Skip to main content

Independent editorial reference · no accreditation and no qualification awarded · general information only, not legal or professional advice

Open Data DeskGalway · IE

Transparency

Publishing standards

The rules this site applies to itself: how accuracy is pursued, how sources are treated, how independence is maintained, and what happens when something is wrong.

Reviewed 12 June 2026 · sources dated in text · general information only

Accuracy as a procedure

Accuracy here means that every factual claim can be traced to a source, and that the sentence around a fact does not extend it beyond what the source establishes. Before publication, claims are traced individually; numerical claims are re-derived from the original document rather than copied from a summary of it; and definitions attached to figures are checked so that a unit, period or population is not quietly changed.

The most frequent way accurate material becomes misleading is not a wrong number but an over-general sentence built on a correct one. Standards therefore apply to construction as much as to content: a claim about eighteen local authorities is not written as a claim about the country.

Source hierarchy

Sources are preferred in a fixed order: primary documents first, including legislation, official guidance, statistical methodology notes, register filings and decisions of courts and regulators; then peer-reviewed research and official statistics, read for their methodology rather than their abstracts; then standards issued by recognised professional bodies; and finally established reporting, used only where it records a fact traceable to one of the earlier layers.

Vendor marketing, undated pages, aggregator summaries of documents that are themselves available, and social posts are not treated as sources, except where a post is the object being described. Every source is recorded with publisher, date and access date, because registers and portals change in place.

Record, inference and open question

Three registers are kept apart throughout the site. A record is what a document establishes. An inference is a reasonable reading that is not established. An open question is what the sources do not settle.

The distinction is carried in the language rather than by labels: records are stated plainly, inferences are marked by construction, and open questions are gathered explicitly at the end of a page. If a reader cannot tell which register a sentence belongs to, that is a defect and a report about it is welcome.

Describing legal and regulatory duties

Passages describing access rights, data-protection duties or media law are checked against the instrument itself, written at a level of generality that cannot function as advice, and accompanied by the standing statement that a specific situation requires the rules in force and a qualified professional.

Where practice varies by jurisdiction, the page says so rather than presenting the Irish position as universal. Where a duty is contested or the guidance is unsettled, that is described instead of resolved.

Independence

No tool vendor, publisher, platform, campaign or public body pays to be mentioned or treated favourably, and none is offered any service by this project. There is no advertising, no sponsored content, no affiliate linking and no paid placement, and the project is funded entirely by the publishing company named on the contact page.

No rankings, reviews or endorsements of tools or services are published. Categories of tools are described together with the checks their use requires, which is information a reader can act on without being steered toward a purchase.

Use of automated tools

Material is written and verified by people. Automated tools may be used in ancillary work: checking spelling and consistency, formatting, converting a table, or generating an initial list of candidate sources that is then read and assessed by a person.

No page is published as machine-generated text, and no factual claim enters a page on the authority of an automated tool. Where such a tool suggests a source, the source itself is consulted before anything is written from it. Responsibility for every page rests with the publisher as a whole.

Titles, descriptions and what they promise

A page's title and its short description are read far more often than the page itself, in a list of results or in a summary produced by another system, so they are held to the same standard as the text. A title names what the page covers and does not announce a conclusion the page does not reach; a description summarises scope rather than advertising a benefit; and neither uses a superlative, a promise or a number that does not appear in the page.

Section headings follow the same rule. A heading is a label for what follows, not a claim to be defended later, and where a section describes a contested question the heading says so instead of resolving it in four words.

Quantitative claims in our own pages

Where this site states a number about itself — how many field pages exist, how many terms are defined, how long a response takes — that number is checked against the site as published before the page goes out, and is corrected when it drifts. It is a small thing, but a reference that is inaccurate about its own contents has no standing to describe how to check anyone else's figures.

Illustrative numbers inside worked examples are labelled as illustrative in the surrounding sentence. No figure is presented as a measurement of the real world unless a source is named beside it with a date, and no invented statistic, quotation, case study or named individual appears anywhere on the site.

Conflicts of interest

The project has no advertising relationship, no affiliate arrangement, no sponsorship and no client relationship with any tool vendor, publisher, platform, campaign or public body, and it provides no paid services to any organisation that appears in its material. There is therefore no interest to weigh against the material as published.

If that changed — if the publisher acquired any commercial relationship with an organisation described here — the relationship would be stated on the page concerned and on the project page, and the page would be reviewed by someone with no part in the arrangement before remaining published. A relationship that could not be disclosed in that way would not be entered into.

Corrections and revisions

How different kinds of change are handled
SituationActionVisible to the reader
Factual errorCorrected and annotatedDated correction note on the page
Source supersededPage revised, source updatedRevision note and new review date
Ambiguous wordingRewritten for clarityNoted where the sense could have changed
Disputed interpretationExamined; disagreement recordedStatement of the disagreement
Page unmaintainableWithdrawnRecorded on the index of pages

Review cycle

Every page carries the date of its last review and is reread at least annually. A review is triggered earlier by a change in legislation or official guidance, a redefinition of a statistical series, the relocation or withdrawal of a public register, relevant new research, or a reader report.

A review confirms, revises or withdraws. Confirmation advances the review date; revision is noted at the foot of the page; withdrawal is recorded rather than leaving a page to decay quietly.

What these standards do not promise

A published standard reduces error without eliminating it. Documents are misread, sources are superseded between reviews, and general descriptions will not fit every case. These standards are a commitment about process, not a warranty about content, and the limits stated on every page continue to apply.