For most of the history of the internet, the predominant computation and user interaction model was to house nearly all the computation, data storage, and other functionality on centralized servers, with client devices serving effectively as simple terminals. Then, as frontend frameworks started to flourish, progressively more data handling and storage began to shift to the frontend. The recent emergence of full-stack frameworks like Next.js and Remix has further blurred the backend/frontend distinction.
Now, however, the development of WebAssembly (WASM) and related technologies has opened the door to yet another paradigm shift in how data is managed and processed. These tools allow client-side code to be written in many languages other than Javascript, but still run directly in the browser. This capability has prompted exploration of a wide variety of client-side processing and computation models for tasks that historically have been handled strictly on the server.
In this Open Source Architect Spotlight event, Roman Yurchak delves into some of these recent tools that are available to enable client-side data storage and processing. He also describes some of the implications for data privacy when these functions are delegated to client devices instead of being handled on centralized servers. Roman is an ML engineer at Symerio and a maintainer of the Pyodide package, and shares from his deep expertise with these tools, their applications, and how they interact with data privacy concerns.
Pyodide: https://pyodide.org/en/stable/
Symerio: https://www.symerio.com/
Roman @ LinkedIn: https://www.linkedin.com/in/ryurchak
Roman @ GitHub: https://github.com/rth
The UK project mentioned in the conversation that is working to provide better children’s services is Social Finance UK.
Blog post: https://www.socialfinance.org.uk/projects/providing-better-childrens-services-data-at-lower-cost
PyData London talk by Tambe Tabitha: https://www.youtube.com/watch?v=g3vX4zKMmrs
October 4, 2023
