It is rare for paediatric palliative medicine physicians to have to break news of a diagnosis of a life-limiting condition (LLC). Usually that has been done weeks or months before. It is much commoner for us to be faced with the question: "how long?" Clearly it is a question that cannot be answered with certainty, and yet a great deal may depend on it. While palliative care (PC) should ideally be available from diagnosis, the need for 'active' practical PC intervention will fluctuate during the course of a child's illness. For most children this will last months or years; often decades. Typically, there will be several periods during which death seems likely before the final terminal episode, particularly among children with non-malignant LLC. Optimal management of all these episodes depends on anticipating the child’s needs in order for timely interventions - or withdrawal of interventions - to be considered. This consideration in turn depends on recognising that such an episode has begun. Providing adequate PC critically depends on making a diagnosis of dying. In this article, we will consider why it is important to make a 'diagnosis of dying', briefly review some of the tools available to help, and examine some of the evidence from published literature in children and adults.