Andy Eick and I developed the SemApp - a system that coherently brings together information from disparate sources to form an Ultra-Large Scale unified compute and storage platform. Built on open source compute and storage infrastructure, SemApp is both a unified data store and a massively parallel computational engine. Domain data models are preserved, yet decoupled from the storage model. Data ingest is distortion-free and fast. The variety of domain-specific interfaces to the data are preserved while the system presents a single unified interface that spans the entirety of its holdings. The unified interface enables entirely new classes of analytics, tools, and applications to be developed for building, discovering, and exploiting connections within and across semantic boundaries. Processing products may be formed and distributed as finished analytic artifacts, such as visualizations and reports, while the content continues to participate in the increasingly rich web of data and processing within the SemApp. When populated with diverse data and processing at scale, the SemApp exposes a new substance to science and serves a laboratory for data-intensive research and development.