FHIR Chat · making canonical urls resolve with simplifier · conformance

Stream: conformance

Topic: making canonical urls resolve with simplifier


view this post on Zulip Simone Heckmann (Aug 24 2016 at 09:00):

Perspectively speaking: If I use canonical urls for all of my profiles, valuesets, namingsystems etc in the form of http://mydomain.de/fhir/<Resourcetype>/<Identifier> and redirect http://mydomain.de/fhir/ to my project's endpoint on Simplifier, then all the urls should resolve, right?

view this post on Zulip Simone Heckmann (Aug 24 2016 at 09:06):

In case of "yes": Is this still going to work, if I want to use subfolders, e.g. http://mydomain.de/fhir/<Resourcetype>/sort-1/<Identifier> and http://mydomain.de/fhir/<Resourcetype>/sort-2/<Identifier>?
How about http://mydomain.de/fhir/sort-1/<Resourcetype>/<Identifier>.
The latter would probably mean I'd have to split my profiles into two Simplifier-Projects and do individual redirects per project?

view this post on Zulip Stefan Lang (Aug 24 2016 at 09:13):

In Apache that would be a RewriteRule or RedirectMatch on the mydomain.de web server.
Either way of translating source to destination is possible (It's basically RegEx).
Apache examples: https://httpd.apache.org/docs/2.4/rewrite/remapping.html

view this post on Zulip Stefan Lang (Aug 24 2016 at 09:20):

Worst case one would have to implement some 20 lines of PHP code to get the redirection correct ;-)

view this post on Zulip Martijn Harthoorn (Aug 25 2016 at 10:46):

Simone, I had a discussion with David Hay in Paris, and his philosophy is that a canonical URL could take any form. It does not have to conform to the FHIR Rest standard. On the other hand, I believe that Gramahe's perspective is that all canonical URLs should be resolvable.

view this post on Zulip Martijn Harthoorn (Aug 25 2016 at 10:48):

We had a lot of discussion here in the Simplifier team about canonical. And the basic problem is that you will probably see a lot of systems, in health, that will want to resolve canonical URL's, but are not allowed to fetch their stuff from 'just anywhere' on the internet. It's more likely that they need an internal server to resolve their profiles.

view this post on Zulip Martijn Harthoorn (Aug 25 2016 at 10:49):

And thus those profiles will never be resolved on their canonical url.

view this post on Zulip Martijn Harthoorn (Aug 25 2016 at 10:49):

except as a value in a parameter.

view this post on Zulip Martijn Harthoorn (Aug 25 2016 at 10:54):

Imagine a patient from a less democratic country visits a doctor, and the software tries to parse the patient resource. And the profile reference (canonical url) is http://hack.ru/leakmyinfo/Patient

view this post on Zulip Martijn Harthoorn (Aug 25 2016 at 10:54):

:)


Last updated: Apr 12 2022 at 19:14 UTC