OPEN ACCESS · SCHOLARLY PUBLISHING · APC · REPOSITORY · CREATIVE COMMONS · RIGHTS RETENTION · OPEN SCIENCE

How Open Access Publishing Works

Open access publishing changes the default question from “Who is allowed to read this research?” to “What conditions, infrastructure and rights are needed so that anyone can read it—and, where the licence permits, reuse it—without a subscription barrier?”

Open access is not one business model. It is a publication outcome that can be reached through several routes.

This distinction is essential. Some open-access journals charge authors an Article Processing Charge. Many do not. Some subscription journals allow authors to make accepted manuscripts openly available in repositories. Some journals sell subscriptions while making selected articles open. Some scholarly communities run journals that charge neither readers nor authors. Some articles are free to read but carry no clear reuse licence. Open access therefore sits at the intersection of access, rights, funding, infrastructure, preservation and publishing governance.

It also sits inside a broader movement toward open science. UNESCO describes open science as making scientific knowledge openly available, accessible and reusable, alongside open infrastructures, research data, software, educational resources and wider engagement with society.

The short answer

RESEARCH
  → MANUSCRIPT
  → PUBLISHING ROUTE
  ├─ OPEN ACCESS JOURNAL / PLATFORM
  ├─ SUBSCRIPTION JOURNAL + OPEN REPOSITORY
  ├─ HYBRID OPEN ARTICLE
  ├─ DIAMOND / COMMUNITY-FUNDED JOURNAL
  └─ OPEN MONOGRAPH / BOOK PLATFORM
  → RIGHTS / LICENCE
  → FUNDING MODEL
  → PEER REVIEW / EDITORIAL PROCESS
  → VERSION OF RECORD / ACCEPTED MANUSCRIPT
  → METADATA / DOI / REPOSITORY
  → PRESERVATION
  → OPEN DISCOVERY
  → READING / REUSE / CITATION
  → PUBLIC KNOWLEDGE

1. Open access begins with a barrier

Traditional subscription publishing places access behind institutional or individual payment. A university library pays for journals, a company buys access, or an individual pays for an article.

Open access removes the price barrier for the reader. The financial question does not disappear; it moves somewhere else in the system.

2. Free to read and open to reuse are different

A paper can be available online at no charge while remaining under restrictive copyright terms. Another can be openly licensed so readers may redistribute or adapt it under stated conditions.

Serious open-access analysis therefore asks two questions: Can I access it? and What am I allowed to do with it?

3. Open licences make reuse explicit

Creative Commons licences are widely used to express reuse rights clearly. Under CC BY 4.0, for example, users may share and adapt the material for any purpose, including commercial use, provided attribution is given and changes are indicated.

Other Creative Commons licences add conditions such as non-commercial use, share-alike or no derivatives. The exact licence matters.

4. Copyright does not disappear in open access

Open access usually works through copyright plus licensing, not through abandoning copyright automatically.

The author or another rights holder grants the public defined permissions. The licence states the boundaries.

5. Rights retention changes the default publishing relationship

Some open-access policies seek to ensure that authors or institutions retain enough rights to make a scholarly manuscript openly available immediately, even when publishing in a subscription venue.

Current Plan S principles emphasise author or institutional copyright retention and open licensing as part of the route to full and immediate open access.

6. “Gold” open access describes the venue route

Gold open access generally refers to articles published openly by the journal or publishing platform itself.

Gold does not automatically mean the author paid an APC. Some gold journals charge publication fees; others are subsidised, institutionally supported or community funded.

7. APCs are one funding mechanism

An Article Processing Charge, or APC, is a fee associated with publishing an article, usually intended to fund editorial, production, platform and dissemination services when readers are not charged for access.

APCs shift part of the payment burden from reading to publishing. That can expand readership while creating new equity problems for researchers without publication funding.

8. Open access does not require an APC

This is one of the most important misconceptions to correct.

Many open-access journals charge no author publication fee. The journal may be funded by a university, learned society, research infrastructure, consortium, library, public agency or philanthropic source.

9. Diamond open access removes both reader and author charges

Diamond open access describes scholarly publishing where readers do not pay to access the work and authors do not pay publication charges.

UNESCO has highlighted diamond open access as a model that treats research communication as public-good infrastructure rather than a transaction charged at either end.

10. Diamond does not mean zero cost

Editorial systems, hosting, copyediting, DOI registration, preservation, indexing and governance still cost money.

The difference is who funds those costs: institutions, consortia, societies, libraries, governments or other shared infrastructure rather than individual readers or authors at the point of use.

11. Green open access uses repositories

Green open access generally refers to making a version of a scholarly work openly available in a repository, often alongside publication in a subscription journal.

The deposited version may be a preprint, author accepted manuscript or version of record depending on rights and policy.

12. The repository is a publishing-adjacent institution

A repository does more than store PDFs. It can manage identifiers, metadata, access rights, persistent URLs, preservation, harvesting and institutional reporting.

The repository route is powerful because it separates long-term public access from the commercial business model of the journal.

13. Institutional repositories preserve local scholarly output

Universities use institutional repositories to collect and expose research papers, theses, datasets and other outputs created by their communities.

In Singapore, DR-NTU, the Digital Repository of Nanyang Technological University, supports curation, storage, access and preservation of NTU research outputs, while DR-NTU (Data) supports research-data sharing.

14. Subject repositories cross institutional boundaries

Some repositories organise around disciplines rather than universities. Researchers in physics, economics, biology and other fields may use subject repositories that aggregate work from institutions worldwide.

Subject repositories create discovery communities independent of journal brands.

15. The version deposited matters

Version	Typical state
Preprint	Author manuscript before formal peer review.
Author Accepted Manuscript (AAM)	Peer-reviewed and accepted text before publisher typesetting.
Version of Record (VoR)	Formal published edition controlled by the journal or publisher.

Repositories should identify the version clearly so readers know what stage of review and production they are seeing.

16. The accepted manuscript is not “unreviewed”

An AAM has normally passed peer review and editorial acceptance even though it lacks the publisher’s final formatting and pagination.

Confusing the AAM with a preprint weakens version literacy.

17. Embargoes delay repository access

Some publishers allow repository deposit but require an embargo before public access begins.

Open-access funder policies increasingly push toward immediate access because an embargo preserves a temporary access barrier even when the repository copy eventually becomes public.

18. Plan S made immediate access a central policy goal

Plan S requires eligible research funded by participating organisations to be published in open-access venues or made immediately available through open-access repositories, without embargo, under the funders’ implementation rules.

The current cOAlition S strategy for 2026–2030 keeps full and immediate open access as a core objective while expanding attention to sustainability, equity, infrastructure, diamond models, preprints and artificial intelligence.

19. Open-access mandates differ by funder

Researchers should not assume one global rule. Funder requirements can differ on permitted journals, repository timing, licence, manuscript version, monographs, data and publication fees.

The correct route is the policy that applies to the actual grant and institution.

20. Hybrid publishing mixes subscription and open articles

A hybrid journal is primarily subscription based but allows individual articles to be published open access, commonly through an APC.

This can give authors an open route inside established subscription journals while raising concerns about whether institutions are paying both subscriptions and publication charges.

21. “Double dipping” is an accounting and transparency concern

If a publisher receives subscription revenue for a journal and APC revenue for selected open articles, institutions may ask whether subscription pricing is adjusted appropriately.

Transparent pricing and contract structures are therefore essential in hybrid systems.

22. Transformative agreements try to change the payment architecture

Libraries and consortia have negotiated agreements that combine reading access with open-access publishing entitlements. These are often described as read-and-publish or publish-and-read arrangements.

The idea is to redirect institutional expenditure from subscription access toward open publication, although the long-term cost and equity effects vary by agreement.

23. Transformative does not automatically mean transformed

An agreement can be labelled transformative while preserving high costs, complex exclusions or limited global equity.

The meaningful question is whether the agreement actually moves the system toward sustainable and broader open access.

24. “Bronze” access is useful but licence-weak

The term bronze open access is often used informally for articles that are free to read on a publisher site but carry no clearly stated open licence.

Because reuse rights are unclear and access can potentially change, bronze is weaker than fully licensed open access under strict definitions.

25. Open access colours are shorthand, not universal law

Gold, green, hybrid, bronze and diamond are useful labels, but communities use them with slightly different boundaries.

Always inspect the actual access, licence, fee and repository conditions instead of relying on colour terminology alone.

26. APC transparency is a publishing-quality signal

The Directory of Open Access Journals Principles of Transparency and Best Practice requires journals to state clearly whether author fees exist, what they are, and whether waivers are available.

DOAJ also states that fee or waiver status should not influence editorial decision-making.

27. Hidden fees are a governance warning

Submission fees, page charges, colour charges, editing charges and withdrawal charges can matter even when the advertised APC looks simple.

Authors should inspect the full fee policy before submission, not after acceptance.

28. Waivers are an equity mechanism

APC waivers or discounts can support authors who lack publication funding.

A credible waiver policy explains eligibility, application timing and which authors in a multi-author paper must qualify.

29. Waivers do not solve the whole equity problem

Researchers may face stigma, administrative complexity or institutional inequality even when waiver systems exist.

Equitable open access asks who can publish, who can read, who controls infrastructure and whose scholarship remains visible.

30. APC inflation can recreate exclusion on the author side

A system can remove paywalls for readers while creating publishing barriers for unfunded authors.

This is one reason cOAlition S’s 2026–2030 strategy now places stronger emphasis on sustainable and equitable models beyond APC-dominated publishing.

31. Open access is not synonymous with commercial publishing

Commercial publishers, university presses, scholarly societies, libraries, universities, research infrastructures and volunteer communities all operate open-access venues.

Business model should be evaluated separately from editorial quality.

32. Open access is not synonymous with low quality

A rigorous open-access journal can use the same peer-review standards as a subscription journal. A weak subscription journal can still publish poor work.

Access model and scholarly quality are different dimensions.

33. Predatory publishing exploits open-access incentives

Some journals imitate scholarly publishing while providing little genuine peer review, misleading authors about indexing or editorial boards, hiding fees or using aggressive solicitation.

The problem is deceptive publishing behaviour, not open access itself.

34. A journal should reveal its governance

Publisher identity and address.
Editorial board.
Peer-review process.
Author fees and waivers.
Copyright and licence policy.
Publication ethics.
Corrections and retractions policy.
Archiving and preservation.
Contact information.

Transparency lets authors inspect the publication system before trusting it with their work.

35. DOAJ is a discovery and transparency layer

The Directory of Open Access Journals indexes open-access journals and records information about licences, fees, peer review, preservation and other publication practices.

DOAJ inclusion is useful evidence about transparency and open-access practice, but researchers should still assess journal fit and scholarly quality directly.

36. Indexing is not the same as quality certification

Databases have different inclusion criteria and update cycles. A journal’s presence in one database does not prove every article is strong.

Use indexing as one layer of due diligence rather than the final answer.

37. Persistent identifiers strengthen open access

DOIs, ORCID iDs, repository identifiers and grant identifiers help machines and readers distinguish articles, people, institutions and versions.

Open content without stable identity can still become difficult to cite and discover.

38. Metadata must be open enough to travel

Title, authors, abstract, licence, funding, publication date, DOI, subjects and repository location support discovery across search engines, library catalogues, scholarly databases and AI systems.

Open access works better when metadata is machine-readable and interoperable rather than trapped inside one publisher interface.

39. Repositories need preservation, not just upload buttons

Long-term access depends on stable storage, backup, fixity, persistent identifiers, metadata, format management and institutional succession.

The Plan S technical requirements for repositories emphasise persistent identifiers, interoperable metadata, machine access and long-term availability as part of a trustworthy open-access route.

40. Open access without preservation can become temporary openness

A journal can disappear. A university server can be reorganised. A platform can close.

Preservation services, repositories and legal-deposit systems help keep openly published scholarship available after the original business or platform changes.

41. Preprints are open dissemination before formal journal review

A preprint makes a manuscript publicly available before formal peer review is complete.

Preprints can accelerate communication, establish priority and invite feedback. They should be labelled clearly so readers do not mistake their publication state.

42. Preprint is a version state, not a quality judgement

Some preprints are excellent; some are weak. The important fact is that the manuscript has not yet completed the journal’s formal review and version-of-record workflow.

Readers should evaluate evidence directly and check whether a later peer-reviewed version exists.

43. Linking preprint and published article prevents version confusion

Repositories and preprint servers should update records when a version of record becomes available.

The scholarly graph becomes stronger when each version points to the others rather than competing as unrelated documents.

44. Open peer review extends openness into evaluation

Some publication systems make reviewer reports, editor decisions or author responses public.

Open access to the article and openness of the review process are separate choices, but they can reinforce one another by exposing more of how the publication was evaluated.

45. Publish–Review–Curate changes the sequence

Traditional publishing often follows review → publish. Emerging models can publish a research object first, organise review openly, and then curate or certify the work later.

The 2026–2030 cOAlition S strategy explicitly identifies Publish–Review–Curate models as one area of interest in a more open and transparent scholarly communication system.

46. Open monographs create a different economics problem

Books cost more to edit, design and produce than individual articles and serve disciplines where long-form scholarship remains central.

Open-access monographs therefore use a mixture of Book Processing Charges, library consortia, institutional funding, press subsidies, grants and community infrastructure.

47. Book Processing Charges are the monograph analogue of APCs

A BPC can fund publication of an open-access book, but it can also create author-side barriers if funding is uneven.

Open-book initiatives increasingly experiment with collective funding so publication does not depend on one author’s grant.

48. Humanities and social sciences need bibliodiversity

UNESCO’s Recommendation on Open Science explicitly supports diversity of formats, disciplines, languages and non-commercial publishing models.

Open science should not force every field into an article-centric model built around the economics of laboratory science.

49. Multilingual open access expands who can participate

Research published only in one dominant language can remain inaccessible even when the price is zero.

Translations, multilingual metadata, abstracts and local-language publishing can broaden meaningful access beyond technical availability.

50. Accessibility is part of meaningful openness

A PDF that screen readers cannot navigate is technically free but not equally usable.

Semantic structure, alternative text, accessible tables and machine-readable formats extend open access to readers with different access needs.

51. Open access and open data are related but separate

An article can be open while the underlying dataset remains restricted. A dataset can be open even when the accompanying article is paywalled.

Open science links these layers while recognising different ethical, legal and technical conditions.

52. Data cannot always be opened

Privacy, consent, indigenous data governance, commercial confidentiality, security and legal obligations can constrain data sharing.

Responsible open science seeks the most open state compatible with legitimate protection—not openness at any cost.

53. FAIR and open are not identical

FAIR data are Findable, Accessible, Interoperable and Reusable under defined conditions. “Accessible” can include controlled access rather than universal public download.

Open data is one possible access state inside good data stewardship.

54. Public funding strengthens the argument for public access

A common open-access principle is that research funded for public benefit should become publicly accessible as rapidly as practical.

UNESCO’s Recommendation encourages policies that require and reward open access to scientific knowledge, particularly around publicly funded research.

55. Public access does not eliminate publishing labour

Editors, reviewers, copyeditors, designers, platform engineers, metadata staff, preservation services and governance all remain necessary.

The open-access debate is therefore partly about how to finance publishing labour without restricting knowledge unnecessarily.

56. Subscription savings do not automatically become open-access infrastructure

Moving away from subscriptions can free funds, but institutional budgets, publisher contracts and infrastructure needs determine where that money actually goes.

Transition requires deliberate reallocation, not merely cancellation.

57. Libraries are major open-access actors

Libraries fund repositories, negotiate publishing agreements, support diamond journals, advise researchers on rights, manage research data and preserve scholarly outputs.

Modern academic libraries are therefore part of the publishing infrastructure, not only consumers of published material.

58. Universities are becoming publishers and infrastructure owners

Universities host journals, presses, repositories, data platforms and preprint services.

This can reduce dependence on commercial intermediaries while creating new governance responsibilities around editorial independence and sustainability.

59. Open infrastructure needs stable funding

Repositories, DOI systems, metadata registries, journal platforms and preservation networks may look free at the user interface while depending on memberships, institutions and public-good funding behind the scenes.

Open infrastructure is sustainable only when somebody funds maintenance, security, governance and succession.

60. Interoperability makes openness scale

Open APIs, standard metadata, persistent identifiers and repository harvesting allow scholarship to move across discovery systems without repeated manual entry.

The more open the content, the more important machine-readable identity becomes.

61. Search engines are only one discovery route

Researchers discover open scholarship through scholarly databases, repositories, citation networks, library catalogues, journal platforms, preprint servers and increasingly conversational AI systems.

Open access should therefore optimise for durable scholarly metadata, not only web search ranking.

62. Open access improves machine access too

When full text is openly licensed and machine-readable, text mining, systematic review, translation, accessibility tools and AI-assisted discovery become easier to build.

Reuse rights still depend on licence terms and lawful processing.

63. AI turns licensing into infrastructure

Generative AI has made machine reuse of scholarly literature a major policy question. A paper may be readable by humans but carry licence conditions affecting automated reuse.

Current cOAlition S strategy specifically identifies AI, open copyright and data reuse as an area requiring shared policy development during 2026–2030.

64. CC BY is deliberately broad

CC BY allows redistribution and adaptation, including commercial use, with attribution and indication of changes.

This breadth is why many funders prefer it: scholarship can be translated, mined, taught, aggregated and reused without seeking permission each time.

65. Non-commercial restrictions narrow machine and commercial reuse

CC BY-NC allows sharing and adaptation but not commercial use.

The restriction can align with an author’s preferences while creating ambiguity around some educational, platform, publishing and AI uses involving commercial organisations.

66. No-derivatives restrictions reduce adaptation

Licences with ND conditions restrict the distribution of adapted material.

That can affect translations, accessible versions, remixing and certain machine-generated derivatives. Authors should choose licences with an understanding of the trade-off.

67. Third-party content can carry different rights

An open-access article may contain a photograph, map or figure reproduced under separate permission.

A CC licence covering the article does not necessarily grant the same rights over every third-party asset unless the publication states that clearly.

68. Licence metadata should be machine-readable

Displaying “CC BY” in a footer helps humans. Embedding structured licence information helps repositories, indexers and AI systems understand reuse conditions automatically.

Open access becomes more reliable when rights information travels with the object.

69. Corrections remain essential in open access

Openly available errors can spread faster because more people and machines can access the work.

Open publishing therefore needs visible corrections, retractions, version relationships and updated repository links just as strongly as subscription publishing.

70. Retraction should propagate across repositories

If a version of record is retracted while an author manuscript remains in a repository, the repository record should be updated to show the changed publication status.

Open availability should not freeze obsolete trust states.

71. Open access creates a larger correction surface

Copies can exist across repositories, author websites, indexing services and scholarly networks.

Persistent identifiers and canonical version links help corrections propagate across that distributed environment.

72. Metrics change when access barriers fall

Open papers may receive broader readership because more users can reach them without subscription authentication.

Downloads and attention still do not equal quality. Open access expands exposure; evaluation must remain evidence-based.

73. Altmetrics capture attention, not truth

News mentions, social sharing and policy references can reveal forms of impact missed by citation counts.

They can also reflect controversy, sensationalism or coordinated promotion. Metrics require interpretation.

74. Open access can strengthen education

Teachers and students outside research-intensive universities can read scholarship that would otherwise require expensive subscriptions.

Open licences can further support teaching packs, translations and adaptations where permitted.

75. Open access can strengthen public reasoning

Journalists, policymakers, patients, engineers, entrepreneurs, teachers and citizens can inspect research directly rather than relying entirely on secondhand summaries.

Direct access does not make every reader an expert, but it makes expert claims more inspectable.

76. Open access does not eliminate expertise

A technical paper can be open and still require advanced knowledge to interpret responsibly.

The goal is to remove unnecessary access barriers, not to pretend specialised knowledge no longer matters.

77. Public access can expose poor scholarship faster

More readers can identify errors, challenge methods and compare claims against other evidence.

Open access can therefore strengthen postpublication review when correction systems are responsive.

78. Openness without governance can spread misinformation faster

A freely accessible preprint or weak journal article can circulate widely before limitations are understood.

Version labels, peer-review status, corrections and responsible communication are necessary companions to openness.

79. Open-access policy should preserve researcher choice intelligently

Mandates can accelerate system change but can also create administrative complexity when different funders, institutions and journals impose overlapping rules.

Good policy makes compliant routes clear, funds necessary infrastructure and avoids pushing costs onto individual researchers.

80. Research assessment should not reward paywalls

Plan S principles explicitly state that funders should value the intrinsic merit of research rather than the publication channel, journal impact factor or publisher prestige.

This matters because open-access reform is weakened if career systems continue rewarding venue prestige above research quality.

81. Open access is a system design problem

WHO PAYS?
WHO OWNS RIGHTS?
WHO CAN READ?
WHO CAN REUSE?
WHICH VERSION IS OPEN?
WHERE IS IT PRESERVED?
HOW IS IT DISCOVERED?
WHAT HAPPENS AFTER CORRECTION?
CAN THE SYSTEM SURVIVE WITHOUT ONE VENDOR?

The answers determine whether openness is durable, equitable and trustworthy.

82. A practical author route

Check funder and institutional open-access requirements before submission.
Identify whether the journal is fully open, hybrid or subscription.
Check APCs, waivers and other fees.
Read the copyright and licence terms.
Know which manuscript version may be deposited.
Check embargo rules.
Use an appropriate repository.
Preserve DOI and version relationships.
Update repository records after final publication.
Monitor corrections or retractions.

83. A practical journal route

State access and licence policy clearly.
Publish all author fees and waiver rules before submission.
Keep editorial decisions independent of payment.
Use persistent identifiers and structured metadata.
Support long-term preservation.
State peer-review policy transparently.
Maintain corrections and retractions.
Expose machine-readable licence and version information.
Provide repository-friendly metadata.
Review accessibility and multilingual discovery.

84. A practical institution route

Fund repository infrastructure.
Provide rights and licence guidance.
Support APC and BPC funding transparently where used.
Support diamond and community publishing where strategic.
Negotiate sustainable agreements.
Track total publishing expenditure, not only subscriptions.
Preserve research outputs and data responsibly.
Align research assessment with open-science goals.
Train researchers in version and licence literacy.

85. Failure modes

Failure	What breaks
Open access = APC	Diamond and repository routes disappear from view.
Free to read = free to reuse	Copyright and licence are misunderstood.
Repository copy unlabeled	Preprint, AAM and VoR become confused.
Hidden fees	Authors cannot make informed submission decisions.
APC without waiver or alternative	Reader access improves while author inequality grows.
Open journal with weak preservation	Access disappears when platform fails.
Preprint reported as peer-reviewed evidence	Publication state becomes misleading.
Licence absent	Reuse rights become uncertain.
Retraction not propagated to repositories	Old trust state remains open and discoverable.
AI uses open literature without version awareness	Retracted or superseded research can be recycled.

86. The deeper model: open access separates knowledge from the gate

Subscription publishing ties access to the reader’s institutional or personal ability to pay. Open access attempts to break that dependency.

But every removed gate reveals another system question: funding, licence, preservation, discoverability, quality, equity and governance.

REMOVE READER PAYWALL
  → SOLVE PUBLISHING COST
  → PRESERVE QUALITY
  → PRESERVE RIGHTS
  → PRESERVE INFRASTRUCTURE
  → PRESERVE VERSION CONTROL
  → PRESERVE EQUITY
  → KNOWLEDGE BECOMES MORE PUBLIC

Open access succeeds when openness is not a temporary feature of a webpage, but a durable property of the scholarly record.

87. Why open access matters to civilisation

Modern societies invest enormous resources in producing knowledge, yet the people who might benefit from that knowledge are distributed far beyond the universities and institutions able to buy expensive subscriptions.

Open access widens the return path. A teacher can read research. A doctor outside a major university can inspect evidence. An engineer can follow a method. A policymaker can examine the source. A student can learn from the original literature. A machine can connect findings across fields when rights and metadata permit.

The larger goal is not that every publication becomes free by magic. It is that the costs and rights of scholarly communication are designed deliberately enough that knowledge can move farther than the institutions that first produced it.

Current authority routes

Continue the Archives and Publishing series

Publication control: Wintour House · eduKate Publishing · evidence, review, rights, licence, access, metadata, preservation, correction and archive gates.

World Return: When you encounter an “open access” article, inspect the whole route: which version is open, who owns the rights, what licence applies, whether anyone paid to publish, where the work is preserved, and how a future reader will know if it is corrected or retracted.

How Open Access Publishing Works | From Subscription Journals and Article Processing Charges to Repositories, Licences, Preprints and Public Knowledge