Stream: IG creation
Topic: IG valueset expansion
John Moehrke (Nov 13 2019 at 15:10):
Did I do something wrong, or should I just expect a valueset in an IG to not be expanded?
No Expansion for this valueset (not supported by Publication Tooling)
see http://build.fhir.org/ig/IHE/ITI.MHD/branches/master/valueset-IHE.formatcode.vs.html
David Pyke (Nov 13 2019 at 17:14):
There may be an issue with it trying to expand that set that is referenced by OID.
Grahame Grieve (Nov 13 2019 at 20:27):
not so much that it is referenced by an OID, but that it is not defined anywhere. The Publication tooling is not magic. If it's meant to refer to http://build.fhir.org/ig/IHE/ITI.MHD/branches/master/IHE.formatcode.cs.xml.html, then you need to use the correct canonical URL
John Moehrke (Nov 14 2019 at 02:54):
Wait... the Publication Tooling is not magic????? This can't be.. it sure seems like magic to me... BUT, I am refering to the CDA code-set given the URL that I find it is defined as...
https://terminology.hl7.org/builds/current/site/CodeSystem-hl7-document-format-codes.html
John Moehrke (Nov 14 2019 at 02:55):
here is a more simple version... I have added this valueset to the FormatCode IG. http://build.fhir.org/ig/IHE/FormatCode/branches/master/valueset-IHE.formatcode.vs.html
Grahame Grieve (Nov 14 2019 at 03:02):
Codesystem:
Defining URL: http://terminology.hl7.org/CodeSystem/hl7-document-format-codes
Grahame Grieve (Nov 14 2019 at 03:03):
in the value set:
Grahame Grieve (Nov 14 2019 at 03:03):
Include all codes defined in urn:oid:1.3.6.1.4.1.19376.1.2.3
John Moehrke (Nov 14 2019 at 03:46):
<compose>
<include>
<system value="urn:oid:1.3.6.1.4.1.19376.1.2.3"/>
</include>
<include>
<system value="http://terminology.hl7.org/CodeSystem/hl7-document-format-codes"/>
</include>
</compose>
John Moehrke (Nov 14 2019 at 03:47):
https://github.com/IHE/FormatCode
Grahame Grieve (Nov 14 2019 at 04:00):
yes. well, what is urn:oid:1.3.6.1.4.1.19376.1.2.3 ?
John Moehrke (Nov 14 2019 at 04:02):
that is the IHE codesystem for FormatCodes... it is in this IG... Do you want me to try the URL form?
Lloyd McKenzie (Nov 14 2019 at 04:06):
I'd make a strong recommendation to the IHE to not use OIDs as the canonical URLs for FHIR artifacts. Yes, there's a cost to translating OIDs to human-readable URLs, but the infrastructure has to exist for all conformant systems regardless and the incremental cost is therefore low. The benefit of moving to human-readable URLs is significant. If you don't do it initially, it's going to be much harder to do it in the future (because the infrastructure exists for v3/CDA <-> FHIR conversion, not for FHIR <-> FHIR conversion.
John Moehrke (Nov 14 2019 at 04:08):
well that did not fix it
John Moehrke (Nov 14 2019 at 04:08):
OIDS are used more universally across many protocols. I understand the FHIR preference.. but it is not a universal preference
John Moehrke (Nov 14 2019 at 04:09):
and this codesystem and valueset have been known by these OIDS for 15 years
Lloyd McKenzie (Nov 14 2019 at 04:12):
I understand that it's known by those OIDs in other protocols for many years. But there are new implementers all the time, and all of them will struggle to recognize and differentiate the OIDs. Whereas those who have been using the OIDs for all 15 of those years will still understand the human-readable URL. (That's the benefit of them being human-readable.) "We've been incurring the learning curve and maintenance cost of OIDs for 15 years" isn't a strong argument for continuing to incur that cost for the next 15+
John Moehrke (Nov 14 2019 at 04:12):
note also that this OID is how the FHIR valueset is created today. So this is not the new part. The new part is that I am bringing in the HL7 defined codes to this valuset
John Moehrke (Nov 14 2019 at 04:13):
so the thing that is not working.. is the new HL7 codesystem
John Moehrke (Nov 14 2019 at 04:13):
I appreciate your distaste for OID.. but that is not the problem
Lloyd McKenzie (Nov 14 2019 at 04:14):
I realize this is tangential to the discussion of what's not working. And it's unlikely (though potentially possible) that the issue is caused by it being an OID. What I'm doing is passing a message to you that I hope you can take back to IHE. Retaining OIDs in IHE's FHIR specs is ill-advised.
John Moehrke (Nov 14 2019 at 04:14):
I still expect the problem is something I am doing wrong. just not clear what that is yet
Grahame Grieve (Nov 14 2019 at 08:31):
ValueSet.compose.include.system is the canonical URL for the Code System to include codes from. How have you provided the code system with the canonical URL in quetsion?
John Moehrke (Nov 14 2019 at 13:10):
In the case of the IHE one, yes it is in the same IG... in the case of the HL7 one I expected that to come from the terminology server.
Grahame Grieve (Nov 14 2019 at 13:12):
but it's not in the same IG. That's where the problem is
John Moehrke (Nov 14 2019 at 13:22):
so... I removed the hl7 one, and it still doesn't work.. as I indicated, I think there is something more simple. Can you look at the vs source?
Grahame Grieve (Nov 14 2019 at 13:25):
I've already told you - you're referring to something you haven't defined:
Include all codes defined in urn:oid:1.3.6.1.4.1.19376.1.2.3
urn:oid:1.3.6.1.4.1.19376.1.2.3 is not defined
Lloyd McKenzie (Nov 14 2019 at 13:31):
The code system needs to be defined in the core spec, in the implementation guide with the value set or in an implementation guide that's an explicit dependency of the IG defining the value set.
John Moehrke (Nov 14 2019 at 13:31):
it is the identifier in the codeSystem...
I have just switched it to the url in the codeSystem
Lloyd McKenzie (Nov 14 2019 at 13:32):
Implementation guides can't reference any resource (value set, extension, etc.) that isn't defined in something they have an explicit dependency on.
John Moehrke (Nov 14 2019 at 13:32):
Lloyd, it is in this IG. that is all that this IG is about is allowing IHE to publish this codeSystem for consumption by the FHIR core spec
John Moehrke (Nov 14 2019 at 13:33):
okay. so NOW the valueset with just the IHE codesystem that is in the IG with it is expanding... How do I add the HL7 codes?
John Moehrke (Nov 14 2019 at 13:35):
this set of codes https://terminology.hl7.org/builds/current/site/CodeSystem-hl7-document-format-codes.html
Grahame Grieve (Nov 14 2019 at 13:38):
that bit was fine - just put it back in and it should be ok
John Moehrke (Nov 14 2019 at 13:51):
I did try that when you recommended it... I just put it back again.. nope, doesn't work
Last updated: Apr 12 2022 at 19:14 UTC
