As the remote work is gaining popularity, so are the applications allowing multiple people to collaborate together in responsive manner. While some well-known tools like Google Docs have kept their status, we're would like to embed similar features into our own products, which would also allow us to work offline, in remote areas, on the data that we own on our personal computers and have full control over.

During this session we'll cover how to make these promises a reality, understand their inner workings and what challenges can we expect. We'll also gain some intuition to consciously deal with trade-offs of local-first software and P2P systems, compared to more traditional service-oriented architectures.


Focus on our talk will be on how to design and implement algorithms that make a same-document collaboration possible, even when we accept that users may work offline for prolonged periods of time. For that we'll go over challenges, tradeoffs and optimizations that make our approach scale in real-life scenarios.

As the point of reference, we'll use Yjs/Yrs, which are one of the most widely used libraries for building collaborative applications on that field, used currently in products such as Evernote and Jupyter Notebooks.