Stream: committers
Topic: Tasks Needed
Grahame Grieve (Feb 19 2018 at 22:54):
I've been working through the approved tasks - I think I'm top of the approved infrastructure tasks - has anyone got GForge tasks tehy want to take on?
Lloyd McKenzie (Feb 19 2018 at 23:00):
You mean tasks they want you to take on?
Grahame Grieve (Feb 19 2018 at 23:18):
yes
Lloyd McKenzie (Feb 19 2018 at 23:20):
Do you want to do the QA spreadsheet stuff and/or pattern alignment stuff for the FHIR-I resources?
Grahame Grieve (Feb 19 2018 at 23:21):
right now, I'm focused on getting ready for ballot. And particularly, looking for tasks that are blockers
Lloyd McKenzie (Feb 20 2018 at 01:04):
For our normative candidates, by current policy, those things are blockers
Grahame Grieve (Feb 20 2018 at 04:19):
@Lloyd McKenzie ok, pattern alignment, i was looking at this last night.... we need some kind of a table that summarises the pattern mappings. Any ideas what it should look like?
Lloyd McKenzie (Feb 20 2018 at 06:17):
It would be a pretty wide table for Event, but it's possible. To keep it simple and relatively consumable, perhaps just a color-coded square - one color for full alignment, one for slight differences (change of name, constraint of data type, constrained binding), one for supported with an extension and white for no-support. Then fly-overs to convey the details.
Grahame Grieve (Feb 20 2018 at 07:15):
ok. checked the first draft in for you to look at - each of the 3 workflow patterns. I'll work on it some more....
Grahame Grieve (Feb 20 2018 at 07:15):
All the conformance resources are definition resources, right?
Grahame Grieve (Feb 20 2018 at 08:53):
ok, can you review the current tables - I've got the infrastructure in place. Now we can refine them
Grahame Grieve (Feb 20 2018 at 08:53):
both the presentation and the errors - we need a table of allowed data type changes
Grahame Grieve (Feb 20 2018 at 08:54):
btw, while I'm looking at the definition... you can go draft -> withdrawn without going active first
Grahame Grieve (Feb 20 2018 at 08:55):
now, the status codes... all status codes are mapped to the W5 master, so it should be possible to enforce status codes to the patterns, right?
John Moehrke (Feb 20 2018 at 14:08):
How about the dragons? just kidding, clearly not that important GF#13367
John Moehrke (Feb 20 2018 at 14:13):
@Grahame Grieve how about FormatCode codeSystem from IHE, rather than internally published... GF#13950, GF#14003, GF#14004, possibly others.
Lloyd McKenzie (Feb 20 2018 at 17:15):
Not all of the conformance resources are definition resources. Most of them aren't actually. To be a definition, it needs to be describing a type of event that can occur. So for us, that's OperationDefinition and Questionnaire
Lloyd McKenzie (Feb 20 2018 at 17:17):
Right now I'm doing validation of the workflow patterns with transforms and I yell if the w5 mappings don't jive. In terms of allowed data type changes, we haven't gone that far. Doesn't mean we couldn't.
Lloyd McKenzie (Feb 20 2018 at 17:19):
I don't understand why you're making changes to cardinality red, but not splitting to multiple elements, changing name, etc. red. It's totally legitimate to have a cardinality that's looser than or tighter than the pattern in some places. It's not a design-by-constraint thing as created.
Lloyd McKenzie (Feb 20 2018 at 17:20):
Also, it would be good to see what's supported by the extensions and I think fly-overs would make for a more readable table.
Grahame Grieve (Feb 20 2018 at 19:29):
why is it legitimate to have a looser cardinality?
Lloyd McKenzie (Feb 20 2018 at 20:05):
Because that's how the resource behaves in that business domain.
Lloyd McKenzie (Feb 20 2018 at 20:05):
The pattern reflects "typical" not "broadest set of what's possible"
Grahame Grieve (Feb 20 2018 at 20:06):
no that doesn't explain it to me. why would the pattern say 0..1 if there's a domain that says 0..*?
Lloyd McKenzie (Feb 20 2018 at 20:07):
If Event says subject is 0..1, a resource where it's core to have multiple subjects can have 0..*. Most resources won't have more than one.
Lloyd McKenzie (Feb 20 2018 at 20:07):
Thus the pattern says 0..1
Grahame Grieve (Feb 20 2018 at 20:11):
so in other words, in your eyes, the pattern cardinality is best practice, not worst case
Lloyd McKenzie (Feb 20 2018 at 20:13):
Not even necessarily asserting "best" practice. More "common" practice, though we've perhaps edged a little bit towards "best". It's certainly not worst case.
Grahame Grieve (Feb 20 2018 at 20:15):
I think that for some of the uses we intend to make of pattern, we need both cardinalities
Lloyd McKenzie (Feb 20 2018 at 20:18):
That would actually be two patterns. The "must fit within" pattern and the "common practice" pattern. The latter could be derived from the former. Mappings to the latter would allow you to figure out the mappings to the former.
Grahame Grieve (Feb 20 2018 at 20:20):
it it's just cardinality, we don't need 2 patterns. I've sometimes thought that it would be good to get 2 sets of cardinalities in the base resources, but the problem is 'common practice' cardinalities are dependent on which practice you're thinking of. I think that will be a problem with the patterns as well
Lloyd McKenzie (Feb 20 2018 at 20:20):
Cardinalities and data types I would expect.
Grahame Grieve (Feb 20 2018 at 22:48):
"Not all of the conformance resources are definition resources. Most of them aren't actually. To be a definition, it needs to be describing a type of event that can occur" - that's not the current definition of Definition
Grahame Grieve (Feb 20 2018 at 22:48):
and if it is the right definition, why isn't it called "ActionableDefinition" or something?
Grahame Grieve (Feb 21 2018 at 03:18):
and @Lloyd McKenzie check the updates to the pattern mappings
Lloyd McKenzie (Feb 21 2018 at 04:08):
They're workflow patterns - Definition, Request and Event. In that context it's clear. If you think they need a different name, you can supbmit a change proposal.
Lloyd McKenzie (Feb 21 2018 at 04:08):
Like the table much better now. We need a legend that explains N, T, C and E
Lloyd McKenzie (Feb 21 2018 at 04:09):
Is it possible to put <br/> in a flyover? That might improve readability a bit.
Grahame Grieve (Feb 21 2018 at 04:10):
Grahame Grieve (Feb 21 2018 at 04:10):
in other words... no
Grahame Grieve (Feb 21 2018 at 04:10):
need javascript... I'm undermotivated
Lloyd McKenzie (Feb 21 2018 at 04:18):
I'm not ultra-concerned with it being a little less pretty in Firefox, but it's not critical in any case.
Grahame Grieve (Feb 22 2018 at 06:10):
want to comment on the definition issues above?
Lloyd McKenzie (Feb 22 2018 at 06:41):
Did a quick skim of "above" and am not sure which issues you're referring to. Only one I noticed not addressed was the state machine issue - and that will need a change proposal.
Grahame Grieve (Feb 22 2018 at 06:51):
the definition of Definition is different in the spec to the one you give here, and that makes a difference as to which of the conformance and terminology resources are definition resources
Grahame Grieve (Feb 22 2018 at 09:03):
@John Moehrke looking at task GF#13950, the resource found at ftp://ftp.ihe.net/TF_Implementation_Material/fhir/CodeSystem/IHE.formatcode.codesystem.xml has a canonical URL of http://ihe.net/fhir/ValueSet/IHE.FormatCode.codesystem. I think it should be http://ihe.net/fhir/CodeSystem/IHE.FormatCode.codesystem
Grahame Grieve (Feb 22 2018 at 09:06):
also, "Note, that the HL7 defined format codes are today used as if they are assigned by IHE. How should that be managed? They really should be HL7 defined." - ah yes, but who owns that? SDC? where did the list come from?
Grahame Grieve (Feb 22 2018 at 09:20):
also, the IHE code system has errors in it - there are codes with a space in them
Grahame Grieve (Feb 22 2018 at 09:31):
also, it has the same OID as the value set
Grahame Grieve (Feb 22 2018 at 10:42):
ah that's another task. ok.
Grahame Grieve (Feb 22 2018 at 10:42):
how does this one get resolved?
Grahame Grieve (Feb 22 2018 at 10:42):
John Moehrke (Feb 22 2018 at 13:49):
The main XDS-I codes are in the IHE codesystem. So pulling the IHE codeSystem will get them.
John Moehrke (Feb 22 2018 at 13:49):
I fixed the URI.. sorry
John Moehrke (Feb 22 2018 at 13:50):
The ValueSet is wrongly using the OID for the CodeSystem... so in IHE I created a new OID for the ValueSet. If you pull the IHE ValueSet you will get that fix
John Moehrke (Feb 22 2018 at 13:51):
ah, spaces... dang.. that might take me a bit to fix. I am new at this codeSystem stuff
John Moehrke (Feb 22 2018 at 13:52):
the problem with the HL7 defined formatCodes is likely going to be a forever problem.. because they are used today that way. This is an accident, mostly because many saw FormatCodes as URN, and therefore never expected to need a code system identifier.. had we been smart back then, we would have used the RFC solution; but we were not smart, so have an OID to live with forever.
John Moehrke (Feb 22 2018 at 13:57):
on the spaces in codes @Grahame Grieve where? I see all my <code> elements don't have spaces.
Grahame Grieve (Feb 22 2018 at 18:01):
search for ": 2010"- there's 3 at the bottom. Or diff with this:
Grahame Grieve (Feb 22 2018 at 18:01):
Grahame Grieve (Feb 22 2018 at 18:02):
HL7 format codes... but who owns them?
Anthony(Tony) Julian (Feb 22 2018 at 18:46):
Question: GF#13845:The link to "CapabilityStatement" needs to be in Camel Case it is currently written as "capabilitystatement" (http://build.fhir.org/messagedefinition.html) 2.51.1
It looks like this link is generated by code - it certainly is not in messagedefinition.xml
Last updated: Apr 12 2022 at 19:14 UTC
