Re: [Sedate] Distinguishing "optional extra" vs "you need to understand this to process the date accurately" extensions

[Bron Gondwana <[e-mail]>]

On Fri, Dec 10, 2021, at 06:55, Justin Grant wrote:
> For the core timestamp format (from RFC 3339) and for time zone annotations (which already exist in java.time and similar libraries like Noda Time for .NET), adding a "this is required" character would break backwards compatibility. I'd be somewhat concerned about formats like 2022-12-19T16:39:57!-08:00[!America/Los_Angeles] for that reason.
> 
> For new extensions like calendar without prior art, it seems more reasonable to have an indicator of what's required vs optional. That said, I'd be concerned that senders might not have a clear understanding of what the receiver is going to do with the data-- so the sender may have to guess about whether to include the "!" or not. Or in the case of storing strings in a database, the sender might not know who's going to be using that data in the future, and for what purpose. In that case, the sender won't really know whether to include the "!".

The point is to only have the '!' on things that are required to accurately interpret the time.   I don't think it should be an optional field where you can either have it or not, it should be for things like "relativity offset" where if you're traveling really fast and a long way from Earth then you need to know that in order to accurately align values.  It's not "you need to know hebrew to understand this timepoint" it's "you won't be able to accurately calculate the timepoint unless you understand this.

> Unless we can define a really clear, deterministic way for senders to know whether to include the "!" or not, then I worry that we'd end up with a somewhat-random patchwork of usage where some senders add the "!" and some don't for the same use case.

Not for my initial suggestion - where it's not optional - it's either a datapoint which has it, or not.

Bron.

--
  Bron Gondwana, CEO, Fastmail Pty Ltd
  [e-mail]