Stream: implementers
Topic: Near location units
Lee Surprenant (May 15 2020 at 21:24):
This came up on PDEX PlanNet track and I was supposed to raise it here, but forgot to until our readout:
currently the base spec is pretty vague on the supported units for the Location.near search parameter.
/cc @Saul Kravitz
Lee Surprenant (May 15 2020 at 21:25):
Basically it says you need to support "km" but without much more guidance on other units
Lee Surprenant (May 15 2020 at 21:25):
our server was expecting UCUM units but the client from MITRE was using "mi" for miles
Lee Surprenant (May 15 2020 at 21:25):
what do the other servers implement for location search units?
Lee Surprenant (May 15 2020 at 21:26):
should it be further clarified in the base spec?
Brian Postlethwaite (May 16 2020 at 05:02):
Those are likely the only 2 real units in use. I dont recall what mine supports, or if I did the ucim conversion using the ucum conversion libs.
Lee Surprenant (May 18 2020 at 13:30):
I tend to agree. Should it state how a client and server can specify miles? mi, miles, or [mi_us]?
Lee Surprenant (May 18 2020 at 14:12):
I opened FHIR#27157 and left my suggested wording:
Servers which support the near parameter SHALL support
kmfor kilometers and SHOULD support bothmi, and[mi_us]for miles; support for other units is optional.
Paul Church (May 18 2020 at 14:21):
Haven't implemented this yet but some standardization would be great. Suggested wording sounds fine to me.
David Pyke (May 18 2020 at 14:23):
I think they should only support smoots. It's the only truly universal measurement https://en.wikipedia.org/wiki/Smoot
Saul Kravitz (May 18 2020 at 15:21):
Lee -- I updated the plan-net RI client to use km as units on the near query, as we discussed, to improve likelihood of portable operation.
Lee Surprenant (May 18 2020 at 18:32):
@David Pyke I was thinking to also support "mil" but make that be https://en.wikipedia.org/wiki/Thousandth_of_an_inch ;-)
Last updated: Apr 12 2022 at 19:14 UTC
