Stream: questionnaire
Topic: Extension names
Lloyd McKenzie (Aug 11 2018 at 04:12):
In Grahame's blog post coming out of the Boston session, he had a number of extensions. The names of most of those extensions will need to change because they're scoped by questionnaire and all of those will start with "questionnaire-". As I was going through doing that, I was wondering about changing the name of "form-filler-parameter" to "questionnaire-context". Passing in these parameters won't necessarily only be relevant for form popluation. It could also drive dropdowns for Reference parameters, or possibly enableWhen logic. As well, this sort of behaves like the "context" that gets passed into CDS Hooks or SMART apps. If anyone doesn't like the name change, please holler in the next 24 hours...
Grahame Grieve (Aug 11 2018 at 04:16):
what would the definition be?
Lloyd McKenzie (Aug 11 2018 at 04:25):
Resources that provide context for form completion that are used in form processing logic (pre-population, flow-control, drop-down selection, etc.) when completing a QuestionnaireResponse
Grahame Grieve (Aug 11 2018 at 04:54):
context for form completion - can it be 'processing'? 'support'?
Grahame Grieve (Aug 11 2018 at 04:54):
I'm ok with the name context
Lloyd McKenzie (Aug 11 2018 at 14:35):
Resources that provide context for form processing logic (pre-population, flow-control, drop-down selection, etc.) when creating/displaying/editing a QuestionnaireResponse
Lloyd McKenzie (Aug 11 2018 at 15:04):
I also want to change populated-value to questionnaire-initialExpression. It's taking the place of the initial.value element, so it would be good to keep the name aligned.
Grahame Grieve (Aug 11 2018 at 16:44):
that's not so good for me because if it's attached to score, it's applied each time the score changes, not just choosing the initial value
Lloyd McKenzie (Aug 11 2018 at 17:26):
No, there's a different extension for that
Lloyd McKenzie (Aug 11 2018 at 17:26):
One for initial value (for pre-pop) that can be overridden/changed by the user. And one for calculated that will keep changing as other questions have their values change
Grahame Grieve (Aug 11 2018 at 17:42):
right. but the extension for tieing the value to the item is the same
Lloyd McKenzie (Aug 11 2018 at 17:48):
You had two different extensions in your blog - populate-value and questionnaire-calculated-value
Lloyd McKenzie (Aug 11 2018 at 17:48):
And I think it needs to be two different extensions. The behavior is different
Grahame Grieve (Aug 11 2018 at 17:49):
you have misinterpreted it then. The calculated value is the twin of variable, not populate-value
Lloyd McKenzie (Aug 11 2018 at 17:50):
Why would you need another extension to calculate and not populate?
Lloyd McKenzie (Aug 11 2018 at 17:51):
You need a different extension for initialValue vs. calculateValue for population. The first can be overridden and doesn't change after a one-time determination. The latter continuously changes and will overwrite anything the user changes it to (and thus will usually be marked read-only)
Grahame Grieve (Aug 11 2018 at 17:52):
because you calculate the score at the level where it has access to the sub-parts that affect the score, which is not necessarily the level at which the item lives. We had a big long discussion about that
Grahame Grieve (Aug 11 2018 at 17:52):
there's no question you need a different extension
Lloyd McKenzie (Aug 11 2018 at 17:58):
Right - but the calculation of the score is just another variable
Lloyd McKenzie (Aug 11 2018 at 17:59):
We only need one variable extension - whether it's querying stuff or calculating stuff. But we need two extensions for population.
Lloyd McKenzie (Aug 11 2018 at 18:17):
Ended up changing most of the extension names. And it turns out we never actually approved the Boston proposal in a change request, so I've created one and applied it. Proposed names can be seen here: GF#17665
Grahame Grieve (Aug 11 2018 at 18:26):
I think I'm ok with making the difference on the binding rather than on the variable.
Paul Lynch (Aug 13 2018 at 15:56):
The change to questionnaire-observationLinkPeriod in GF#17665 does not look right to me, or at least I disagree with comment next to it that says, "because it's not creating a link, it's specifying a period". The point of questionnaire-observation-link as I understand it was to create a link to a particular Observation, and the period part was just the means of specifying which Observation.
Eric Haas (Aug 13 2018 at 16:08):
it is just specifying how long the look back period is
Lloyd McKenzie (Aug 13 2018 at 16:14):
The linkage is through the Questionnaire.item.code
Paul Lynch (Aug 13 2018 at 16:38):
Questionnaire.item.code does not specify that there is a link to an Observation. It is the presence of the extension that indicates such a link exists and defines what the link is to.
Lloyd McKenzie (Aug 13 2018 at 16:54):
The extension indicates that the link exists, but it doesn't indicate what the link is to - that's done by the code. The extension just communicates the duration
Paul Lynch (Aug 13 2018 at 17:08):
Well, since there can be multiple observations with the same code, I think the period is just as essential as the code in defining what the link is to. Anyway, to get back to what my original comment said, I think that the comment in the ticket that says "because it's not creating a link" is the part that I most object to. Since you just wrote, "The extension indicates that the link exists", perhaps we can agree that that comment (about it not creating the link) will not show up in the description for the extension, and then I am okay with leaving the name as questionnaire-observationLinkPeriod.
Lloyd McKenzie (Aug 13 2018 at 19:51):
I wasn't planning to include anything that indicated it wasn't creating a link.
Grahame Grieve (Aug 14 2018 at 01:19):
I think paul is right: the presence of the assertion about the link period implicitly defines that there is an operational link rather than this just a definition link
Lloyd McKenzie (Aug 14 2018 at 02:19):
Understand. It does establish the link, but the value is not the link. Neither the old name nor the new one are super-intuitive about what the extension actually does...
Last updated: Apr 12 2022 at 19:14 UTC
