Trieste LSP implementation

Since Trieste understands scopes and bindings, I wonder if it would be possible to generate simple language servers from Trieste language specifications. The selling point would be that LSP tooling is one of the things you get for free (or at least for cheap) when using Trieste. You could imagine a well-formedness specification that a language can target and that contains the required information for a language server to answer queries. It could also be a separate tool that understands arbitrary Trieste trees and that a Trieste language can hook into.

Most likely we would have to solve incremental parsing to make it performant enough, but a proof of concept to get the idea of the groupnd could be a possible thesis project.

The issue names no files or tests. Start by surveying Trieste's language specifications, tree representation, and scope and binding APIs, then examine how incremental parsing would affect an LSP proof of concept. Done would require a defined approach and working proof of concept, but the issue does not specify acceptance criteria.