Annotation Model - Testing Annotation Keys

We have 3 'tests' up on Shane's test development site:

1. http://testdev.spec-ops.io:8000/annotation-model/annotations/annotationMusts-manual.html

2. http://testdev.spec-ops.io:8000/annotation-model/annotations/annotationOptionals-manual.html

3. http://testdev.spec-ops.io:8000/annotation-model/annotations/annotationAgentOptionals-manual.html

You can also get to these tests (and 1 earlier draft test that is superseded by test 1) by going to http://testdev.spec-ops.io:8000/tools/runner/index.html
 putting /annotation-model/annotations in the Run tests under path box and clicking start.

Either way you will need to paste in the json of an annotation you'd like to 'test' once for each test. Please give it a try (feel free to use annotations from the spec). Hopefully you will conclude that these latest versions of the tests are ready for the production WPT site and to be seen by implementers.

A valid annotation will past all 14 assertions of Test 1 (annotationMusts).  Note, passing assertions 7-14 does not mean that the json you pasted in has implemented these features, only that if it has, the json is corrrect.

Tests 2 (15 assertions) and 3 (16 assertions) report on what recommended and optional features have been implemented (correctly). Most annotations will fail most of these assertions; failure of an assertion means that the feature associated with the assertion was not implemented at all (the annotation is still valid) or was not implemented correctly.

These tests only check keys that are direct children of the annotation or that are associated with Agents (creators or generators) involved in the Annotation itself.  We are working on a second set of tests that will report on implementation of body and target features.  The definition schemas are all written, working on finalizing the assertions and test scripts themselves.

Feedback on the functionality of this first set of annotation-level tests will help inform finalizing the body and target level tests, which will be similar in structure and design, but obviously testing things like selectors and choice, etc. rather than @context and the like.

Thanks,

Tim Cole