The paper provides a comprehensive review of the development of generic user modeling systems over the past two decades. It traces the evolution from early user modeling shell systems to commercial user modeling servers, highlighting their purposes, services within user-adaptive systems, and design requirements. The paper discusses various architectures, including shell systems, central server systems, and future user modeling agents. It reviews several implemented research prototypes and commercial systems, such as GUMS, UMT, BGP-MS, DOPPELGÄNGER, TAGUS, and commercial tools like Group Lens, LikeMinds, Personalization Server, Frontmind, and Learn Sesame. The characteristics of these systems, such as generality, expressiveness, inferential capabilities, and client-server architecture, are analyzed. The paper also explores the future of generic user modeling systems, considering mobile user models, user models for smart appliances, and multiple-purpose usage. It concludes by discussing the need for standardization and the potential for a wide variety of specialized user modeling systems in the future.The paper provides a comprehensive review of the development of generic user modeling systems over the past two decades. It traces the evolution from early user modeling shell systems to commercial user modeling servers, highlighting their purposes, services within user-adaptive systems, and design requirements. The paper discusses various architectures, including shell systems, central server systems, and future user modeling agents. It reviews several implemented research prototypes and commercial systems, such as GUMS, UMT, BGP-MS, DOPPELGÄNGER, TAGUS, and commercial tools like Group Lens, LikeMinds, Personalization Server, Frontmind, and Learn Sesame. The characteristics of these systems, such as generality, expressiveness, inferential capabilities, and client-server architecture, are analyzed. The paper also explores the future of generic user modeling systems, considering mobile user models, user models for smart appliances, and multiple-purpose usage. It concludes by discussing the need for standardization and the potential for a wide variety of specialized user modeling systems in the future.