Case Study · Scholarly Publishing

From Fragmented Archive to Editorial Platform

A complete rebuild of the scholarly journal platform: reconstruction of a ten-year archive from fragmented sources, bilingual architecture, a custom Payload CMS, structured search, and standardized scholarly metadata.

Migration

Database

Localization

Payload CMS

Cloudflare

Educational Discourse scholarly journal website presented on desktop and mobile devices

Project Snapshot

Field

Academic publishing

Archive period

2017–2026

Project type

Rebuild and data migration

Languages

Ukrainian and English

Infrastructure

Next.js, Payload CMS, Cloudflare Workers, D1, R2

Introduction

A scholarly journal is far more than a collection of published pages

Educational Discourse already had a decade of published research. What it did not have was a single reliable system connecting the public website, article files, issue records, author information, translations, and external scholarly indexes.

The archive had accumulated across a Joomla website, an older OJS installation, SQL databases, server directories, PDF files, and Crossref records. Each source preserved a different part of the publication history, but none contained the complete picture.

The project therefore began with the archive rather than the interface.

Inside the reconstruction

How every piececame together

Follow the reconstruction step by step and see how each decision shaped the final result.

We rebuilt the journal one layer at a time - from restoring its publication structure to giving editors and readers a clear way to work with it. Each step solved a specific problem and made the next one possible.

1
Archive analysis
The content existed. The system around it did not
The legacy website could not be migrated as a set of pages. Publication data had been added inconsistently over the years, without a shared structure.
We reconstructed the archive by cross-checking every issue and article against databases, PDF files, server files, and Crossref records, identifying and completing missing records and fields, and restoring the correct relationships.
No consistent issue structure
Different naming conventions
Articles disconnected from issues
Files stored separately from metadata
No reliable archive search
Legacy URLs embedded in external records
This work turned disconnected source material into a consistent publication model.
2
Data migration
A ten-year archive moved as complete publication records
The migration covered 59 issues and 575 articles published between 2017 and 2026. Publications moved together with their authors, language versions, identifiers, and original files.
Each article was placed in its correct issue and retained its available Ukrainian and English content, author information, DOI, and PDF.
59
Issues
575
Articles
821
Author credits
634
PDF files
574
DOI identifiers
2663
Linked records
Migration principle
Context
Metadata
File
One record
We migrated publications, not files
3
Editorial CMS
Future publishing was placed in the editors’ hands
Once the archive had been reconstructed, the next objective was operational: give the editorial team a reliable way to maintain it without developer assistance.
Payload CMS was configured around the way the journal actually prepares a publication. Contextual guidance is built into the interface, explaining how to complete fields, attach files, and enter identifiers correctly. Where possible, values are selected from convenient dropdown lists instead of being entered manually each time.
Publishing an article
01Select the appropriate issue
02Enter the publication details
03Complete both language versions
04Add and connect the authors
05Attach the PDF
06Review the public record
07Publish with the completed issue
Beyond publications
The same workspace also lets the editorial team update navigation, policies, contacts, and every other section of the website, including the logo, favicon, and social sharing image.
One workspace for publishing the journal and maintaining the website
4
Archive Search
A decade of research. One clear way to find it.
The reconstructed archive needed to be useful not only to the editorial team, but also to readers. Finding an article should not require knowing its exact title, publication year, or issue number.
A single search field now covers article titles, author names, abstracts, keywords, and DOI values. Readers can search with the information they already have and then narrow the results by year or issue.
Before
How search works
One field
Titles, authors, abstracts, keywords, and DOI values are searched together.
Closer matches first
Complete phrases and closely related words appear before scattered matches.
Filters stay available
Readers can narrow the archive by year and issue without starting over.
Readers search by what they know – not by how the archive is stored
After
5
Publication records
Every article became part of a connected scholarly system
Finding an article is only useful when the result leads to a page that can be read, identified, cited, and verified. Each publication page now brings its bibliographic context, research content, files, and navigation together in one place.
DOI and ORCID identifiers lead directly to the corresponding scholarly records, while internal navigation connects every article to its issue and to the other articles published alongside it.
Connected routes
DOI and ORCID
Direct links open the corresponding external scholarly records.
Article to issue
Readers can return to the complete issue without restarting their search.
Article to article
Previous and next navigation continues through the issue in publication order.
Before
What readers had to work with
Ukrainian and English versions overlapped
Article pages broke at desktop width
Metadata had no clear hierarchy
DOI and issue details were hidden inside PDFs
The PDF link looked like a decorative image
PDFs opened inside the page and broke its layout
ORCID appeared as a raw URL
Articles formed one continuous list and had to be found by paging through it manually
After
Anatomy of a record
Identity
Title, complete author information, DOI, and linked ORCID identifiers.
Publication context
Volume, issue, year, page range, and received, accepted, and publication dates.
Research content
Abstract, keywords, formatted bibliography, and recommended citation.
Access
The article PDF and its complete issue remain directly available from the record.
The PDF is the document. The page is the scholarly record.
6
External Metadata
The website moved. Its scholarly identity had to move with it
Moving the journal to a new domain did not update the external systems that identify and index its publications. DOI records still pointed to the previous website, while DOAJ contained only part of the archive.
We matched the reconstructed publication records with their external counterparts, verified the new destinations, corrected legacy discrepancies, and prepared complete updates for both registries.
One publication. Three connected records
Journal website
Article page, publication metadata, and PDF
Crossref
DOI and the address to which it resolves
DOAJ
Publication metadata used for indexing and discovery
Crossref
MatchedVerifiedUpdated
563
DOI records updated
562
Unique pages checked
100%
Successful responses
Incorrect and shifted DOI assignments were checked against article titles, issue structure, page ranges, and existing Crossref data before correction.
DOAJ
Backed upCompletedValidated
289
Records backed up
575
Complete article records
XML
Replacement validated
Existing DOAJ records and newly recovered publications were combined into one complete replacement dataset for the next registry update.
The archive and the external indexes now describe the same publications.
7
Technical Architecture
The journal kept its domain. Everything behind it was rebuilt
The journal remained on its established domain, preserving continuity for readers, citations, search engines, and external scholarly records. Behind the same public address, the platform was completely rebuilt on Cloudflare.
Cloudflare now stands between public traffic and the application, providing DNS, encrypted connections, edge delivery, and network-level DDoS protection before requests reach the journal platform.
Encrypted HTTPS connections
Network-level DDoS protection
Fast loading for readers around the world
Stable operation during traffic spikes
journal-discourse.com
Established domain retained
The established public address was retained
Citation continuity was preserved
Search history remained connected to the domain
External scholarly records still identify the same journal
The address stayed the same. The system behind it became protected, structured, and maintainable.
Platform structure
Cloudflare Workers + OpenNext
Runs and delivers the application
Next.js + next-intl
Public website, archive, search, article pages, and localized routes
Payload CMS
One editorial source for issues, articles, authors, translations, and website content
Cloudflare D1
Structured publication metadata and editorial relationships
Cloudflare R2
Article PDFs, complete issues, and website media
8
Design system
The system was rebuilt. The design gave it a coherent public face
Once the archive, editorial workflows, search, publication records, and infrastructure had been brought into one system, the final task was to give that system a clear and consistent public form.
The interface takes its character from academic publishing without reproducing the limitations of a printed journal. Deep navy provides structure, warm brass marks important actions, and calm ivory surfaces support long-form reading.
Navy
Structure
Brass
Emphasis
Ivory
Reading surface
Editorial typography
Academic character without imitating print
Measured spacing, clear hierarchy, and restrained typography keep long abstracts, bibliographies, and publication metadata comfortable to scan.
Interface language
← Back to archive
2025Issue 52No. 1–2
Author information
Mariia Umryk
ORCID: 0000-0002-0396-0045
One visual language across the archive, article records, editorial content, and both locales.
Mobile reading flow
Not a compressed desktop
Content reorganizes for the screen width instead of shrinking the entire page
One reading column
The article, metadata, and author information follow a clear sequence
Hierarchy retained
Navigation, citation, and document access remain in their expected places
PDF outside the layout
The document opens separately instead of breaking the mobile page
Structure made the journal reliable. Design made that reliability visible.
Before
After
9
Result
Ten years of publishing becamea system built to continue
The project preserved the journal’s publication history while changing how that history is organized, accessed, and maintained. Issues, articles, authors, files, translations, and identifiers now exist as parts of one connected bilingual system.
Readers received a complete and searchable scientific archive. Editors received a practical workflow for publishing what comes next without rebuilding the system around every new issue.
/01
10
Years
/02
59
Issues
/03
575
Articles
/04
2,663
Records
External records aligned
563
Crossref records updated
575
DOAJ records validated
The journal kept its history and gained a way forward
Preserved
/01
Ten years of publication history
Article and issue PDFs
DOI, author, and language relationships
Established domain and its search authority
Ready to continue
/02
New issues and articles
Bilingual editorial workflow
Website and external metadata maintenance
Independent updates without developer assistance

WHEN CONTENT OUTGROWS THE SYSTEM

Trapped in a legacy platform
that accumulated problems for years?

We preserve the value of accumulated content, bring order to complex archives, and turn outdated websites into platforms teams can manage without constant developer support. Data migration, a practical CMS, multilingual publishing, and reliable infrastructure come together in one system built to support the organization’s next stage of growth.

✦Content & Data Migration

✦Custom CMS Development

✦Multilingual Platforms

✦Cloudflare Architecture

Discuss Your Project

Visit Educational Discourse Explore More Work

From Fragmented Archive to Editorial Platform

A scholarly journal is far more than a collection of published pages

How every piececame together

Archive analysis

The content existed. The system around it did not

Data migration

A ten-year archive moved as complete publication records

Editorial CMS

Future publishing was placed in the editors’ hands

Archive Search

A decade of research. One clear way to find it.

Publication records

Every article became part of a connected scholarly system

External Metadata

The website moved. Its scholarly identity had to move with it

Technical Architecture

The journal kept its domain. Everything behind it was rebuilt

Design system

The system was rebuilt. The design gave it a coherent public face

Result

Ten years of publishing becamea system built to continue

Trapped in a legacy platformthat accumulated problems for years?

Quick links

Services

Contacts

Trapped in a legacy platform
that accumulated problems for years?