An attempt will be made at taking a look at the current state of the art of PHM and venturing some guesses as to where the discipline is headed; some modest suggestions will be offered. The following issues will be addressed, or at least evoked, with a goal to stimulating discussion. - Sensors: how to capitalise on technological progress for PHM? What progress is required? - Intelligent data processing. What are the advances and the challenges on fault detection, fault separation (diagnostics) and prognostics? - The various approaches to prognostics: reliability-based, model-based, statistics based; in particular, what are the challenges and potential breakthrough with adaptive modelling? How can model-based and statistical approaches be made truly complementary? How can slow degradations be sensed effectively enough to permit prognosis of remaining useful life? - Data architecture and MMI aspects. - What is meant by validation in the context of PHM? - What are the added challenges in moving from equipment PHM to fleet PHM?