Gordon Mickel just posted an interesting, fun and interactive quiz game to learn openEHR concepts. Test your knowledge, earn badges, and become an OpenEHR integration master!
Game: OpenEHR Integration Quest
GitHub: GitHub - gmickel/openehr-quest: A fun and interactive quiz game to learn OpenEHR concepts. Test your knowledge, earn badges, and become an OpenEHR integration master!
Oh i really liked this one 
reminds me of a similar one made for fhir
i liked that the mention of smoking archetypes reflects the points from the presentation made by @heather.leslie at the openEHR day in Arnhem.
In the last question, the AQL is not correct. I will contact the owner.
I was surprised about that one. In general openEHR CKM has a maximal modelling strategy. At this level this question is confusing to me.
I also donât remember the message from @heather.leslie in Arnhem. And reviewing her ppt didnât help.
Iâm curious for some elaboration.
I got that one wrong too?
The way i understood the question is that: the aim of the archetype is always to get the max dataset but in specific cases such as the case of smoking, you have to divide it in different archetypes - you can have an archetype for consumption summary other for use, etc. But if theres a better explanation or if i understood it wrong, its always possible to open a issue in github and ask for clarification. I remember seeing a similar pattern in the alcohol case.
Ah I see. I think it will be too confusing for the audience. First one to open an issue gets a beer in reading?
I already opened one yesterday for the AQL, and was fixed very quickly by the developer.
So I guess I am counting the days for the 
Nice quiz-initiative! I had to click XML to see if the quiz premitted both valid answers.
The quiz should be updated to not exclude any of the allowed representation formats (or to not list XML in the multiple choise). Read the full chapter under âData representationâ in Overview openEHR specs
We sometimes find the canonical XML format to be at least as useful as the canonical JSON format for some integration use cases, especially when the source system or some transforming part uses XML and you want to have e.g. an XQuery based transformation. Ask the Veratech people for example, right @yampeku & Co?
Yep, tried the same thing 
i remember that i have also sent it in xml format before on the server from better, that question could have 2 valid answers.
Thank you for the quiz, just did it with my colleague and it was so much fun!
Hi everyone!
I want to start by thanking @borut.jures for sharing this quiz and @vanessap for the valuable feedback. Itâs great to see it being used here!
A bit of background: I originally created this quick quiz for my co-workers before giving a presentation on OpenEHR for developers. Itâs wonderful to see it reaching a wider audience.
Regarding question 7, youâre absolutely right. Iâll be working on allowing multiple answers to make it more accurate.
In other news, Iâve taken openehr-quest and developed it into a generic quiz maker called CodeQuest. If youâre interested in creating your own quizzes, feel free to check it out here: GitHub - gmickel/CodeQuest: An interactive quiz game framework to build your own quests!
Thanks again for your interest. Itâs great to be part of such an engaged community!
Expert mode: Getting the complex AQL questions right 
God mode: Pointing out errors in the complex AQL questions 
Thanks @gmickel for a fun challenge! 
Well, you donât have to, but trying to cram too much stuff into a single archetype will create a lot of avoidable headaches for both modellers, implementers and information consumers.
I for one agree completely with the answer given in the quiz: We asymptotically approach a maximal data set for each concept. But all aspects of smoking spans several distinct concepts.
I got the answer âwrongâ!
I donât use the word âmaximalâ any more to describe what we do, rather aiming for âinclusiveâ.
Further complicated by the approach to achieving âinclusiveâ is not always as simple as being limited to only one archetype per concept.
And of course, the final solution requires consideration on how to best represent the whole concept within a reusable ecosystem of archetypes.
Time for a new blog post Heather? I think the question of how to âselectâ data elements for an archetype is very relevant. Given the FHIR 80/20 rule and the increasing collab. Iâve always liked the simplicity of âmaximalâ to me defined as âanything some clinician wants to record about a specific conceptâ. But I now see thatâs simplistic. At Nictiz the idea came up to make a âmind mapâ (but less hierarchical) of all clinical data elements and their relations and selecting an area of related elements for a specific concept. They called it âbetekenismodelâ (~semantics model). I hated the idea because it was very loosely defined yet made into a silver bullet. And because itâll never be finished and you have nowhere to start doing implementable work. But given your and Siljeâs described experience it may be more helpful then I thought. What do you think?
(Maybe letâs split this in a new topic)
I agree, I also prefer âinclusiveâ to âmaximalâ, I just forgot what the word was. I blame four weeks of not thinking about work 
Also very true!
I use the term âinclusiveâ to challenge the misconception that achieving 100% representation is not feasible or not necessary; rather it is setting the expectation that stakeholders who identify reasonable and semantically sensible requirements should be able to have them represented. Conversely, we donât set a nominal target like an 80/20 split, since determining when 80% is achieved is equally as challenging and unrealistic as deciding when we have reached, or will we ever reach, 100%.
The ideal is that a published archetype to be approaching what we consider (sometimes guess) to be 100%, but circumstances and resource limitations often complicate achieving this. Each published archetype in CKM, no matter how rudimentary, is only published when it is clear what the plan is or pattern it will align with, drawing on past experiences and similar modelling patterns, and grounded by the rigor of the archetype ontology. It is not necessarily well documented at the moment, but if the way forward is not clear for a new archetype or archetype pattern it will be set aside until we can identify a viable path forward. This approach isnât foolproof, but it is pragmatic and has so far served us well.
Iâm not sure I understand what you mean by âbetekenismodelâ. But keen to understand more. Sounds like a new thread will be useful.
