If you don’t synthesize knowledge, scientific journals become spare-parts catalogues for machines that are never built.