Proposal of a clean architecture for a crowdsourcing-based system for missing persons search in Mexico

Abstract

Crowdsourcing systems applied to the search for missing persons face significant challenges in terms of maintainability, performance, and security, particularly when human validation of reports is required to mitigate false information. Since the system has not yet been implemented or subjected to stress testing in a real-world environment, this article is limited to a methodological proposal for architectural design and evaluation prior to implementation, aimed at managing citizen reports of missing persons using geolocation. The process begins with three architectural proposals, which are compared using the Architecture Tradeoff Analysis Method (ATAM) [1], with the sole objective of identifying the most suitable alternative. The evaluation is based on a utility tree constructed from quality attributes prioritized by the project stakeholders. The results indicate that Clean Architecture provides better domain isolation and significantly facilitates unit and integration testing; however, it introduces greater initial complexity in the definition of ports and adapters. In addition, this work contributes an architectural evaluation methodology for social-impact systems developed under small-team constraints and requiring human validation.