Stream: committers
Topic: Maintaining/enhancing IG templates
Lloyd McKenzie (Feb 18 2021 at 20:37):
Carrying over discussion from here: https://chat.fhir.org/#narrow/stream/179294-committers.2Fannounce/topic/Helping.20maintain.2Fenhance.20IG.20templates
Lloyd McKenzie (Feb 18 2021 at 20:38):
I was remiss in publishing the recording from last week. Content is here:
https://hl7-org.zoom.us/rec/share/j4QQEJCDuaFnrsxPnyMpojM7PHMA-xL7SzJ4gqB6eurILMW4nra-HYzej6bWuw9X.3CxsMBGPPuxchTbb
Access Passcode: &WCp4YMr
Jose Costa Teixeira (Feb 18 2021 at 20:39):
is there a call today?
Lloyd McKenzie (Feb 18 2021 at 20:40):
I'm listening to the recording to figure out what we decided :) Give me another 60 seconds...
Lloyd McKenzie (Feb 18 2021 at 20:42):
Agreement was to have a call 2 weeks past the 11th, so that'll mean the 25th.
Jose Costa Teixeira (Feb 23 2021 at 17:41):
What did we decide on the branching? I want to make some changes, what I presume is:
- we make a branch on the base-template, with a different version number in the package, say
0.2.1.mine - we make a branch on the sample / guidance ig, pointing at
fhir.base.template#0.2.1.mine
Jose Costa Teixeira (Feb 23 2021 at 17:41):
is that it?
Jose Costa Teixeira (Feb 23 2021 at 17:42):
I want to test that with the BPMN stuff
Lloyd McKenzie (Feb 23 2021 at 20:24):
I think that's right, y
Jose Costa Teixeira (Mar 03 2021 at 23:50):
do we have a call tomorrow?
Lloyd McKenzie (Mar 04 2021 at 00:26):
Jose was the only one who came last week. We got some stuff set up in Git, but no other changes. I'll admit that my preference is to wait until next week to talk again.
Jose Costa Teixeira (Mar 04 2021 at 07:46):
I agree, not having / seeing any pressing topics*
Jose Costa Teixeira (Mar 04 2021 at 07:46):
* (besides the 100000000 improvements we can do, but we can do those anyway, anytime, no call needed)
John Moehrke (Mar 04 2021 at 13:07):
yes, please skip this week. sorry I missed last week, conflict with one of the day-jobs.
Lloyd McKenzie (Mar 04 2021 at 14:24):
Ok, with 3 votes against and no votes for, we will not hold our IG template call today. Note that one of the things we were going to do pos our call 3 weeks ago was exercise the proposed process a bit, so feel free to do exactly that prior to next week's call. :)
Lloyd McKenzie (Mar 11 2021 at 21:16):
Reminder - call today at 6 Eastern
Lloyd McKenzie (Mar 11 2021 at 23:01):
@Eric Haas @Sean McIlvenna @John Moehrke ??
Eric Haas (Mar 11 2021 at 23:10):
Zoom line?
Lloyd McKenzie (Mar 11 2021 at 23:11):
Regular FHIR-I link
Lloyd McKenzie (Mar 11 2021 at 23:11):
(Monday call or SDC or whatever)
Jose Costa Teixeira (Apr 07 2021 at 15:31):
Perhaps for the call tomorrow (or please tell me if I am missing something):
The rendering of extensions for the core is much neater than the rendering of extensions for the IG publisher. Is it possible to use the same kind of rendering?
for example, https://www.hl7.org/fhir/extension-cqf-relativedatetime.html looks better than just the slicing rendering
Lloyd McKenzie (Apr 07 2021 at 15:49):
Looks like both views are provided in core. Have you checked to see whether the 'summary' fragment gets generated by IGs? If it is, we should be able to make it show up in the templates.
Jose Costa Teixeira (Apr 07 2021 at 16:22):
It seems the summary fragment just contains the boring text-based super short summary:
<p>Fixed Value: 1 element<br/> Prohibited: 1 element</p><p><b>Slices</b></p>
<p>This structure defines the following <a href="http://hl7.org/fhir/R4/profiling.html#slices">Slices</a>:</p>
<ul>
<li>The element Extension.value[x] is sliced based on the value of type:$this</li>
</ul>
Jose Costa Teixeira (Apr 07 2021 at 16:22):
oh wait there's more stuff
Jose Costa Teixeira (Apr 07 2021 at 17:09):
no, the closest I have in the fragments is the profile-like rendering
image.png
Lloyd McKenzie (Apr 07 2021 at 20:22):
@Grahame Grieve - is this something that we can expose in the IG publisher too?
Grahame Grieve (Apr 12 2021 at 23:06):
I could
Jose Costa Teixeira (May 06 2021 at 17:23):
Which one of these summaries do we want on profiles?
Jose Costa Teixeira (May 06 2021 at 17:23):
Lloyd McKenzie (May 06 2021 at 17:33):
The first, though we might need to turn the table into 2 rows to convey everything needed
Jose Costa Teixeira (May 06 2021 at 17:54):
perhaps then i'll add something to it:
image.png
Jose Costa Teixeira (May 06 2021 at 17:56):
the only things taht are missing are the name, the version and the copyright. The url is lower in the page, so perhaps we can put all of them in a single table?
I presume that the table would be done by the publisher?
Lloyd McKenzie (May 06 2021 at 19:22):
Could be, though I think all of that information can/should be surfaced in a way that would allow the table to be constructed by the templates - and that might be better to allow for customization (e.g. if some organizations don't want to surface standards level or FMM)
Title should already be in the page heading, so no need for it to be surfaced. I'm skeptical that the 'computable' name should show up in the human-readable view, though I suppose if it didn't, no one would review. If we expose it, should definitely be labeled "Computable name". Moving the canonical URL up makes sense - and could express the version as part of that.
Lloyd McKenzie (May 06 2021 at 19:25):
Copyright, I lean towards not having in the table, but instead having as a text section below the definition/description. I think you'd want to read the details of what the resource is for before you read about the rights constraints on it - also, copyright/legal info could sometimes be several paragraphs and sticking that in the table is going to be ugly.
Elliot Silver (May 06 2021 at 19:52):
Lloyd McKenzie said:
I think all of that information can/should be surfaced in a way that would allow the table to be constructed by the templates
I've had this thought about IGs in general. How much of the content generated by the publisher could actually be moved into the templates? On the one hand it might improve maintainability by allowing others than Grahame to fix issues, but it might lead to more divergence in IG presentation/content. I'm sure this was debated thoroughly before I started paying attention to IG development, but it was something I've wished for a couple times in the last year.
Lloyd McKenzie (May 06 2021 at 21:02):
Well, things like the table views I think would be pretty horrendous to generate in templates. Most of the resource views, actually...
Grahame Grieve (May 06 2021 at 21:45):
the code is open source; anyone can make PRs around the views. Doing it in a harder to execute language isn't going to make it any easier to do that
Copyright must be displayed. I know some of you don't like the tables, but they do make specified pieces of information easy to find across all the pages
Lloyd McKenzie (May 06 2021 at 22:48):
The templates will display copyright - but it's as important to display information in an attractive easy-to-consume format as it is to display it - and the current tables definitely aren't that.
Jose Costa Teixeira (May 06 2021 at 23:07):
All the code is in the publisher, which implies that everyone gets the same flavour of rendering.
I guess the reason to putting it in templates would be to allow local variations - do we want to prevent that?
Grahame Grieve (May 06 2021 at 23:23):
there's a lot of logic
Grahame Grieve (May 06 2021 at 23:23):
what would make it easier to consume? Cause I don't feel it
Lloyd McKenzie (May 06 2021 at 23:26):
There isn't really any logic for the intro table stuff. There's a ton of logic for some of the other views.
Grahame Grieve (May 06 2021 at 23:45):
There's still quite a bit of logic. So what's the issue? the table? or the things that are in the rows?
Grahame Grieve (May 07 2021 at 00:25):
but how I about we do this:
Elliot Silver (May 07 2021 at 01:56):
My comment about moving more into the templates was just a wild idea. I don't want to distract the discussion from what Jose and Lloyd are trying to solve.
Grahame Grieve (May 07 2021 at 02:10):
config.json will have a new optional entry "summary-rows" which is a comma delimited list of names from the list below. If none of the templates specify rows, then all the rows will appear, otherwise on the specified rows will appear (based on what the last template to mention summary-rows says).
The possible names are:
- url
- version
- name
- title
- status
- definition
- publisher
- committee
- copyright
- maturity
- formats
- content
- oid
- cs.vs
Lloyd McKenzie (May 07 2021 at 02:34):
I don't want the summary view to appear at all. All of those elements should be available in the resources.json data file. The template should render them appropriately. I'm not saying you can't spit out the table if you wish, but I don't want the official template to use it.
Lloyd McKenzie (May 07 2021 at 02:35):
Some of those elements shouldn't appear in table form. Some shouldn't appear at all. And some should appear in a much more concise layout than the 2-column name+label format the table adopts.
Grahame Grieve (May 07 2021 at 05:54):
well, I'm skeptical. I think that stuff will just get lost, and people won't be able to find it when they look for it
Lloyd McKenzie (May 07 2021 at 13:09):
Title is already in the page heading. url, version, name, status, maturity, publisher (and contact link for publisher) should show up in a 'concise' table. committee and publisher should be one and the same. definition and copyright should show up as specific text sections - the latter should have a sub-header. Formats should be tabs. I guess OID can show up in the table, though it should be grey with a fly-over saying "for mapping purposes only". content and cs.vs should be part of the rendering of the value set, not metadata shown at the top.
Last updated: Apr 12 2022 at 19:14 UTC
