This submission presents the graduate metadata courses offered by the X university and reports a case study of an experiment in the design of an advanced graduate metadata course. We made the changes to course design to facilitate the skill-building component and provide more efficient link between content-based learning and skill-based learning. Our goal was to ensure metadata learners develop not only the in-depth understanding of and interest in advanced metadata topics, but also the ability to clearly see the connections between these topics. In particular, the experiment included integrating the learning of the process of designing a local metadata application profile with evaluation of the quality of metadata, including assessing the ability of a standard metadata scheme or an application profile to capture and adequately represent important and unique attributes of information objects in a special collection. The benefits of this approach to the quality of learning were measured and are discussed.