Medical journals have many possible functions, but the main one for most is publishing science. They are actually better at campaigning and agenda setting, rather like the mass media. Journals are now beset with problems, including failing to include data, publishing lots of poor quality material, being slow to publish, publishing research that is either not reproducible or fraudulent, encouraging waste in the system, failing to be transparent, and exploiting academics. New ways of publishing science are appearing, and a better system would be for the grant proposal, protocol, and full data to be published on a database with the whole process transparent.