The science of psychiatric diagnosis and classification has its critics. Reference Seligman, Snyder and LopezSeligman (2005), the father of positive psychology, has even suggested that we ‘develop a nosology of human strengths – the “unDSM–1”, the opposite of DSM–IV’. Nevertheless, as ‘diagnosis is intended to be one of the strongest assets of a psychiatrist’ (Reference TyrerTyrer 2009), clinicians need to think about and be involved in the forthcoming revisions and harmonisation of the two major classifications ICD and DSM. Sartorius (pp. 2–9) gives a behind-the-scenes view of the revision process. There are many vested interests: not just clinicians, but governments and NGOs, lawyers, researchers, public health practitioners, Big Pharma and patient groups. Vast sums are at stake – everything from welfare benefits and compensation claims to research budgets. Concerns include the use of national classifications to facilitate political abuse and of diagnostic labels that are seen as stigmatising or are used to stigmatise. Like Sartorius, Thornicroft (pp. 53–59) singles out chronic fatigue syndrome, ‘bitterly contested in terms of its status as a physical, psychiatric or psychosomatic condition’ and viewed by healthcare staff as a ‘less deserving’ category.

Should the classifications use categories or dimensions? A dimensional approach seems impractical, although dimensions could be used to augment categorical definitions, as with severity of depression in ICD–10. If categories, should revision involve lumping or splitting and on what basis? Goldberg (pp. 14–19) is critical of the proliferation of diagnoses necessitating ‘a diagnostic paella’ of multiple comorbidities. He tantalises us with his proposal of a small number of large groups of disorders, each group based on aetiologically related criteria. But there are difficulties. For example, should bipolar disorder be considered a mood disorder, a psychotic disorder or separate from both? Craddock (pp. 20–22) is not convinced that there is (yet) sufficient evidence for Goldberg's categories, believing that it will take at least 5–10 years before our understanding of brain dysfunctions will justify major change in classification. The complexities involved in clarifying neurobiological mechanisms are well illustrated in a robust exchange in our Correspondence (pp. 76–80). Gillman stresses the importance of understanding the pharmacodynamics and pharmacokinetics of interactions of antidepressants. Palanniyappan et al agree, but reach an only too familiar conclusion: insufficient clinical data and a lack of biological markers of pharmacological mechanisms.