Problems and goals for interoperability

One of the first initiatives that we worked on under the interoperability panel was an attempt to articulate the fundamental problems that actually needed to be solved to deliver on Solid’s aim to decouple applications from data. We continue to iterate on these as we progress, but I wanted to share them here for anyone interested in the core problems that we’re trying to tackle. We tried to keep these fundamental on purpose, because we could very quickly end up with a list of hundreds of problems and that wouldn’t be a great compass for focusing and prioritizing the panel’s work.

The current version of the full writeup is available here, but here’s a short TLDR; with links to sections in the full text.

Problem #1 - Disparate applications cannot practically interoperate over the same data safely and effectively.

  • Goal - Seamless data interoperability across disparate applications.

Problem #2 - People must be able to understand what data they have to effectively use it or share it safely.

  • Goal - Use intuitive data boundaries to provide human to machine interoperability.

Problem #3 - Disparate entities using different applications must be able to safely and effectively interoperate within their scope of authorization.

  • Goal - Secure collaboration over intuitive data boundaries.

Problem #4 - People shouldn’t need to think about how to physically organize their data to use Solid.

  • Goal - Standard and flexible organizing model for data.

Problem #5 - When data is stored and secure, it should be queryable using any mechanism, and able to be requested in different representations.

  • Goal - Flexibility and support for different and evolving query mechanisms.

Problem #6 - These problems must be solvable using the fundamentals of Solid as it exists today.

  • Goal - Optimize and extend, but do not change Solid fundamentals.

Great points were raised in this thread to include interoperability with other non-Solid stacks, ecosystems, and storage mediums. Looking forward to diving into those some more. I do think its legitimate to include those in the fundamentals. More input, feedback, suggestions welcome!


Love this problem. I can’t help to think that it all comes down to how the SOLID PODs will be hosted (what would the hosting platform/architecture will look like) and who will host it. Will the PODs contain only verifiable data or just about anything related to the users. Conventional methods like replication algorithms (for HA and data consistency) and middleware to manage and route traffic safely might be something to consider in that architecture. Though I am still struggling to see how these data can be queried by businesses for decisions in Low latency/realtime. We might need new analytics meant to query large amounts of small data very quickly (currently our tools are meant for big data). Nevertheless love to follow this conversation/discussion…