Abstract
With the rising popularity of graph-based approaches to data management, exposing the content of traditional, often relational, sources as (knowledge) graphs becomes more and more relevant. In such scenarios, Direct Mapping approaches are often used to automatically transform such sources into graph-like formats. A “fundamental” property of these transformations is to be information preserving, that is, it should be always possible to (algorithmically) reconstruct the content of the original database. Information preservation, along with other “fundamental” or “desirable” properties proposed in the Semantic Web literature, has never been put into correspondence with over 40 years of extended literature coming from the traditional database perspective. In particular, to the best of our knowledge, it is unknown how classical results on information capacity, dominance, and equivalence, tailored towards specific tasks such as query answering or data update, relate to the results and definitions from the Semantic Web world.