Rationale (from mailing list): " In any case as we're using fully qualified artifact id there's no possibility of clashing and it's kind of duplicating the information to put the hierarchy information both groupId and artifactId. Also, the groupId really represents the project. Note: If we were to use different groupIds using the dotted notation, the m1 repository would look pretty crowded as we'll have the following at the top level: ibiblio: org.codehaus.cargo/ org.codehaus.cargo.core/ org.codehaus.cargo.core.api/ org.codehaus.cargo.core.containers/ [...] "