- From: Erik Wilde <erik.wilde@dret.net>
- Date: Sun, 11 Sep 2022 12:30:11 -0700
- To: Roberto Polli <robipolli@gmail.com>, Tommy Pauly <tpauly@apple.com>
- Cc: Mark Nottingham <mnot@mnot.net>, HTTP Working Group <ietf-http-wg@w3.org>
hello roberto.
On 2022-09-09 18:38, Roberto Polli wrote:
> 1. a non human-readable serialization makes SF-Date something
> different from Date;
> 2. if this is clear to everyone, it's ok to me, since Date has a lot
> of machine-processing use cases (e.g. caching).
>
> For new fields, I think that using this data type should not be a
> MUST. I'd like to be free to standardize a new field
> that conveys a date in ISO format (e.g. as a string).
that's possibly a slippery slope. i don't think a MUST is even feasible,
but it would be good to have some clear guidance. it could be something
along of the lines of recommending that:
- the general recommendation for timestamps is to use SF-Date which is
easy to parse and understand for machines.
- if human-readability is a concern, then please use X (RFC 3339?) for
timestamps.
cheers,
dret.
--
Erik Wilde | mailto:erik.wilde@dret.net |
| https://youtube.com/ErikWilde |
Received on Sunday, 11 September 2022 19:30:36 UTC