- From: Grahame Grieve <grahame@healthintersections.com.au>
- Date: Wed, 26 Sep 2018 11:18:27 +1000
- To: HTTP Working Group <ietf-http-wg@w3.org>
- Message-ID: <CAG47hGZ+CMTqqLf86kjSdsxgUtM1cLYS12TasPLOJNrOieqe0g@mail.gmail.com>
Hi We've run into a problem with the FHIR HTTP API to do with timezones (FHIR = HTTP API for healthcare). TImezones are hard to deal with, and we can't nail everything down (duh!). There's cases where the server needs to know the client's timezone to perform certain tasks correctly on behalf of the client, and the client wants to know the server's timezone to fully understand certain outputs The classic response is for this to be handled by javascript client side, but that doesn't work for a data (json) API. Note also that in principle, servers should not have a timezone, but in practice, every clinical system in the world does have a timezone, and it matters a whole lot. If we care about legacy data, we have to have the client knowing what the server defaults to... and we really care about legacy data. And various clinical protocols have operational ties to timezones, not just offsets (giving medications at the right time really does matter) Has anyone got any comments on this? I like the header described in https://tools.ietf.org/html/draft-sharhalakis-httptz-05 and would specify it for both directions, but that's not blessed by the http-wg? Grahame -- ----- http://www.healthintersections.com.au / grahame@healthintersections.com.au / +61 411 867 065
Received on Wednesday, 26 September 2018 01:19:02 UTC