Two example files are not valid JSON

[jgeluk]

Summary

Two files under examples/ are not valid JSON and therefore cannot be parsed by
any JSON or JSON-LD processor. Both are syntax errors, not modelling problems.

Nothing in the build parses the examples, so these have gone unnoticed.

1. examples/dprod-example.json:12 — missing comma

"outputPort": {
  "type": "DataService",
   "id": "https://www.ekgf.org/data/dataservice/permid-organization"
  "title": "Organization Endpoint",

The id member is not terminated with a comma before title.

json.decoder.JSONDecodeError: Expecting ',' delimiter: line 13 column 7

Fix: add the comma. (The stray leading space on that line is worth
normalising at the same time.)

2. examples/data-rights/example.jsonld:28 — trailing comma

  "action": "odrl:use",
}

JSON does not permit a trailing comma before a closing brace.

json.decoder.JSONDecodeError: Illegal trailing comma before end of object: line 28 column 27

Fix: remove the comma.

Verification

The other four parse cleanly:

file	status
examples/dprod-example.json	fails — line 13
examples/data-rights/example.jsonld	fails — line 28
examples/data-quality/example.jsonld	ok
examples/data-schema/example.jsonld	ok
examples/equity-trade/example.jsonld	ok
examples/sba-pool-rates/example.jsonld	ok

Suggested follow-up

Add a build-time check that parses every examples/**/*.json and
examples/**/*.jsonld and fails the build on a syntax error. Without one this
recurs silently, since the spec generator only reads each example directory's
README.md and never touches the JSON payloads.

That check is also a prerequisite for the round-trip expansion test proposed in
#246 — the examples have to parse before they can be expanded against the
context.

Prior reports

Neither syntax error was reported directly, but the surrounding problem was, and
the validation that would have caught both was proposed nearly two years ago:

• validate all examples by converting them to turtle/trig #92 (@VladimirAlexiev, 2024-09-25, still open, OMG DPROD-44) — asks for
  exactly the check proposed above: extract every example, convert them to
  Turtle/TriG in a workflow, and verify the output. Had that landed, both files
  would have failed the build on the day they were committed. validate all examples by converting them to turtle/trig #92 should be
  treated as the canonical issue for the validation workflow; this issue is the
  concrete breakage it would have caught.
• Can not parse example.jsonld #119 (@peio, 2025-05-09, reopened 2026-08-03) — examples fail to parse in
  Jena and GraphDB. Different root cause (@context redefinition rather than
  malformed JSON), but the same class of defect: examples are published without
  ever being run through a parser.
• 7.6 Data Rights - Example 2 #83 (@riannella) — the correctness thread for the data-rights example,
  through which the trailing comma reached develop.

Provenance

Both errors were introduced by us, not inherited:

• examples/dprod-example.json — the missing comma has been present since the
  file was added in a514e98 ("Add files via upload", 2025-02-07). It has never
  parsed. The unresolved <how do we search by entity type???> placeholder noted
  below arrived in the same commit.
• examples/data-rights/example.jsonld — the trailing comma was introduced by
  31a05ee (2026-03-04), applying the Ballot 4: update data-rights example.jsonld (#83) #165 data-rights variant on ballot/4.

Both survived because nothing in the build parses the examples, which is the
substance of #92.

Notes

Two other things spotted in the same pass, neither a parse failure, so listed
here only for the record rather than as part of this issue:

• examples/dprod-example.json:14 contains an unresolved authoring placeholder
  in a published example:
  "endpointURL": "https://api-eit.refinitiv.com/permid/search?q= <how do we search by entity type???>"
• examples/data-rights/README.md:18 fences a mixed Turtle/JSON illustrative
  snippet as ```json. It reads as intentional pseudocode; only the fence
  language tag is inaccurate.