Proposal: graphql for retrieving, updating turtle documents

I know this might be a bit controversial (it is maybe seen as redundant burden) but it has some advantages:

1. search

Currently implementing a search would require to search all documents for the specific values to conform the sparql requirements. Implementing a search in graphql is very easy:
you tag every document with some metadata and can search through the metadata

2. custom actions (mutations)

The sparql Syntax allows just a limited set of operations. In graphql you can add Operations (mutations).
This might be handy for creating actions like "encrypt content for pod with public keys specified in pod" or "ping owner of pod".

3. privacy

Sparql searches are very expressive. This way it is harder to track which operations should be restricted and how to not leak metainformations. In graphql you can customize the search like in rest.

• a solution would be to prefilter the query and only allow rdf operations on the filtered query

4. reuse libraries, simpler implementation

There are already some libraries for graphql which can be easily adapted for own needs. Also you can define which APIs you expose and how. This makes authentication easier and also allows triggers for events like: "document was queried".
Also it allows an bottom-up approach which is helpful for creating prototypes.

5. documentation

Graphql is self-documenting. You can have an autogenerated documentation.

Nomenclature

• Document: a turtle file
• SPARQL: typically used query language for rdf documents.
• GraphQL: a query language created by facebook

Opinions?

Is there a way to extend sparql for supporting custom actions?

• Yes there is a way

まず、リポジトリに既存する SPARQL API と、Turtle ドキュメントの取得・更新方法を確認します。要求されている GraphQL のクエリとミューテーションを、現在の API サーフェス、認証、プライバシー制御と比較します。完了とするには、GraphQL サポートの合意済みのスコープと実装計画が必要ですが、この issue ではファイル、テスト、具体的な最初の変更は示されていません。