Every time you open an e-journal article, download a government report as a PDF, or click through to a digitised manuscript, a quiet system is working behind the scenes to make that resource findable. That system is metadata. As libraries shifted from shelves of printed books to vast collections of web-based content, the old ways of cataloguing alone could no longer keep pace. Three standards now dominate the work of describing electronic resources: MARC21, Dublin Core, and the Text Encoding Initiative (TEI). Each one solves a different problem, and together they form the backbone of how digital collections are organised, indexed, and preserved. This post walks through how each standard works and where it fits in the everyday practice of resource description.
Table of Contents
- The need for standardised electronic resource description
- MARC21 for electronic resources
- What field 856 does
- Subfields and repeatability
- Dublin Core for web resources
- The fifteen elements
- How it supplements traditional cataloguing
- TEI for digital humanities
- What TEI is
- Why it matters for preservation and scholarship
- How the three standards work together
The need for standardised electronic resource description
The volume of online information has grown beyond what any single library can manually track. A digital collection may hold scanned books, audio lectures, video files, datasets, and born-digital documents, all gathered from many different sources. Without a shared way to describe these items, search and retrieval break down. A user searching for a title would have no reliable way to know whether the resource exists, where it lives, or how to reach it.
Standardised metadata schemes solve this problem. They define a consistent structure for recording who created a resource, what it is about, when it was made, and how to access it. This consistency is what allows different systems to talk to each other. The National Digital Library of India (NDLI), developed at IIT Kharagpur, is a clear example. It harvests metadata from over two hundred sources and faces serious difficulty integrating records that arrive in different formats, with noise and duplication. Its solution was to build a schema combining three globally accepted standards, namely Dublin Core, the Learning Resource Metadata Initiative, and the Electronic Thesis and Dissertation schema, so that more than 87 million items could be indexed and searched through a single window.
The lesson is simple. Standardisation is not bureaucratic box-ticking. It is the precondition for resource discovery across collections, languages, and institutions. The three standards below each contribute to that goal in their own way.
MARC21 for electronic resources
MARC21, short for Machine-Readable Cataloging, is the standard most academic and research libraries use for bibliographic records. It is detailed, structured, and built for library catalogues. When a resource exists online, MARC21 needs a way to record not just what the resource is, but where it sits and how to reach it. That job belongs to field 856.
What field 856 does
Field 856, titled Electronic Location and Access, contains the information required to locate and access an electronic resource. According to the Library of Congress MARC standards, the information in this field is sufficient to transfer a file, subscribe to an electronic journal, or log on to an electronic resource. In practice, it most often holds the URL that links a catalogue record to the actual digital object.
The field uses two indicators to add meaning. The first indicator specifies the access method, such as email, FTP, remote login (Telnet), dial-up, or HTTP. The value 4 signals access through HTTP, which covers most web resources today. The second indicator describes the relationship between the location and the resource the record describes. A value of 0 means the link points to the same resource covered by the record, while 1 indicates it is an electronic version of an item described in the record, as set out in the CONSER cataloguing guidelines.
Subfields and repeatability
Within field 856, subfields carry the specific pieces of access information. Subfield $u records the URL or URI and is the most commonly used element. Other subfields handle the host name ($a), the path ($d), the electronic file name ($f), link text for display ($y), and public notes ($z). The field is repeatable, which matters for electronic resources. As the Library of Congress guidelines explain, a record may carry more than one 856 field when a resource is available by more than one access method, when mirror sites exist, or when different formats live at different URLs.
This repeatability gives MARC21 real flexibility for web content. A single e-journal can be linked through several providers, and a non-working link can even be retained and flagged in subfield $z to preserve clues about former titles or providers. For a library already running a MARC-based catalogue, field 856 lets electronic resources sit alongside printed books in one unified system, which is exactly why it remains central to electronic resource description.
Dublin Core for web resources
MARC21 is powerful but heavy. Its records are complex and demand cataloguing expertise. For the open web, where ordinary content creators and digital repositories need a lighter approach, Dublin Core fills the gap. It was originally developed in 1995 as a core set of elements for describing the content of web pages and enabling their search and retrieval, as the Dublin Core Metadata Initiative records.
The fifteen elements
Simple Dublin Core consists of fifteen elements that capture the most basic descriptive features of a resource. These are Title, Creator, Subject, Description, Publisher, Contributor, Date, Type, Format, Identifier, Source, Language, Relation, Coverage, and Rights. Every element is optional, repeatable, and can be used in any order, which makes the scheme easy to learn and easy to apply. This simplicity is its greatest strength. A resource can be described with a handful of plain attribute-value pairs rather than a complex structured record.
Dublin Core has been formally standardised internationally, including as ISO 15836 and as ANSI/NISO Z39.85, which gives institutions confidence that records will remain interoperable over time.
How it supplements traditional cataloguing
Dublin Core does not replace MARC21. It supplements it. Where MARC21 serves the deep needs of a formal library catalogue, Dublin Core enables broad resource discovery across domains. Its elements can be embedded directly in the head section of an HTML page using meta tags, which helps search engines index and surface online resources, improving discoverability. It is also the baseline format required for sharing records through the Open Archives Initiative Protocol for Metadata Harvesting (OAI-PMH), the mechanism repositories use to expose their metadata for harvesting.
This is why the National Digital Library of India built Dublin Core into its core schema. The simplicity and extensibility of Dublin Core let a repository describe video, audio, images, and text in one consistent vocabulary, then combine those records with other standards when more detail is needed. For digital libraries handling many resource types from many sources, that flexibility is essential. When a project needs greater precision, Qualified Dublin Core adds refinements and encoding schemes to sharpen the meaning of each element without abandoning the simple base.
TEI for digital humanities
MARC21 and Dublin Core both describe a resource from the outside. They tell you about a text without recording what is inside it. For scholars in the humanities, that is not enough. A historical manuscript, a literary work, or an ancient inscription carries meaning in its very structure, its corrections, its marginal notes, and its physical features. Capturing that requires a different kind of standard. This is where the Text Encoding Initiative comes in.
What TEI is
TEI is a set of guidelines for encoding humanities texts using XML, maintained by an international consortium that has operated continuously since the late 1980s. Its purpose, as the TEI Consortium states, is to specify encoding methods for machine-readable texts, chiefly in the humanities, social sciences, and linguistics. Rather than describing how a text looks, TEI describes what its parts mean. A tag marks a paragraph as a paragraph, a name as a person, or a passage as a quotation, so that software can process the text intelligently.
The TEI Guidelines define a large set of XML elements and attributes. Most projects do not use the full range. Instead, as documentation on TEI notes, they create a customised subset suited to their material, and a simplified version called TEI Lite offers a manageable starting point. This customisation mechanism keeps the standard usable across very different kinds of texts.
Why it matters for preservation and scholarship
TEI offers three lasting advantages for scholarly work. It focuses on meaning rather than layout, it is software-independent, and it is supported by an open scholarly community. These qualities make it well suited to long-term digital preservation. A text encoded in TEI is not tied to any single program or display format, so it can be migrated, searched, and analysed for decades. The Women Writers Project at Northeastern University describes how the largest use of TEI is in building digital library collections, often of rare or fragile materials that would otherwise be inaccessible.
For students of literature and history, TEI opens doors that print never could. An encoded edition can record every variant reading of a poem, link a transcription to images of the original page, or expose the deep structure of a text for computational analysis. Tools and databases used widely in research, such as JSTOR and EEBO, draw on TEI principles to power their search functions. In this way TEI does more than preserve. It turns a static text into a resource that can be queried, compared, and studied at scale.
How the three standards work together
It helps to see these standards not as rivals but as a layered toolkit. MARC21 with field 856 connects formal library catalogues to electronic resources and links records to the digital objects themselves. Dublin Core provides a lightweight, interoperable vocabulary for describing web resources and harvesting metadata across repositories. TEI encodes the internal structure and meaning of humanities texts for deep scholarship and preservation. A single large repository may use all three at once, choosing each for the task it does best, just as the National Digital Library of India combines multiple schemas to serve a national audience across many languages and formats.
Understanding when to reach for each standard is a core skill in information processing and retrieval. The choice depends on the resource, the audience, and the depth of description required. A catalogue record, a harvestable web record, and a scholarly digital edition all call for different tools, and these three standards cover that full range.
What do you think? If you were designing the metadata strategy for a new digital archive of regional literature, which standard would you lead with, and why? And as collections grow ever larger, do you think simple schemes like Dublin Core or rich ones like TEI will prove more valuable for the future of discovery and preservation?
References
- https://cacm.acm.org/research/national-digital-library-of-india/
- https://www.loc.gov/marc/bibliographic/bd856.html
- https://www.itsmarc.com/crs/mergedprojects/conser/conser/module_31.15._electronic_location_and_access_(field_856).htm
- https://www.loc.gov/marc/856guide_august1999.html
- https://www.dublincore.org/specifications/dublin-core/dces/
- https://www.iso.org/standard/71341.html
- https://www.techtarget.com/whatis/definition/Dublin-Core
- https://tei-c.org/
- https://en.wikipedia.org/wiki/Text_Encoding_Initiative
- https://wwp.northeastern.edu/outreach/seminars/tei.html
- https://libguides.exeter.ac.uk/digitalhumanities/tei

Leave a Reply