Author
A review method for a public dataset distribution whose download link works but whose license and access terms are not clearly stated. It separates a license document from an access-rights statement, places both at distribution level, records them in separate fields, and returns conflicting or missing terms to the owner instead of inferring reuse permission. The cited vocabulary is used within its metadata scope; no legal interpretation, crawling, indexing or citation outcome is promised.
A dataset page can offer a working download link and still leave the reader unable to answer a key question before reuse: what am I actually allowed to do with this file? The decision this review addresses is narrow. For a particular public dataset distribution, which statements must a reader be able to find, and why is a reachable download link not one of them?
In the cited vocabulary, the license property identifies the document setting the terms of availability for a distribution. The separate access-rights property describes access conditions. We record those distinct metadata roles without treating them as a combined permission decision. That separation is the load-bearing distinction for this review. A license document speaks to the terms under which the distribution is made available. An access-rights statement speaks to how the distribution is reached. Neither one substitutes for the other, and a page that supplies only one of them has answered only half of the reader’s question.
We treat file reachability separately from the distribution’s stated terms. Check the actual access-rights statement and license document rather than infer their contents from a link. Point readers to those authoritative statements and identify their respective roles.
This is a different question from whether a link should open a readable page or trigger a download. That destination choice is about what the reader receives after a click, and it is covered in the existing SHMLANG note on the download attribute versus a readable destination page. Here the object of review is the statement attached to the distribution, not the anchor behaviour.
Related reading: Download links and readable destination pages
Start with the terms for the particular distribution, even when the page also shows dataset-level information. The cited vocabulary recommends associating licensing and rights information with distributions. It allows additional dataset-level information while keeping distribution-level statements in place. The practical consequence is that a reader who only reads the dataset description may never see the statement that governs the file they are about to use. The review therefore starts by identifying the specific distribution: which format, which version, which access route. The statement has to be attached to that object, not to the dataset as a loose whole.
We do not assume that every format or version of a dataset shares identical terms. The vocabulary’s own placement rule points the other way: terms belong with the distribution, and a dataset-level statement that differs from a distribution-level statement is something the guidance advises against because it can create legal conflicts. Treating one distribution’s terms as covering its siblings is an inference the source does not support.
The record is where this review becomes checkable. The table below is a proposed field set for a distribution-level statement record. It is a recommendation, not an observation about any real dataset, and it contains no filled example, no named license and no approval state.
| Field | What it records |
|---|---|
| Dataset identity | The dataset the distribution belongs to |
| Distribution / version | The specific format or version the statement governs |
| Access statement | The rights statement concerning how the distribution is accessed |
| License document URL | The link to the legal document under which the distribution is made available |
| Relevant rights statement | Any additional rights information the owner supplies for the distribution |
| Unresolved owner question | What still needs an owner decision before publication |
The access and license fields are separate because the vocabulary gives them distinct metadata roles. A license document URL answers one question; an access statement answers another. Merging them into a single "terms" cell would recreate exactly the confusion the review is meant to prevent. The distribution/version field identifies which record the statements belong to. We recommend a separate record for each distribution, following the documented placement guidance and its warning about inconsistent rights descriptions.
Document a disagreement between dataset-level and distribution-level rights statements and ask the owner to resolve it. The cited vocabulary warns that inconsistent rights descriptions at those levels can create legal conflicts. The review records that discrepancy without deciding the legal outcome.
We recommend a fixed escalation rather than an editorial judgment. Record both statements as they appear, mark the distribution as unresolved, and route the question to whoever owns the dataset. Do not choose the more permissive wording, do not harmonise the two by editing one to match the other, and do not compose a license statement where none was supplied. The same handling applies to a missing statement. An absent license or access statement is an open question for the owner, not a default permission and not a default prohibition. This is metadata review; it is not legal interpretation, and the reviewer’s job is to surface the gap, not to resolve it.
The publication check has two parts, and neither is about anchor or destination selection. First, does the visible license and access statement match the distribution-level metadata recorded for that distribution? Second, does the license link resolve to the authoritative document the owner identified?
The first part borrows a consistency principle from the cited search guidance, which calls for structured data to match the visible text on the page. Applied here, the visible statement and the machine-readable distribution metadata should say the same thing. The second part applies the vocabulary’s placement rule: the statement the reader sees should be the distribution-level statement, not a dataset-level paraphrase that may differ from it.
What a passing check proves is limited, and the limit should be stated plainly on the page’s own review record. It proves internal consistency between the visible statement and the metadata, and it proves that the license link resolves. It does not establish that the license permits model training, commercial reuse or AI citation. Those are questions for the license document and the owner, not for a consistency check.
Nor does a passing check establish that the cited search service will crawl, index or serve the page. The cited guidance is explicit that meeting all requirements, best practices and policies does not mean the service will crawl, index or serve the content. A clean statement check and a search outcome are separate facts, and the review record should keep them separate.
What remains outside this website review is the substance of the terms themselves. The review can confirm that a distribution has a license document URL and an access statement, that the two are recorded separately, and that the visible page matches the metadata. It cannot decide whether a given license permits a given use, and it should not be presented as legal clearance.
The owner questions are therefore specific. Which distribution does this statement govern? Is there a license document, and where is its authoritative location? Is there a separate access-rights statement, and what does it say? Where the dataset-level and distribution-level statements differ, which one is authoritative? Each answer comes from the owner, not from the reviewer’s reading of the page.
This work fits inside a publishing structure that can carry it. SHMLANG describes a CMS-ready publishing structure so future SEO and GEO work compounds instead of restarting each time, and connects question baselines, approved business facts, useful pages and source research in one GEO service scope. A distribution-level statement record is a business fact of that kind: it needs an owner’s approval before it is published, and it needs a page structure that can hold it. The review stops at the boundary of metadata and publication. It does not certify a license, does not interpret legal terms, and does not govern private account or vendor permissions. Those belong to the owner and to whatever specialist advice the owner chooses to obtain.
No comments yet. Be the first!