As mentioned in #103, it might be interesting to be able to persist / serialize the IndexedSet to be able to cache expensive indexing. The mentioned scenario would be in Blazor WASM applications, where you could precalculate everything on the server.
Risks to address
- Scope Creep / Time for maintenance
- Serialization: The library allows anything to be a key but it should not be a fully fledged serializer
Requirement
- Deserialization needs to be much faster than recreation (i.e. serialization of all elements and creating a new set with the deserialized elements)
Alternatives to check
- Speed up trie building (fulltext / prefix indices) for batch insertions
As mentioned in #103, it might be interesting to be able to persist / serialize the IndexedSet to be able to cache expensive indexing. The mentioned scenario would be in Blazor WASM applications, where you could precalculate everything on the server.
Risks to address
Requirement
Alternatives to check