Hi,
I'm Omar from the London Java Community. We've just completed our early review of this JSR and we think it's great but had following suggestions.
I was expecting this JSR also to focus on autogenerating classes from JSON schema similar to JAXB using XML schema to generate classes that can be serialised and deserialised. That can be very useful for efficient data validation even prior to deserialisation.
The JSR only seems to focuses on serialisation and deserialisation. In that perspective, I think it is fairly comprehensive. Generics is a challenging area because of erasure as mentioned in the JSR. While comprehensive measures to infer types from deserialised JSON documents have been specified, it may be possible to use annotations to guide this inference.
I shall be grateful for advice and suggestions on the above.
Thank you and best wishes
Omar