For higher-level reuse, activities from preservation to knowledge reutilization yield long-term benefits but are often perceived as tedious due to their lack of immediate payoff. These processes, including knowledge harvesting, storage, retrieval, and reuse, can seem unappealing. However, integrated development environments for lower-level reuse, such as source code, have shown how to enhance these activities through completion mechanisms and recommender systems, making community knowledge readily accessible to programmers. Unfortunately, similar support is lacking for modeling. This work addresses challenges related to representation, harvesting, evolution, and retrieval within modeling. We propose a tailored approach for modeling with UML or similar class diagrams, resulting in a knowledge-based recommender system utilizing property graphs and metagraphs for broader applicability. Additionally, we offer a cookbook for developing such systems, including schemas for model recommendation production in operation-based model recommenders, informed by our experiences with HEREMES. By incorporating contextual information from monitored modeling operations, this system can be characterized as an operation-and-knowledge-based recommender system that (semi-)automates tedious tasks.
