The concept of the locative web has been around for a while. We should expect the web principles of openness and uniform access to provide a solid alternative for developing locative systems, but the locative web has never succeeded. In this study, we extend the original concept of locative web with a conceptual framework to support the large-scale publication of locative web content by many independent publishers. To pursue this vision, we explore the concept of a locative resource as a small, data-centric web resource that represents the information and actions meaningful for a specific locative situation. These resources can be automatically generated in large numbers from datasets, and each can serve a very concrete situation. Our research aims to study the tools and processes that could support the convenient and cost-effective transformation of locative datasets owned by multiple independent entities into a global web of locative resources. As part of the experimental work, we developed a prototype implementation of a locative sites toolkit. We also developed four deployment case studies, in which we used that toolkit to generate locative resources from diverse types of datasets. The key contributions are a set of novel concepts for locative web systems and a set of guidelines for dealing with data heterogeneity and allow many independent entities, all of them with their particular datasets and publication goals, to conveniently publish their data as locative resources while promoting their convergence as part of a global locative web.