Cross collaboration with JSON5

[DonaldTsang]

There are others who would like to write a new specification for a new a JSON.
Might be a good idea to discuss in regards to binary encoding. json5/json5#190

[AnyCPU]

@rkalla Did you see that?

[rkalla]

I did - this isn't something I have the capacity to take on though. I'm sorry.

…

On Wed, Jul 24, 2019, 4:06 PM M1xA ***@***.***> wrote: @rkalla <https://github.com/rkalla> Did you see that? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#105?email_source=notifications&email_token=AACG26D26C5XCEVH6D36KGLQBCY37A5CNFSM4HZFOTA2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2XOW6Y#issuecomment-514780027>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AACG26FEJ7Z6RQIIQKZPUI3QBCY37ANCNFSM4HZFOTAQ> .

[DonaldTsang]

@rkalla what do you mean by "take on"? It would be great to talk about some of your opinions, not necessarily contributing code.

[edgar-bonet]

The aim of JSON5 is to be a more human-readable flavor of JSON. UBJSON is binary, quite the opposite of human-readable. These projects are way too far in scope for any collaboration to be meaningful.

[DonaldTsang]

@edgar-bonet I would say that between these should a rise a new meta-JSON standard that covers both binary and "human" JSON in some way (maybe interchangable conversions?).

[DylanYoung]

I think the main thing here would be that ubjson can support the additional constructs in JSON5 (and existing semi- standard json parsers). In particular, comments for round-tripping and debugging. It doesn't look off the top like any other feature of json5 requires additions to ubjson.