Integrating cloud with fog/edge is a main trend in networking. Many cloud computing applications have been shifted to the edge/fog domain. Such paradigm shift offers new opportunities for pervasive computing. An example is AgriTalk, an Internet of Things (IoT) application development platform for smart agriculture. By integrating cloud with edge/fog, this article describes how AgriTalk addresses six issues for developing edge/fog agriculture applications. These issues include device domain development, application generation and bug detection, sensor failure detection and calibration, big data management, Artificial Intelligence (AI) provisioning, and data privacy. We show how AgriTalk integrates fog/edge applications and use rice blast detection and piglet crushing mitigation as two examples to demonstrate that fog/edge computing is a better solution than cloud computing. Compared with cloud computing, fog/edge computing reduces the delays by 50 percent in AgriTalk. Through the low-code no-code approach, AgriTalk allows the farmers to create and maintain fog/edge agriculture applications by themselves.