What are books for these days? The expansion of electronic and often freely available material in the sciences provides overwhelming choice for the reader. Traditional textbooks may seem quaint, especially works which seek to be comprehensive sources of knowledge in a large field. Most readers in Western countries are likely to be affiliated to a healthcare education provider which will provide free access to the major journals. The problem for the reader now is having the skill to negotiate the mass of information which is out there, and to understand hierarchies of information.

Yet textbooks can still be relevant. In this fourth edition of the admirably simply titled Dementia, the editors bravely attempt once again to assemble a plethora of subject area experts for a book that seeks to cover the full range of scientific and care overviews. All 80 chapters are discrete, with no subchapters. I approach this book as a reference work to help with tricky patients, to get a quick baptism in areas I know nothing about, and to help me update lectures to non-expert but awkward audiences like medical students, who will invariably ask difficult and important questions. Others will use it for different reasons, but will surely find its scope and authority useful.

Every major area of scientific enquiry is covered here, with an emphasis on those which are clinically relevant. Full referencing is provided for each chapter – essential in my view if this book is to maintain its relevance. Chapters on stem cell therapies and on biomarkers for Alzheimer's disease may have seemed scientific esoterica in the past; now they are directly relevant to questions that patients and families ask in our clinics. There are always problems with multi-authored textbooks such as varying ability in written English and more or less scientific approaches. For example, the degree of evidence base available to inform a chapter on physical design and dementia is much less than that for a chapter on health economics or on the neuropathology of Alzheimer's disease. This book is meant to be an attempt to be comprehensive and meet the needs of a wide audience – so we end up with essentially an expert's subjective view of what physiotherapy offers in dementia followed soon after this by a dispassionate review of existing systematic reviews and randomized controlled trials of non-pharmacological therapies for behavior problems in dementia. These differ greatly. Does it matter? If read one after the other (as I did) it seems to matter as we seem to be reading from two different books. But most readers will approach this book a chapter at a time; as such, it fills an ideal gap on the bookshelf, and I have just made at least a dozen changes to a regular lecture based on its content and I will use it for the other reasons above. No-one would (or should) try to read a full book like this; it's the ideal accessible and genuinely useful reference work for a broad range of readers.