> shouldn't an abstract say something about the findings? One counter-example is in security / defence work where the existence of a project is unclassified bu

> shouldn't an abstract say something about the findings? One counter-example is in security / defence work where the existence of a project is unclassified but the details are protected in some way. Not reporting the findings allows the paper to be indexed in libraries / registries without 'contaminating' the library indexes with classified information. The other example (rightly or wrongly) is where a publisher wants you to be able to find via bibliographic search tools a paper hidden behind a paywall but still wants you to pay to access the paper. If the abstract revealed too much, it defeats the paywall, from the publisher's perspective.