Xiangjun Kong, Xia Wang, Xishi Tai
Carboxylate-ligand-based coordination polymers and metal complexes form structurally adaptable crystalline systems, spanning discrete complexes, one-dimensional chains, two-dimensional layers, three-dimensional frameworks, and MOF-like architectures. This adaptability arises from diverse carboxylate binding modes and metal coordination preferences; auxiliary N/O donors are included only when carboxylate coordination remains central. However, the increasing availability of structural and electronic descriptors has not always been matched by equally rigorous validation of structure-function relationships. Hirshfeld surface analysis, energy-framework analysis, density functional theory, adsorption simulation, molecular docking, and machine learning are descriptor-generating methods; neither a method nor its output alone establishes descriptor-guided design. This review critically examines carboxylate-based coordination systems from the perspective of descriptor-guided function-oriented design. We discuss ligand-level regulation, metal-center effects, dimensional evolution, and what supramolecular, electronic, adsorption-related, and biological descriptors can and cannot prove. Functional studies covering luminescence and sensing; catalysis, adsorption, and small-molecule transformations; and bioactivity are evaluated by evidence strength. A claim-centered Descriptor-to-Function Evidence Ladder distinguishes structure reporting, descriptive descriptor use, post hoc association, controlled comparative trends, mechanism-supported validation, and prospective experimental confirmation. It evaluates a specific descriptor-function relationship rather than overall paper quality. Future progress should rely on standardized reporting, comparative structural series, function-specific validation, and transferable descriptor-function relationships.