Meeting 2 (2014, Oct 21) chat and notes
The meeting was conducted via the meeting room https://connect.sunet.se/openEHR Oct 21, 2014, at 12:00-13:00
Partly edited notes
Main agenda points:
- Define your main participation-interests for the project as a whole and suggest starting points where you would like to contribute initially. Those with limited time can of course monitor/review/comment work items as they evolve (instead of developing them), but then please tell us a bit about what items/subjects you want to focus on.
- How do we best make this a joint effort with e.g. the CIMI and 13606 communities?
(Interests and roadmap have been copied and edited into the project wikipage)
Interests:
-------------
Samuel Frade:
- Transformations e.g. OPT -> HTML etc (both HTML trees structures and more "clinicaly friendly" ones)
Erik Sundvall:
- Help research and test "Operational Transformation" possibilities/frameworks (for undo/redo and collaboration)
- Interest in helping out with GUI considerations (not lead GUI development)
- Interest in pedagogical RM-visualization (at some later stage of project when more important things already work..
Koray Atalag & Aleksandar Zivaljevic
- Annotation tools - formal "GUI Directives"
Marand - Bostjan, Borut, Marko
- Contributes ANTLR grammar and Java implementation of ADL 2.0 parser
Tom:
- providing webservice/rest interface to OPT generation??
- single file template e.g. https://github.com/openEHR/adl-archetypes/blob/master/Example/openEHR/single_file_template/templates/openEHR-EHR-COMPOSITION.t_clinical_info_ds.v1.adls
- OPT 1.4 and OPT2 should be in specs
Ian:
- Help validate ADL2.0 -> 1.4.OPT generation
- Explore optimal repository organisation ncluding integration with git/githib
Roadmap
-------------
- first get a rough componentization up and running
-- something that can output ADL as one of the first serialization formats in order to feed that into a backend/serverside conversion chain (based on the ADL workbench to begin with)
-- as a start it is good if the tool can read from (and write to) Git repositories
-- make sure the solution is modular and extendable so that people can build different tools, UI on top of it
- get a framework with basic editing capabilities up and running before November 2015 (hopefully a lot sooner) otherwise consider ending this project and hope other actors step in.
- develop ADL 1.4 migration strategy for implementers
Partly edited chat transcript
Erik tried to add some incomplete real-time summaries in the chat below. Sadly he forgot to start the audio recording this time...
------------------------------------- (10/21/2014 11:51) -------------------------------------
Erik Sundvall: Preparing for new meeting... Welcome!
------------------------------------- (10/21/2014 12:02) -------------------------------------
Seref Arikan: sorry can't talk - in an open office env. - @ UCL
Ian McNicoll: the UCL walls have ears ;-)
Seref Arikan: nah, but half a dozen phd students sitting around me do
Samuel Frade: Hello :)
Dong-Won Choi: hello ~ :) sorry can't talk
------------------------------------- (10/21/2014 12:06) -------------------------------------
Erik Sundvall: Ian is interested in OPT-generation, how to manage repositories, GIT is interesting but might be tricky for clinicians if UI etc is not simplified
Erik Sundvall: Tom says, use the current ADL WB to get ideas
Thomas: https://github.com/openEHR/adl-archetypes
Erik Sundvall: The international CKM is mirrored to git.
------------------------------------- (10/21/2014 12:09) -------------------------------------
Erik Sundvall: Tom says: There are logical "libraries" inside the directory structure. Self-describing text files starting with _ (underscore). Write once, read many.
------------------------------------- (10/21/2014 12:11) -------------------------------------
Erik Sundvall: Tom says: Good if all tools used similar method of treating repos. Download latest interrim build of the workbench and try it out!
Thomas: https://github.com/openEHR/adl-archetypes/blob/master/ADL15-reference/_repo_lib.idx
------------------------------------- (10/21/2014 12:14) -------------------------------------
Erik Sundvall: Tom says: a roadmap may help focus peoples interests
------------------------------------- (10/21/2014 12:18) -------------------------------------
Erik Sundvall: Samuel describes HTML transformations
------------------------------------- (10/21/2014 12:19) -------------------------------------
Seref Arikan: I'd like to see a bottom up roadmap. Some model driven core components, parsers, AOM etc. UI and functionality can go into 100 different directions
Seref Arikan: people can build different tools, UI on top of these
Seref Arikan: our problem with current tools, especially TD is that is not designed for extension
Seref Arikan: so if we have some core components, we can collaborate better
------------------------------------- (10/21/2014 12:27) -------------------------------------
Seref Arikan: All tooling for future openEHR work (adl 2) would require 1: a parser, 2: a flattener (opt) 3: XSD for these (including AOM)
Seref Arikan: opt is flattener output now is it not Tom?
Seref Arikan: the super archetype...
------------------------------------- (10/21/2014 12:31) -------------------------------------
Erik Sundvall: Tom says: Some work left to do defining "switches" etc in OPT generation regarding terminology bindings etc
------------------------------------- (10/21/2014 12:33) -------------------------------------
Seref Arikan: Tom: would it be possible to transform opt from adl 2.0 to 1.4 ?
Seref Arikan: though it does not solve the problem of having to maintain and use 1.4 artefacts
Erik Sundvall: Tom says: AWB reads ADL at the moment
Ian McNicoll: Hi Seref - I agree - we need to start by showing that we can give implemeters an easy transition to ADL 2.0.
------------------------------------- (10/21/2014 12:36) -------------------------------------
Seref Arikan: I think there is an important point we're missing: when vendors/implementers say we want better tools, as of today they mostly mean we want better tools for 1.4, now
Seref Arikan: there are short term concerns with higher priority then using adl 2.0
Erik Sundvall: Seref, Temlplate should that
Erik Sundvall: should we start witht OET or ADL 2?
Seref Arikan: Erik: no, actual deployments & project require that we edit templates, create new ones every day. They're all 1.4. good to have 2.0, but deployments will require work on 1.4 artefacts for years
Erik Sundvall: Tom speaks abput "single-file template"
Seref Arikan: yep, but that is all 2.0
Seref Arikan: I hve no objection to 2.0 stuff, but it won't help people who can't make the jump, that is all openEHR vendors with deployments out there
Thomas: sorry
Thomas: https://github.com/openEHR/adl-archetypes/blob/master/Example/openEHR/single_file_template/templates/openEHR-EHR-COMPOSITION.t_clinical_info_ds.v1.adls
------------------------------------- (10/21/2014 12:41) -------------------------------------
Ian McNicoll: @seref - We can't fix the issues in 1.4 templates without going to ADL 2.0. Thisi not just a tooling issue - it is the underlying formailsm that is flawed
Seref Arikan: Ian: that depends on what going to ADL 2.0 means
Seref Arikan: What I dont' know is: would a 2.0 based tool let me modify my current models and generate TDDs for them?
Seref Arikan: that's what I need today :)
Seref Arikan: yep, 1.4 OPT is minimum
Seref Arikan: because both TDO and TDS depends on that
------------------------------------- (10/21/2014 12:45) -------------------------------------
Ian McNicoll: @serf - agree, as long as the ADL 2.0 can generate 1.4 .opt, we can gradually improve the ADL 2.0 tooling from usability perspective
Erik Sundvall: Tom says: it would be good to get 1.4 OPT generation from the AWB - Tom might need help on this
------------------------------------- (10/21/2014 12:46) -------------------------------------
Seref Arikan: for the vendors there is only OPT
Erik Sundvall: Samuel says students do survays, he did one. 3 systems used OPT and others used OET and other forms of templates.
------------------------------------- (10/21/2014 12:47) -------------------------------------
Erik Sundvall: Ian says nowadays there are more vendors using OPT
------------------------------------- (10/21/2014 12:49) -------------------------------------
Dong-Won Choi: I also have used OPT but we can't find detailed spec. on OPT. So I analysed OPT file itself...
Ian McNicoll: Hi Dong-Won - the schema are available
------------------------------------- (10/21/2014 12:51) -------------------------------------
Dong-Won Choi: But for implemeters, fixed OPT spec. should be presented... I think...
------------------------------------- (10/21/2014 12:53) -------------------------------------
Dong-Won Choi: and... 1.4 OPT vs 2.0 OPT also be presented...
Erik Sundvall: Samuel talks about standardisation of web services (SOAP/REST)
Ian McNicoll: @Samuel see hopdscape-hopd.rhcloud.com
------------------------------------- (10/21/2014 12:56) -------------------------------------
Ian McNicoll: That demonsrates AQL firing against 2 different openEHR servers - the code is at github.com/handihealth
shinji kobayashi: agree, Dong.
Samuel Frade: ok thanks Ian
Ian McNicoll: oops wrong calll should be http://minimal-hopd.rhcloud.com
Samuel Frade: but the servers are from the same implementer?
Seref Arikan: thanks, have a nice day everyone
Iago: thank you every one, cheers
Dong-Won Choi: bye!
shinji kobayashi: Good night for me. haha. bye
------------------------------------- (10/21/2014 13:00) -------------------------------------
Erik Sundvall: Now closing the spoken part of the meeting. Please add your thoughts to chat or the agenda/speakerlist document since I forgot to start audio recording, sorry.
------------------------------------- (10/21/2014 13:04) -------------------------------------
Erik Sundvall: I obviously forgot the agenda point "How do we best make this a joint effort with e.g. the CIMI and 13606 communities?"
End of chat transcript