering principles in the analysis of its disruption data, instead.

This requires taking a holistic view of the overall disruption data and designing redundancies in the network, such that it remains dependable despite sub-system failures.

Disruptions due to external factors could represent human interface issues that should be analysed for preventive measures or redundant features to ensure continuity of service.

Such data should not be discarded.

While an accurate breakdown of technical failures would help engineers pinpoint efforts for improvement in specific areas, the authorities must not lose sight of the overall disruption data that can point to other latent issues that affect the network's reliability.

Ang Kheng Wee