ADL 2 update - single file templates; finalisation of AOM and ADL for community process.

Technical (archive)
1

For those interested in the progress of ADL/AOM 2, I have updated the wiki pages explaining the single-file template format, and showing examples. This page shows some screenshots and generated OPTs.

Single-file templates are now implemented completely in the ADL 2 Workbench and will be available in the next release, which should be in the next few weeks. (Anyone desperate for an earlier build, please check the AWB home page regularly, there will be new interim builds every few days).

The AOM has been updated to reflect these and other modelling changes; only the Archetype and Terminology packages are affected. Draft versions of the specs with these new models are available here. These documents will change in the short term as follows:

  • review and update explanatory text
  • replace UML model diagrams with tool-based diagrams

In a slightly longer timespan, the specification document will be converted to a new form, probably ODF or Asciidoc. Nevertheless, the UML models and main explanatory text is a reasonably correct reflection of the AOM and ADL 2.

These technical changes are the final major updates I propose to make to the formalism. There are various minor technical adjustments likely, including to:

  • exact details of OPT formats
  • annotations - we probably need to make the ‘value’ part of every key/value pair any AOM primitive data rather than just String
  • AOM may need a proper URI type; still under investigation
  • the descriptive meta-data may well need some final adjustments, possibly based on / related to ISO Tech Spec 13972.
  • some small adjustments around terminology code representation, based on CIMI / IHTSDO input.

These will most likely come out of the forthcoming Specifications Program editorial process. This process is open, and comments are welcome from anyone in the community at any time (we will be establishing a dedicated Jira tracker for this purpose very shortly).

So - we more or less at the end of ADL 2 development. There is of course work to do on working out migration strategies for ADL 1.4 based systems and tools, and of course much more to do on ADL 2 tooling. Interim forms of the specifications between 1.4 and 2 may be required for ISO standardisation. But the

A lot of people have provided input on these specifications, and as ever we are indebted to them. Their names are recorded in the revision history sections of the specification documents (if anyone has been left out, please contact me). Input from various people in CIMI, Mayo Clinic, Intermountain Healthcare, Polytechnic University Valencia, and of course the openEHR community is much appreciated.

  • thomas beale
2

For those interested, I have added a new version of the XML OPT for the demonstration template. This converts more of that AOM class properties to XML attributes, reducing size a bit, and arguably improving processability. The XML of the core definition in this style looks as follows:

I encourage XML specialists to have a look at these formats and suggest improvements (preferably on the wiki page).

thanks

  • thomas
3

Any one getting a BOM character encoding error when using oXygen or similar editor to open the files from the wiki can select to replace the marker. Apparently the file was edited with something that saved it marked as UTF-8 but used another BOM character. Probably a MS Windows editor, may the archetype workbench?

(attachments)

jfhgdajc.png

4

Tim,

The original generation is not generating the BOM leader. To my knowledge, for UTF-8 it was not needed, which seems to be the case, if we believe this <http://stackoverflow.com/questions/2223882/whats-different-between-utf-8-and-utf-8-without-bom&gt;\.

When I open the files in Oxygen 16, I get no error relating to BOM. And I don't see anything unusual opening with gvim (unix mode/utf-8) on Windows.

I'll have a Linux build of some of these XML related changes up in a day or so, you may be able to test the output more directly there.

- thomas

5

I didn’t state the problem well. The issues is a character in the file, that is not a UTF-8 character and tools like oXygen think it is a byte order mark.
I see Ian noticed a similar issue. It really isn’t a big deal. Maybe it happened in the upload? You will only see it depending upon your settings in oXygen maybe you have IGNORE or REPLACE selected instead of REPORT?
http://www.oxygenxml.com/doc/ug-editor/index.html#topics/preferences-encoding.html

BOM is mainly used for backwards compatibility with legacy systems, at this point.

6

Ah you're right, at the moment I have not turned quoting on of the source text data i.e. what goes between the tags. I think two types of processing is needed - one to do normal XMl quoting (things like '<' to '&lt;' etc) and then a utf-8 conversion that detects wrong characters and hopefully converts them to a UTF-8 equivalent. I think the wrong chars in this template are actually from the demographic templates that Sergio and Rigoleta did a long time ago...

- thomas

7

There are Windows, Linux and Mac builds of the next release 2.0.6 of the ADL WB available here, which contain:

  • single file templating support (make sure your clone of adl-archetypes is up to date, see the library adl2_template

  • rewritten export feature (Archetypes > Export)

  • performance improvements

  • thomas

8

Hi Thomas,

Does it have ADL 1.5 / 2 editing capabilities?

9

partial. You can create a new archetype, template, specialised archetype by right-click on the appropriate node in the explorer of the library you are in. You can then either a) use the ADL source tab and/or b) right-click on the new archetype/template in the explorer and do ‘edit’ (the orange icon) and then in the definition view, some right click operations work. It’s pretty alpha, and is being slowly worked on. - thomas

10

I will have some free time next week, do you want me to beta test it and create a small report for you? (bugs, improvements , etc).

11

anyone is welcome to do this - please report problems here <https://openehr.atlassian.net/browse/AWBPR/?selectedTab=com.atlassian.jira.jira-projects-plugin:issues-panel&gt;\.

thanks

- thomas