A single term may have different definitions, that is, not different meanings but different angles (customer, terminologist, student, manager, SME).

It would be interesting to add a <definitionAlt> element based on topic/section in the termBody in order to implement it.

The processing of a semantic net must be public, because it has to be injected into the webhelp processing, but creating a new transtype does not work, because the plugin must not depend on the webhelp plugin. Therefore the processing should shadow properties using the (conditional property) <local> Ant task.

A term should not have a deprecated <termNotation>, that is an allowed <termNotation> of another term. This should be validated after the preprocessing phase.

term 1

<fullForm language="en-GB" usage="notRecommended">
   <termVariant>foo</termVariant>
</fullForm>

term 2

<fullForm language="en-GB" usage="preferred">
   <termVariant>foo</termVariant>
</fullForm>

Currently, the status is set using the usage attribute:

<fullForm usage="preferred" language="en-GB">
  <termVariant>my term</termVariant>
</fullForm>

Because the usage may different in another term domain, the usage should be expressed with an element. Of course, this should be avoided, because the terminology should be harmonized.

<fullForm usage="preferred" language="en-GB">
  <usage status="preferred" termdomain="foo"/>
  <usage status="notRecommended" termdomain="bar"/>
  <termVariant>my term</termVariant>
</fullForm>

In that case, the status and the termdomain attributes should be mandatory for the <usage> element.

The relation names for thesauri are normed in ISO 2788, see Wikipedia: Thesaurus.

For our Oxygen terminology checker add-on we use a simple XML format to encode the rules:
https://www.oxygenxml.com/doc/ug-editor/topics/terminology-checker-addon.html
An XSLT stylesheet could produce such XML files from the original termentry terminology files.

Currently the message is

The term foo is not allowed.

Better is probably:

The term foo is a not recommended term and should be replaced with the preferred term bar.

The semantic net does not work any longer.

Test and proof-read the README.md.

This makes it easier to add annotations in <termBody> and <termNotation> and use all elements, that are allowed in a <div> element.

Also rebuild samples with a containermap

Authoring a new term entry will be much easier in oxygen author mode when term relation structure is predefined within the template.

To increase the performance, an approach can be to compile the XSLT and export it in the compiled form. Then, this compiled stylesheet can be used from other Schematron file, or from other XSLT via xsl:use-package.

On a 4K display the termbrowser displays text that's too tiny to read, in some cases the letters overlap. Would be nice to use this on my new 32" display :-). Thx.

Implement a transformation switch to select, whether termchecker rules should be generated for <source>, <target> or both elements.

...

The language codes should no longer be maintained in the tementry.mod. This is not flexible. If someone is not happy with those values, he/she should just remove/change the Schematron file, but should leave the MOD file untouched.

Solution:
Update file: org.doctales.terminology/xsl/terminology2semanticnet.xsl
go to:

<!-- partOf -->
<xsl:for-each select="document(./$filename)/descendant::*[contains(@class, ' termentry/partOf ')]">

change order of value-of elements to:

<xsl:for-each select="document(./$filename)/descendant::*[contains(@class, ' termentry/partOf ')]">
                <xsl:text>{id: '</xsl:text>
                <xsl:value-of select="$key"/>
                <xsl:text>2</xsl:text>
                <xsl:value-of select="@keyref"/>
                <xsl:text>', from: '</xsl:text>
                <xsl:value-of select="$key"/>
                <xsl:text>', to: '</xsl:text>
                <xsl:value-of select="@keyref"/>
                <xsl:text>', arrows: 'to', label: 'is part of'},</xsl:text>
</xsl:for-each>

A <termgroup> element should be integrated into the <termmap> to group termentry topics. That offers the possibility to implement features like sorting terms without affecting other <topicref> elements.

Example

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE termmap PUBLIC "-//DOCTALES//DTD DITA DOCTALES Termmap//EN" "termmap.dtd">
<termmap id="termmap">
 <title>Termmap</title>
 <termgroup>
  <termref href="term1.dita" keys="term1"/>
  <termref href="term2.dita" keys="term2"/>
 </termgroup>
 <topicref href="other.dita"/>
</termmap>

If I set the associated framework location to org.doctales.terminology/framework/
Oxygen (18.1 + 19) finds the termentry doctype association only but not the termmap doctype association.
Because of that, Oxygen does not provide the author with any transformation scenario.

It seems that Oxygen ecxpects that multiple frameworks must located in separate subdirecories of the added framework directory.
If I try to safe a new external framework in the top level frameworks dirctory the following message occours:

The "*.framework" file must be saved in a subdirectory of one of the configured frameworks directories..."

Moving the framework and referenced ressources in a subdirectory solves the problem.
Associated ransformation scenarios are available after that.
However, I facing another transformation failure
java.lang.ClassNotFoundException: org.dita.dost.module.GenMapAndTopicListModule,
so I am not shure, if other things are to do regarding this issue.

<sqf:fix id="BoLB/L"> <!-- This is invalid -->
    <sqf:description>
        <sqf:title>Replace with an allowed acronym: 'B/L'</sqf:title>
    </sqf:description>
    <sqf:stringReplace regex="(\b(BoL)\b)" select="'B/L'"/>
</sqf:fix>

throws

The value 'BoLB/L' of attribute 'id' on element 'sqf:fix' is not valid with respect to its type, 'NCName'.
--

We had this remark from one of our users:

 We have looked at Doctales’ terminology plugin (https://github.com/doctales/org.doctales.terminology). 
 This has great functionality as it offers not only the possibility to check for incorrect terms, but also to manage terms and concepts like in a terminology database. 
 However, this plugin is seemingly not maintained any more and does not work with DITA-OT 3.6.

The CDN has changed. Use:

<link
  rel="stylesheet"
  href="https://cdn.jsdelivr.net/gh/lipis/[email protected]/css/flag-icons.min.css"
/>

The CSS syntax has changed. Use:

<span class="fi fi-us">

Add an MS Excel template with VBA macros. Each row should represent a <termentry> topic.

It should be possible to create output that only contains allowed terms in source and target language and no other information like definitions, etc. Unfortunately CAT tools do not fully support TBX-Basic in many cases.

Change the order of <termNotation> child elements so that <termVariant> is always the first child. That would simplify working with the hierarchy view in . Test with all transformations.

BEFORE

<fullForm language="en-GB" usage="preferred">
  <termSource>
    <sourceName>source</sourceName>
  </termSource>
  <termVariant>term</termVariant>
</fullForm>

AFTER

<fullForm language="en-GB" usage="preferred">
  <termVariant>term</termVariant>
  <termSource>
    <sourceName>source</sourceName>
  </termSource>
</fullForm>

The Schematron termchecker is working, but it has been hammered quickly. This has to be refactored for quality and performance reasons.

If termchecker rules are generated for terms with a region code, e.g. en-GB, than terms without a region code, e.g. en should be offered as replacements.

<fullForm usage="preferred" language="en">
  <termVariant>anti-lock braking system</termVariant>
</fullForm>
<fullForm language="en-GB" usage="notRecommended">
  <termVariant>anti-skid braking system</termVariant>
</fullForm>

The terminology statistics are broken, fix them.

The current language-country code pairs are not sufficient.

Implement a new termbrowser, that generates a termbrowser by only using the html5 transformation and flavoring this with some JavaScript and CSS libraries.

Eliot K. roughly summarized, how this is done for the OT docs:

- Set nav-toc to "full" or "partial" depending on your preference
(http://www.dita-ot.org/dev/parameters/parameters-html5.html)
- Set args.copy-css to "yes"
- Set args.css to "dita-ot-docs.css"
- Set args.cssroot to "docsrc/resources" (this is relative to the root of
the OT. You may need to make this absolute. For example, in oXygenXML 18
you can do "${configured.ditaot.dir}/docsrc/resources".