Publishing works when a manuscript or knowledge object is transformed into a stable, rights-cleared, edited, identifiable and discoverable publication whose authorship, version, format and later corrections remain clear enough for readers, libraries, markets and future researchers to know what was published, by whom, in which form and what changed afterward.
Publishing is not merely pressing “post”, printing pages or uploading a PDF. It is a chain of decisions that turns private or provisional work into a public object with a defined version, audience, rights position, metadata, distribution route and long-term record.
The publishing product is not only content. It is content plus identity, version, rights, discoverability, distribution and a correction trail.
Quick Read: The Whole Publishing Mechanism
IDEA / KNOWLEDGE → AUTHORSHIP → MANUSCRIPT → RIGHTS / PERMISSIONS → EDITORIAL ASSESSMENT → DEVELOPMENTAL EDITING → FACT / REFERENCE CHECK → COPYEDIT → PEER REVIEW WHERE APPLICABLE → AUTHOR REVISION → ACCEPTANCE → PRODUCTION → DESIGN / TYPESETTING → ACCESSIBILITY → IDENTIFIERS → METADATA → PROOF / QA → RELEASED VERSION → DISTRIBUTION → DISCOVERY → BOOKSTORE / LIBRARY / DATABASE / PLATFORM → READER USE / CITATION → ERROR / NEW EVIDENCE → CORRECTION / RETRACTION / NEW EDITION → PRESERVATION → FUTURE DISCOVERY
The governing RFE is:
Can an author’s or researcher’s work be transformed into a stable, rights-cleared, edited, identifiable, accessible and discoverable publication whose version, provenance and corrections remain clear enough for readers, libraries, markets and future researchers to know what was published, by whom, in which form, and what later changed?
1. Writing and Publishing Are Different Jobs
Writing creates a text or other intellectual work. Publishing turns that work into a public version that other people can identify, acquire, discover, cite and return to later.
A manuscript can be excellent and still be unpublished. A published work can be widely distributed and still be poor.
writing ≠ publishing; publication ≠ quality.
2. Authorship Identifies Intellectual Responsibility
Authorship tells readers who is responsible for the intellectual work. In books this may be straightforward. In research, reference works, edited volumes and collaborative projects, responsibility can be distributed among authors, editors, translators, illustrators, data creators and other contributors.
Publishing should preserve those roles rather than compress every contributor into an undifferentiated byline.
author ≠ publisher; editor ≠ author by default.
3. A Manuscript Is a Candidate Publication
A manuscript may be a novel, textbook, scholarly article, report, manual, children’s book, monograph or another work awaiting editorial and production decisions.
Until a version is formally released, the manuscript can change substantially. The publishing system therefore needs to distinguish draft, accepted manuscript, proof and released publication.
4. Rights Must Be Cleared Before Distribution
Text, photographs, illustrations, maps, datasets, music and quotations may carry copyright or contractual restrictions. Publishers need to know which rights the author controls, which permissions are required and which licence will govern the released work.
Rights can differ by territory, language, edition, format and channel.
publication access ≠ permission to reuse everything inside it.
5. Editorial Assessment Decides Whether the Work Fits the Publisher’s Job
Editors assess audience, purpose, originality, quality, fit, market or scholarly contribution, legal risk, production cost and whether the work can be improved enough to justify publication.
Rejection does not automatically mean a work is false or worthless. Acceptance does not prove it is true. Editorial selection answers whether this publisher chooses to invest its resources and reputation in this work.
6. Developmental Editing Repairs the Structure
Developmental editing can reshape argument, sequence, scope, examples, chapters, narrative structure and audience fit.
Good developmental editing improves the route through the ideas without silently taking ownership away from the author.
editing ≠ authorship.
7. Fact and Reference Checking Test Particular Claims
Depending on publication type, editors or specialist checkers may verify names, dates, quotations, references, calculations, captions, URLs and factual assertions.
Fact checking reduces error. It does not turn the publisher into an omniscient guarantor of every proposition.
8. Copyediting Repairs Language and Consistency
Copyediting addresses grammar, spelling, style, terminology, internal consistency, references, tables, headings and house conventions.
A perfectly copyedited false claim remains false. Linguistic polish and evidential reliability are separate dimensions.
9. Peer Review Is a Specialist Filter, Not a Truth Certificate
In scholarly publishing, peer reviewers may evaluate novelty, method, argument, evidence, interpretation and relevance before an editor decides whether to accept a paper or book.
Review can identify serious weaknesses, but reviewers work with limited time and information. Errors, misconduct and weak inference can survive review. Later replication or evidence can overturn a published claim.
peer review ≠ proof; published ≠ true.
10. Revision Is Part of Publication, Not Evidence of Failure
Authors may revise repeatedly after editorial comments or peer review. The process improves a work by exposing it to questions before release.
The important control is version clarity: readers should be able to distinguish a private draft, accepted version, proof and published version where those distinctions matter.
11. Acceptance Creates a Production Commitment
Acceptance means the publisher has decided to move the work into production, subject to remaining contractual, editorial and quality checks.
In scholarly systems, an accepted manuscript may circulate before the final formatted version appears. Those versions should not be silently treated as identical.
12. Production Turns the Work Into a Defined Publication
Production can include typesetting, layout, cover design, image preparation, figure conversion, table formatting, pagination, eBook packaging and preparation of print or digital files.
Production is where abstract content becomes a specific product form.
13. Format Is Part of Publication Identity
A hardcover, paperback, EPUB and PDF can contain nearly the same intellectual work while being distinct publishing products.
The International ISBN Agency states that the ISBN identifies a particular publisher or imprint together with a specific title, edition and format, and that different product forms such as paperback, EPUB and PDF should normally be identified separately.
same work ≠ same product form.
14. ISBN Identifies a Book Product; It Does Not Create Copyright
An ISBN is a 13-digit International Standard Book Number used across publishing, bookselling and library supply chains to identify monographic publications.
The International ISBN Agency explicitly states that an ISBN is an identifier and does not itself convey legal or copyright protection.
ISBN ≠ copyright ≠ quality certificate.
15. DOI Identifies a Research Object Through Persistent Metadata
Crossref members register DOIs together with metadata describing scholarly content. Crossref explains that the DOI remains associated with the work even if the work moves to another website, while the owner maintains its associated metadata and resolution URL.
Crossref does not receive the publication itself during normal content registration; it receives metadata that persistently points to the content on the publisher’s site or hosting platform.
DOI ≠ content; DOI ≠ validation of the claims inside the content.
16. Metadata Makes Publications Discoverable and Exchangeable
Publishing metadata can include title, author, publisher, date, language, edition, format, identifiers, abstract, subject, licence, accessibility features, contributors and relationships to other versions.
Booksellers, libraries, search engines, citation systems and databases depend on this metadata to know that a work exists and how to route users toward it.
metadata ≠ publication; metadata is the routing record around the publication.
17. Accessibility Must Be Designed Into the Publication
A digital book can technically open and still be unusable to readers who depend on keyboard navigation, structured headings, alternative text, captions, semantic reading order or screen readers.
W3C’s EPUB Accessibility specifications combine content accessibility requirements with accessibility metadata so readers and distribution systems can discover which publications meet particular needs.
18. Standards Have Versions Too
As of August 2026, W3C lists EPUB 3.3 as a Recommendation, with an updated Recommendation dated 13 January 2026, while EPUB 3.4 remains on the draft standards track. EPUB Accessibility 1.1 remains a Recommendation, while EPUB Accessibility 1.2 is progressing on the Recommendation track and is not yet a final Recommendation.
This distinction matters because publishers should separate a current normative standard from a developing future specification.
19. Proofing Checks the Actual Product, Not Only the Manuscript
Proofreading happens after layout or digital production so errors introduced during production can be caught: broken line endings, missing images, wrong page references, malformed equations, font problems, inaccessible reading order or incorrect metadata.
A manuscript can be correct while the produced file is wrong.
20. Release Creates a Version Readers Can Point To
Publication creates a defined public state: a particular edition, journal version, web release or other version that readers can acquire or cite.
The “version of record” in scholarly publishing is a stable published version. Stability does not make its claims immune to correction.
version of record ≠ immutable truth.
21. Distribution Moves the Publication Into Supply Networks
Print books move through printers, warehouses, wholesalers, retailers and libraries. Digital publications move through hosting platforms, online retailers, aggregators, institutional subscriptions, repositories and direct sales.
A finished file sitting on a publisher’s computer is not meaningfully distributed.
22. Distribution and Discovery Are Different
A publication can be available for sale or access and still be hard to find because its metadata is incomplete, keywords are weak, library records are missing, databases do not index it or search systems rank it poorly.
distribution ≠ discovery; discoverability ≠ access.
23. Libraries Turn Publications Into Long-Term Knowledge Infrastructure
Libraries acquire or license published works, describe them, preserve them and make them discoverable to readers. National libraries may also receive legal-deposit copies to preserve a country’s published record.
See How Libraries Work for the collection → metadata → discovery → access → preservation mechanism.
24. Open Access and Public Domain Are Different
An open-access scholarly article may be freely readable while remaining copyrighted under a licence that defines permitted reuse. A work in the public domain is not protected by copyright in the same way.
free to read ≠ free of copyright; open access ≠ public domain.
25. Self-Publishing Still Requires a Publishing System
An author can act as publisher, but the jobs do not disappear. The author-publisher still needs editorial review, rights clearance, product identifiers where appropriate, metadata, design, accessibility, distribution, pricing, legal deposit where required and long-term version control.
self-publishing ≠ absence of publishing machinery; it means one actor owns more of the machinery.
26. Popularity Measures Reach, Not Truth or Lasting Quality
Sales, downloads, bestseller lists and citation counts can reveal attention or reach. They cannot by themselves determine whether a book is accurate, whether research is methodologically sound or whether a work will retain long-term value.
bestseller ≠ best evidence; citation ≠ correctness.
27. Corrections Preserve the Record While Repairing It
Minor errors can sometimes be corrected within a digital work with an explicit revision history. In scholarly publishing, editorially significant corrections are often issued as separate notices linked to the original publication.
Crossref’s update infrastructure allows publishers to register relationships between original publications and later corrections, withdrawals or retractions.
correction ≠ pretending the earlier version never existed.
28. Retraction Signals That a Publication Should No Longer Support Its Original Claim
A retraction is stronger than an ordinary correction. It tells readers that serious problems make continued reliance on the original work inappropriate for its intended scholarly claim.
Retraction should remain linked to the original record so the scholarly history is visible. Erasing the original publication can hide what happened and allow uncited copies to circulate without warning.
retraction ≠ erasure.
29. New Editions Solve a Different Problem From Corrections
A new edition may substantially revise content, structure, chapters, data or pedagogy. A reprint may reproduce substantially the same edition.
Readers therefore need edition information when the differences affect meaning or citation.
edition ≠ reprint.
30. Preservation Keeps Published Knowledge Reachable After the Market Moves On
Commercial platforms can close, links can break, file formats can become obsolete and publishers can disappear. Libraries, repositories, legal deposit and preservation systems protect published knowledge beyond its immediate sales life.
Identifiers, metadata and archival copies become especially important when the original distribution channel no longer exists.
31. AI Changes Production but Not the Need for Responsibility
AI can assist drafting, translation, editing, indexing, alt-text generation, metadata preparation and production. It can also fabricate references, misstate facts, imitate living writers, obscure contribution and create rights uncertainty.
The publishing system still needs a responsible human or institution that can answer for what is released, verify important claims and disclose AI involvement where policy or reader trust requires it.
AI-generated prose ≠ verified authorship ≠ evidence.
32. Publishing and Journalism Overlap but Are Not the Same System
News organisations publish, but journalism has additional real-time reporting, source verification, public-interest and newsroom-accountability mechanisms. Book and scholarly publishing operate on different time scales and editorial structures.
Existing How News Works material therefore keeps journalism and news integrity rather than being absorbed into this generic publishing mechanism.
33. Worked System 1: A Children’s Science Book Reaches a Library
An author writes a science book for children. The full chain can be:
MANUSCRIPT → SCIENCE / AGE-LEVEL REVIEW → DEVELOPMENTAL EDIT → IMAGE RIGHTS → COPYEDIT → DESIGN → ACCESSIBILITY CHECK → ISBN FOR EACH PRODUCT FORM → METADATA → PROOF → PRINT / EPUB RELEASE → DISTRIBUTOR → LIBRARY CATALOGUE → CHILD / PARENT DISCOVERY → BORROW / READ → ERRATA / NEW EDITION.
The science must be accurate, but the publication also has to be readable, identifiable, discoverable and appropriate for the intended learner.
34. Worked System 2: A Scholarly Paper Moves Into the Research Record
A research team submits a manuscript to a journal.
RESEARCH MANUSCRIPT → EDITORIAL SCREEN → PEER REVIEW → REVISION → ACCEPTANCE → PRODUCTION → DOI + METADATA REGISTRATION → VERSION OF RECORD → INDEXING / LIBRARY DISCOVERY → CITATION → REPLICATION / NEW EVIDENCE → CORRECTION / RETRACTION IF REQUIRED.
The DOI stabilises identity and routing. Peer review adds scrutiny. Neither one proves the scientific claim independently of evidence.
35. Worked System 3: An Accessible eBook Has to Advertise Its Accessibility
An EPUB can contain semantic headings, text alternatives, a logical reading order and navigation suitable for assistive technology.
But if accessibility metadata is absent, a reader or retailer may not know those features exist before purchase or download.
The route becomes:
ACCESSIBLE CONTENT → EPUB CONFORMANCE → ACCESSIBILITY METADATA → PRODUCT METADATA → DISTRIBUTOR → RETAIL / LIBRARY DISCOVERY → READER FILTER / SEARCH → ACCESSIBLE READING SYSTEM → USE.
36. Worked System 4: A Serious Error Appears After Publication
A published scholarly article is later found to contain a major analytical error that undermines its central conclusion.
A trustworthy route is:
ERROR REPORT → EDITORIAL INVESTIGATION → AUTHOR RESPONSE → INDEPENDENT REVIEW WHERE NEEDED → CORRECTION OR RETRACTION DECISION → PUBLIC NOTICE → DOI / METADATA RELATIONSHIP UPDATED → DATABASE / LIBRARY PROPAGATION → FUTURE READER WARNING.
The system improves by making the correction visible rather than silently replacing history.
37. Hostile Test: “It Has an ISBN or DOI, Was Professionally Published and Peer-Reviewed, Therefore It Is True”
The conclusion mixes several distinct jobs.
- An ISBN identifies a book product.
- A DOI persistently identifies and routes a research object through metadata.
- A publisher performs selection, editing, production and distribution.
- Peer review adds specialist scrutiny.
- Evidence and later testing determine how strongly the claims should be believed.
Before accepting the content, ask: What evidence supports the claim? What methods produced it? What conflicts or limitations exist? Has later work confirmed or contradicted it? Has a correction or retraction been issued?
publishing infrastructure tells us what object we are looking at and how it travelled; truth still has to answer to evidence.
Where Publishing Explanations Commonly Break
| Failure | What goes wrong | Repair question |
|---|---|---|
| Writing-publishing collapse | Creating text becomes the whole publishing system | How is the public version identified, produced, distributed and preserved? |
| Author-publisher collapse | Intellectual creator and publishing operator become one role by assumption | Who owns which responsibility? |
| Editing-authorship collapse | Editorial intervention becomes hidden authorship | Which changes belong to the author and which to editorial support? |
| Peer-review truth halo | Review becomes proof | What evidence survived later testing? |
| Published-truth collapse | Public release becomes factual validation | Which claims are independently supported? |
| ISBN-copyright collapse | Identifier becomes legal protection | What rights and licence actually govern the work? |
| DOI-content collapse | Persistent identifier becomes the research itself | Where is the publication and what metadata does the DOI route to? |
| Metadata-publication collapse | Catalogue record becomes content | What underlying publication does it describe? |
| Edition-reprint collapse | A substantially revised edition is treated as the same product | Did content change materially? |
| Format collapse | Paperback, EPUB and PDF are treated as one supply-chain object | Which product form is being identified? |
| Distribution-discovery collapse | Availability becomes findability | Can a real reader discover it through metadata and indexing? |
| Discovery-access collapse | Search result becomes usable access | Can the reader legally and technically obtain it? |
| Bestseller-quality collapse | Sales become truth or durability | What does popularity actually measure? |
| Citation-truth collapse | Citations become correctness | Why is the work cited and what evidence supports it? |
| Version-of-record absolutism | Stable published version becomes immutable truth | What later correction or retraction exists? |
| Correction-erasure collapse | Error repair silently rewrites history | Can readers see what changed? |
| Retraction-erasure collapse | Retracted work disappears without warning | Is the retraction visibly linked to the original? |
| Open-access public-domain collapse | Free reading becomes unrestricted ownership | What licence governs reuse? |
| Self-publishing shortcut | Removing a traditional publisher is assumed to remove editorial and metadata jobs | Who now owns those jobs? |
| AI-output authority | Generated text becomes verified authorship or evidence | Who checked the claims and accepts responsibility? |
Singapore: Publishing Connects Directly to National Memory
Singapore’s National Library provides a clear final stage in the publishing system. Its Legal Deposit function, under the National Library Board Act, requires two copies of every physical publication and one copy of every digital publication distributed in Singapore to be deposited with the National Library.
The National Library describes this collection as a repository of Singapore’s creative and intellectual output. Publishing therefore does not end at sale: selected national publishing infrastructure carries released works into long-term cultural and research memory.
NLB also operates the Singapore ISBN function and related publisher services, connecting local publications into international identification and library-supply systems.
How to Read Any Publishing System
- Work: What intellectual object is being published?
- Authorship: Who accepts responsibility for it?
- Rights: Who controls publication and reuse?
- Editorial selection: Why did this publisher accept it?
- Editing: What structural and language work occurred?
- Review: Was specialist or peer review appropriate?
- Version: Draft, accepted manuscript, proof or released version?
- Production: Which print or digital forms were created?
- Accessibility: Can intended readers use the publication?
- Identifier: ISBN, DOI, ISSN or another identifier?
- Metadata: Can systems describe and exchange it?
- Proof: Was the produced product itself checked?
- Release: Which version is public?
- Distribution: Which channels carry it?
- Discovery: Can readers and libraries find it?
- Access: What purchase, subscription or licence condition applies?
- Citation: Can readers point to the same version?
- Correction: What happens when errors appear?
- Preservation: Who keeps the publication reachable later?
- World return: What new evidence would force revision, correction or retraction?
Current Evidence and Standards Anchors
- International ISBN Agency — What Is an ISBN? for publisher, edition and format identification and the explicit separation between ISBN and copyright.
- Crossref — Content Registration for DOI registration, persistent metadata, publisher-maintained URLs and research-object relationships.
- Crossref — Registering Updates for metadata relationships among originals, corrections, withdrawals and retractions.
- W3C — EPUB 3.3 for the current EPUB 3.3 Recommendation.
- W3C — EPUB Accessibility Standards and Drafts for the distinction between EPUB Accessibility 1.1 Recommendation and the developing 1.2 specification.
- National Library Singapore for the current Legal Deposit requirement and preservation of Singapore’s published heritage.
Where This Fits in the eduKate Architecture
This article owns the manuscript → editorial / production control → identifiable public version → distribution / discovery → correction / preservation mechanism.
- How Information Works owns information representation and transmission generally.
- How Knowledge Works owns how information becomes usable understanding.
- How Libraries Work owns collection, metadata, discovery, access and preservation after publication.
- How Archives Work owns activity-generated records, provenance, custody and evidentiary continuity.
- How Scientific Research Works owns how research questions become evidence and correctable knowledge.
- Existing How News Works material keeps journalism, newsroom verification and news-system integrity.
- How Publishing Works owns the generic process that creates and maintains identifiable published objects.
What This Article Does Not Claim
- It does not make professional publication proof of truth.
- It does not make peer review a truth certificate.
- It does not make an ISBN or DOI a quality score.
- It does not make metadata identical to the publication.
- It does not make every format one identical publishing product.
- It does not make distribution equivalent to discovery.
- It does not make discoverability equivalent to access.
- It does not make citations proof of correctness.
- It does not make open access equivalent to public domain.
- It does not make a retraction erase the publication’s history.
- It does not make self-publishing exempt from editorial, rights, metadata or preservation work.
- It does not treat AI-generated content as verified evidence by default.
Observable Mastery Test
Choose one book, research paper, report or eBook and reconstruct:
authorship → manuscript → rights → editorial review → revision → production → format → accessibility → identifier → metadata → proof → release → distribution → discovery → reader / citation → correction path → preservation.
Then ask five correction questions:
- Which identifier tells me exactly which version or product I am using?
- Which publishing signal is being mistaken for evidence of truth?
- Can a reader actually discover and access the released version?
- What correction, retraction or new-edition mechanism exists if the work is wrong?
- Who will still be able to identify and retrieve this publication after the current platform disappears?
Publishing fails when polished production hides weak evidence, identifiers are mistaken for truth, metadata cannot route the work, distribution ends before discovery, or corrections become invisible.
Publishing is not understood when we can say that a book or paper was released. It is understood when we can trace how a provisional work became a defined public version, how that version can be identified and found, what rights govern it, how readers know which version they are using, and how the record remains correctable when the published work turns out to be wrong.