In the Christmas time till now (22 January) I have been working on a product that has many of the technical ideas that I have been thinking of.
I have given a demo to Jamie and Simon. It is an RDF based system that is < 100kb with offline support and it has a way of scaling if needed to a SPARQL backed database. It is not finished and operatable and I have no intentions of letting it be used by LightNet as this is just not realistic and it was not the reason why I made it.
I have made it to prove my ideas. I would like to open source this. It has no reference to LightNet in code. All code is clean and newly written.
It has a couple concepts which might be interesting for LightNet:
- Offline is done via the Cache API and not IndexedDB. This lowers complexity. When the app ‘boots’ in offline mode it loads an in memory store and fetches all the cached documents. Querying always goes over the store, in offline mode the store is in front-end, in online mode the store is in the back-end. When the data would be too much for an in memory store I intend to make it possible to hook in a SPARQL based database.
- Preact is excellent and I can recommend to test it out if it is compatible.