Provde some sort of NullType encoder

[cjnolet]

Thought I'm not advocating for allowing tuples to take on a Null value at any point- I think it would be useful for users to be able to represent that some type is null (and search for it as well). If we were to provide some sort of null type, it would allow nulls in json to propagate to the storage level and to be queried back.

There is an active project which is requesting this feature- which is the reason it's a concern.

[eawagner]

I sort of have concerns with this. Having an object to represent a null value (Null Object Pattern) seems like a hack as we don't really advocate its use in the rest of the code.

I do agree with you though, that we shouldn't try to support raw null values in either tuples or in the type encoders. Things can get messy very quickly.

I would like to investigate an alternative to this that might meet the use case. For instance, is there a reason that this active project couldn't create their own null type and implement their own encoder?

[cjnolet]

I mean that's one possibly but it creates a barrier (however simple it may be) for adoption. The json spec supports nulls as a type. We need to be able to say we take in json and produce the exact same json and as long as there is no way to represent any type of null value, we cannot do that. If we were to use something like NullType to represent this value, it would at least allow us to decouple the type from having a null object state- which would be very bad, I agree.

I'll ponder this for awhile as well to see if I have any other ideas.

[eawagner]

I mean if the intent is to start allowing us to model Json in tuples (ala #131) then I would be more inclined to have a "valueless" tuple implementation.