The QCon Global Software Development Conference, held in Shanghai, brought together industry leaders from Alibaba, Tencent, Intel, and other major companies to discuss the opportunities and challenges of front-end development in the new era. With the rise of large models and the advancement of technology, front-end development is experiencing unprecedented opportunities and challenges, contrasting with the pessimistic view that front-end is dead.
Alibaba, Tencent, and Intel Lead the Discussion
The conference featured a special session titled Front-End Opportunities and Challenges in the New Era, organized by InfoQ. The panel was moderated by Wu Ziqi, the Technology Responsible Person for Alibaba Tmall Group’s Vision Pro, who has extensive experience in terminal architecture, performance optimization, and cross-platform container technology. He is also the founder of Walle, an industry-leading and academically advanced end-intelligence engineering system.
Innovations in Architecture and Performance Optimization
The conference highlighted several key topics, including architecture innovation, performance optimization, and stability enhancement in the context of emerging technology platforms and complex business scenarios.
1. Tmall’s Vision Pro Architecture Practice
Tmall’s XR Engineering Technology Team quickly adapted to Apple Vision Pro, offering a unique immersive shopping experience. Jiang Zhiyuan, a senior R&D engineer at Tmall, shared his insights on the architecture design of Tmall’s Vision Pro version, which successfully supported its development and launch. He discussed the technical challenges and innovative solutions faced by Tmall in adapting to the Apple Vision Pro platform, including engineering architecture, software design, and spatial content development. Jiang’s presentation provided valuable references and inspiration for developers facing similar platform architecture design challenges.
2. Tencent QQ’s Performance and Stability Anti-Degradation Practice
Tencent QQ’s team has developed an innovative design to build a performance and stability anti-degradation system, significantly enhancing the overall performance of terminal applications. Qiu Shaoyong, a senior development engineer at Tencent, shared his insights on how QQ’s team effectively addresses performance issues in complex scenarios with a large code volume and frequent iterations. He detailed QQ’s use of new technologies like xctrace for performance collection, startup time optimization, and anti-degradation strategies, as well as the application of AI in automatic analysis and repair of performance issues. His presentation provided valuable experience and inspiration for developers facing similar terminal performance challenges.
3. Kuaishou’s Smart UI Production System
Kuaishou’s front-end engineer, Ren Yuehua, discussed the challenges of improving UI design and development collaboration efficiency in modern software development. He shared his insights on building an efficient UI production system from design to code implementation through the combination of large models and collaborative design tools. Ren’s presentation covered the construction of visual稿 delivery platforms, the exploration of AI-generated UI code, and the automation of design-to-code conversion tools. His presentation provided valuable insights into how to improve UI design and front-end development collaboration efficiency in an intelligent context.
4. XiaoHongShu’s Performance Optimization on HarmonyOS
Wang Jinpeng, a HarmonyOS engineer at XiaoHongShu, discussed the performance optimization challenges faced by developers in the context of rising application complexity and the rise of multi-platform development. He shared his insights on how XiaoHongShu, as a deep partner of Huawei’s HarmonyOS, explored and overcame these challenges. Wang detailed the use of unique tools and methods to optimize performance in IO scenarios, cold start, and scrolling scenarios on HarmonyOS. His presentation provided valuable insights into HarmonyOS’s unique performance optimization strategies and tools, as well as how to compare traditional Android/iOS dual-end solutions in a multi-platform scenario.
5. WebNN’s Evolution and Practice in the AIGC Wave
Fu Junwei, a senior engineer in the Intel System Software Department, discussed the challenges of efficiently utilizing local terminal resources to run powerful AI models. He presented on WebNN, a new Web API that provides developers with a bridge to call dedicated AI accelerators through browsers, enabling more efficient AI model inference. Fu’s presentation provided insights into how to run popular AI models like Stable Diffusion, Segment Anything, and Whisper through WebNN. He discussed the architecture design of the WebNN API, demonstrated how to achieve efficient operation of classic Hugging Face models by combining JavaScript ML frameworks with Chrome and Edge browsers, and shared performance comparisons across multiple platforms. His presentation provided valuable insights into how to use WebNN to achieve near-native performance AI model inference, as well as the application scenarios of this technology in browsers.
Conclusion
The QCon Global Software Development Conference provided a valuable opportunity for developers to learn about the latest trends and challenges in front-end development. With the advancement of technology and the rise of new platforms, front-end development continues to evolve and offer exciting opportunities for innovation and growth.
Views: 0