Extensible Software Synthesis for Embedded Ubiquitous Learning Systems

Embedded systems are generally designed for specific purpose such as alarm systems and learning systems. These devices work in particular scenario and repeatedly perform a few particular tasks, which result in the shortage of extensible ability to execute diverse tasks. This work proposes software synthesis algorithms for generating code which can execute on embedded ubiquitous learning system. Moreover, the extensible problem is solved by extensible software synthesis (E2S) technique. The E2S comprises CEO, CEOAS, CBOAS and CTOAS procedures that are developed for synthesizing software with extensible ability. Each procedure first separately synthesizes its extensible component, so-called EC, and then applies it to construct the various tasks. Next, the E2S synthesizes various ECs to generate codes to serve complex and diverse functionalities. The synthesis codes are verified on an ARM emulation platform. Experimental results on benchmarks show the E2S achieves not only extensible ability but also the time of software synthesis which meets the requirement of specification.