After more than a decade of unprecedented growth of centralised cloudcomputing facilities, scalability and cost are the main drivers to shiftdata processing closer to the data path, including at the origin ofdata-producing devices. In large-scale sensing scenarios, including insmart cities and in billion-device telecommunications, concentrating alldata processing purely in cloud-hosted software units is not feasibleanymore. Several concepts (edge computing, fog computing,near-memory/near-data computing) have emerged, each with their owncommunity and with, thus far, fuzzy and overlapping terminology. Inthis project, we propose to investigate a first software framework forprogrammable continuums, based on recent gains in developer-friendlyassisted application decomposition into microservices, to enablescalable and optimally distributed data processing applications. Withthis applied research approach, we aim at giving an answer to thecurrently open question: How do we engineer software for continuums?