Does Docling preserve the relationship between page-bottom footnotes and their references?
Docling did not preserve the relationship between footnotes and their references in Markdown. It included all 13 footnote texts, but placed them in the body flow without associating them with their markers.
1 of 1 test case failed
Every test case in this scenario has a result.
The test case
Each test case is judged on its own: Pass, Fail, Not gradable, or Untested. The scenario result above is based on this test case.
Footnote — out of body flow, associatedFailEvidence
The Braithe Estuary Order 1974 is usually described as a consenting instrument, and for most of its life it has been used as one. Article 6(2), on which everything in this article turns, requires the Harbour Board to obtain the written consent of the riparian owners before works are carried out on the intertidal foreshore. 1 What the article does not say, and what the Board's practice for thirty years assumed it said, is that a consent once given runs with the land.
Markdown output: the converted article and its footnote text
Copied from Output file (Markdown), line 5
Open the Markdown file ↗None of this would matter if article 6(2) were merely procedural. The argument of this article is that it is not: that the consent requirement was the price of the compulsory powers conferred by article 4, and that reading it as a formality which passes with the land empties the bargain the Order struck. 1 Braithe Estuary Order 1974, art. 6(2). The Order is reproduced in full in the appendix to R. Sandiman, The Braithe Estuary: A Documentary History (Vardenne, 1988), 211-24. 2 Minutes of the Harbour Board, 14 May 1973 (Braithe Record Office, BHB/3/17), 4. 3 The point was not argued in Tarbert v Braithe Harbour Board [1981] 2 EstLR 88, where both parties proceeded on the assumption that art. 6(2) applied. 4 The Board's file on the 2011 consents (Braithe Record Office, BHB/11/4) contains three letters returned undelivered and a note of a telephone call to a solicitor in Winnipeg. That changed with the sediment recharge programme. Placement of dredged material on the upper foreshore is works within the meaning of article 3, as the Board has always accepted, and the programme requires placement in the same locations year after year. A consent that expires with a conveyance is an expensive thing to administer, and by 2009 the Board was asking whether article 6(2) meant what it appeared to mean. 5
Copied from Output file (Markdown), lines 17–27
| ✗ | Expected: Footnote text is present, kept out of the body flow, and associated with its reference.Found: All 13 footnote texts appeared as ordinary paragraphs at page breaks. Their markers were bare digits, with no footnote references associating the text with them. |
The conversion succeeded, but the Markdown did not keep footnotes outside the body or attach them to their references. The result is a failure of the requested structure, not an unsuccessful conversion.
Author line heading. The author line becomes a level-two heading.
Page furniture dropped. Running heads, the journal line, and page numbers are dropped as page furniture.
Configuration and setup
How this tool was set up for the run and what the test needed in place. Each row is a fact from the run's records; a fact the records do not hold is left out, not guessed.
How this scenario is graded
How we decide Pass, Fail and Not gradable. The same rules apply to every tool tested on this scenario.
Each test case gets one result per tool: Pass, Fail or Not gradable. A test case we haven't run yet shows Untested. There are no partial results.
- Pass: the tool did everything the test expected, and nothing it said contradicts the correct answer.
- Fail: at least one expected behaviour clearly didn't happen; the test case's result says which one and quotes the tool.
- Not gradable: our evidence couldn't settle the outcome (for example a record we needed is missing). It is never counted as a fail, and the test case's result says what's missing.
Where this sits in the benchmark
This page is one cell of a larger study: one tool, one scenario. Only this benchmark's frame appears here.
| Level | Name | Scope |
|---|---|---|
| Benchmark | Converting a complex PDF into clean Markdown with an open-source library → | 12 scenarios · 10 tools |
| Capability | Heading & Section Structure → | |
| Scenario | Footnotes at the bottom of the page → | |
| Tool | Docling → |
Global scenario definition → · Global capability definition → · Docling product page →
History of this result
What has happened to this result since it was first published. Runs and grades are never overwritten: a retest or a re-grade publishes a new result and keeps the earlier one readable.
Act on this result
Nothing filed here edits the run or the grade. A challenge opens a review, and a review can produce a new run or a re-grade — which becomes the current result and leaves this one in the history.
You run the same kind of test against your own setup and get the same behaviour.
Agree →Yours behaves differently. Tell us what you got, with a screenshot if you have one.
Disagree →Something here is wrong — a reference value, a transcription, a grade.
Report an issue →Tell us what changed and we schedule a rerun of the failing test case. The old result stays as history.
Vendor notice →The same record is available as structured data through the AI Demos MCP server, with the counts, the coverage and every per-test-case reason carried as fields.
Verify the proof files
These files support this result. Open a file to inspect the original evidence.
File fingerprints (SHA-256)
A fingerprint identifies the exact file used for this result.