14
Business
STONE GENERATIVE AI: REVOLUTIONISING HVAC & R SOFTWARE DESIGN
By Eamonn Ryan, based on a CAREL webcast
A recent CAREL webcast positioned STone IDE, empowered by generative AI, as an invaluable tool for software designers in the HVAC & R industry.
CAREL’ s innovative solution, STone IDE, represents a comprehensive ecosystem. As highlighted in the webcast- featuring Gabriele Putzolu, CAREL group solution manager: HVAC commercial, and Gianpaolo Dall’ Occo, CAREL software tools product unit manager: digitalTools software platform- STone IDE is not just another Integrated Development Environment; it ' s a paradigm shift towards streamlined, collaborative and secure unit development.
Putzolu emphasised STone ' s holistic approach, evolving from a state-of-the-art IDE into a complete ecosystem designed to simplify unit development across its entire lifecycle, from initial design to ongoing field maintenance.“ STone ' s foundational principles are: fostering efficient teamwork through structured, tested, and standardised languages- all while offering a rich library of resources. STone is designed to support a team of professionals to work together on the same project, increasing the efficiency of everyone involved through flow and team-based development," Putzolu stated. The platform ' s inherent modular architecture allows for simplified coding and the creation of custom libraries, enabling developers to reuse proven routines across various applications, saving valuable time and enhancing control.
Integration with popular version control systems ensures seamless collaboration, even across geographically dispersed teams. Furthermore, STone IDE incorporates on-target debugging and a comprehensive test suite, including unit, automatic and non-regression tests, facilitating robust and reliable software development. Putzolu summarised the core values of STone, emphasising " speed, teamwork, and high-quality results " achieved through its efficient, modular architecture. Cybersecurity, a critical concern in today ' s industrial landscape, is a central tenet of STone, integrating features aligned with market standards to safeguard against unauthorised access and potential risks.
Putzolu emphasised the strategic integration of emerging technologies like artificial intelligence and advanced simulation systems within STone. " This allows programmers to develop highvalue software efficiently, opening up new possibilities for innovation and optimisation in the HVAC & R systems," he asserted, setting the stage for the demonstration of STone ' s generative AI capabilities.
Dall’ Occo took listeners through practical examples of STone ' s AI-powered features, starting with the often-tedious task of code commenting. He showcased STone Copilot ' s ability to automatically generate comprehensive and accurate comments based on the existing code. He demonstrated how a simple right-click could prompt AI to add detailed headers and explanations for function blocks, even identifying and describing the specific variables used by the developer.
Dall’ Occo showcased how a natural language prompt, such as " create a function block that counts from zero to 1 000 every second ", could instantly generate the corresponding code. He illustrated AI ' s ability to identify and correct minor coding errors, like a missing semicolon, further highlighting its intelligent assistance. Code quality is a common concern, especially with junior developers. Dall’ Occo further demonstrated STone Copilot ' s optimisation capabilities, showcasing how it could refactor less efficient code, like initialising a large array, into a more streamlined and readable format, even intelligently retaining and highlighting potentially critical values within the code.
Perhaps the most compelling demonstration focused on the generation of unit tests. Dall’ Occo illustrated how STone Copilot could automatically generate test classes and methods for existing functions, significantly reducing the substantial time developers typically spend on this crucial aspect of software development.
14
RACA Journal I June 2025 www. refrigerationandaircon. co. za