NewsCodes Guidelines

Using controlled vocabularies rather than simple keywords allows for a consistent coding of news metadata across news providers and over the course of time. NewsCodes IDs can apply not only to text content but to photographs, graphics, interactives, audio and video files and streams.

All IPTC Controlled Vocabularies are available on our CV server (cv.iptc.org) in multiple formats, both machine-readable and human readable.

For the news industry – but also far beyond – it is required to be able to state something about the content of a news item, to apply so-called metadata. This could be achieved either by free-text human language (e.g. by a headline or a caption) or by codes that represent concepts. Codes have the advantage that they can be shared across providers, and that each code can have a consistent and comprehensive definition, avoiding amiguity and misunderstandings.

NewsCodes are also language agnostic, thus the code is the same for describing content in different languages, only the label and definition of the code need to be translated to help with understanding its meaning.

We have codes representing subjects of news stories, aspects of news production such as contributor type and coverage status, technical terms such as video encoding types and layout orientations, and many vocabularies relating to sports results such as team type, player statistics and position names for many different sports.

A single vocabulary containing all types of possible codes would be unwieldy and inefficient.

Most NewsCodes vocabularies change rarely. The Media Topics vocabulary, however, is regularly updated as we improve labels and definitions,

We don’t expect that users update their systems as soon as a new version is released - we don’t make "breaking changes" such as changing the meaning of a term of re-using a term ID, we only make corrections and clarifications such as improving the label or definition of a term, moving it to a more appropriate parent, or adding a new translation or mapping to an external source such as Wikidata. We also add new terms regularly based on suggestions from members and other interested parties - if you have suggestions for new terms, please contact us!

If we decide to stop using a term, we never remove it from the CV completely, we only mark it as "retired", usually with a "note" describing which term or terms should be used instead.

Updates to the Media Topics (and to all NewsCodes) are announced on the public iptc-newscodes@groups.io discussion list. Joining the public group is free — simply visit https://groups.io/g/iptc-newscodes and sign up.

Yes, they can. Any NewsCode provided by the IPTC can be used at any stage of a news workflow without any royalty fee. But if one includes IPTC NewsCodes in an application, the intellectual property and the copyright of the IPTC must be explicitly included.

Copyright © 2020 IPTC, International Press Telecommunications Council.

The IPTC NewsCodes Guidelines document is published under the Creative Commons Attribution 4.0 license - see the full license agreement at http://creativecommons.org/licenses/by/4.0/

By obtaining, using and/or copying this document, you (the licensee) agree that you have read, understood, and will comply with the terms and conditions of the license.

Materials used in this guide are either in the public domain or are available with the permission of their respective copyright holders. All materials of this IPTC standard covered by copyright shall be licensable at no charge.

This document is the result of a team effort by members of the IPTC NewsCodes Working Group of the International Press Telecommunications Council (IPTC), with input and assistance from other contributors.

Contributors to these guidelines are (in alphabetical order):

Join the public IPTC Newscodes discussion group:

https://groups.io/g/iptc-newscodes/

Submit a message on our website: https://iptc.org/about-iptc/contact-us/

Visit IPTC’s website: https://iptc.org/standards/newscodes/

Follow IPTC on Twitter: @IPTC

The IPTC, based in London, brings together the world’s leading news agencies, publishers and industry vendors. It develops and promotes efficient technical standards to improve the management and exchange of information between content providers, intermediaries and consumers. The standards enable easy, cost-effective and rapid innovation and include the Photo Metadata standard, the Video Metadata Hub, the news exchange formats NewsML-G2, ninjs, SportsML-G2 and NITF, rNews for marking up online news, the rights expression language RightsML, and NewsCodes taxonomies for categorising news.

IPTC is a not-for-profit membership organisation registered in England - find more about membership.

Our contact address is:

Sets of terms used to describe the contents of news items and other media, such as subject (Media Topics and the older Subject Codes), genre to describe types of news item such as obituary, feature, review and sports report, scene to describe types of video shots, and worldregion for regions of the world (for countries and continents, we recommend using use the ISO Country Codes).

Terms used in the production and distribution of news content, such as technical aspects of media (colorspace of an image or codecs of audio and video files), urgency and the code representing a particular news provider.

The transmission NewsCodes are used in the transmission of news content from a news provider (such as a news agency or photo library) to a consumer (such as a newspaper, broadcaster or archive).

For all G2 standards, a mapping of Scheme URIs to aliases is required: we call this a Catalog. The IPTC provides a Catalog file for all the NewsML-G2 controlled vocabularies listed below. Find more information on this file on a special Catalog page.

For G2-Standards a mapping of Scheme-URIs to aliases is required – a so-called Catalog. The IPTC provides a Catalog file for all the NewsML-G2 controlled vocabularies listed below. Find more information on this file on a special Catalog page.

These NewsCodes support specific functionalities of IPTC NewsML version 1.x, the predecessor to NewsML-G2.

The IPTC metadata schemas for photos “IPTC Core” and “IPTC Extension” require a few specific controlled vocabularies:

As the Media Topics are simply codes, labels and definitions in a tree structure (or hierarchy), they can be used in many applications. Most commonly, Media Topics are built in to newsroom content management systems, library systems or syndication tools. By building IPTC Media Topics into editorial systems, vendors and system implementers create a future-proof system for maintaining content descriptions and annotations that can evolve over time as IPTC Media Topics are maintained and extended.

In addition, content tags can be automatically mapped to Wikidata and translated to other languages based on the built-in Media Topics language translations. Currently, IPTC Media Topics are available in Arabic, British English, French, German, Portuguese (Brazilian and for Portugal), Spanish and Swedish.

Generally the more technical NewsCodes CVs are only used when transmitting news content from one news organisation to another using a format such as NewsML-G2 or ninjs. These standards require various aspects of the news content (such as codecs, priorities or editorial staff roles) to be expressed in terms of codes from a controlled vocabulary. This is often built in to newsroom software so individual journalists and editors can simply choose values form drop-down lists and the correct codes will be used automatically.

In 2017, IPTC extended Media Topics through the use of facets which can express multiple aspects of a concept. For example, to represent all Olympic sports medals in MediaTopics CV we would need hundreds of separate terms for each weight category, age group, gender mix etc for different sports. But by using facets, we can create one set of terms for gender with the values "male", "female" and "mixed", which can then be used in conjunction with other terms to identify different sports.

Facet values can be either from a CV (such as the NewsCodes CV "asportfacetvalue") or could be literal strings of text such as "200m".

So for example, the women’s 200 metre breaststroke swimming event can be described with:

For an example of how this can be expressed in NewsML-G2, see the example in the NewsML-G2 example documents repository.

More information on how to represent facets in NewsML-G2 documents is available in the Faceted Concepts section of the NewsML-G2 Guidelines.

All IPTC NewsCodes, including Media Topics, are defined using the W3C’s SKOS vocabulary, which is commonly used to describe taxonomies. We have extended SKOS slightly to add some extra properties to handle retiring concepts and specifying which concepts can have facets. We call our extensions IKOS (IPTC Knowledge Organisation System).

The vocabularies are created using NewsML-G2 Knowledge Items and their built-in concept, conceptId, related and broader properties. Because these are generic concepts, we can convert them into SKOS/IKOS in RDF/XML, RDF/Turtle and JSON-LD formats.

For more information on how concept definitions work in NewsML-G2, see the Concepts and Concept Items section of the NewsML-G2 Guidelines.

The IPTC NewsCodes Working Group often receives requests from agencies and publishers who like the Media Topics vocabulary but also want to extend it for their own use. For example, they might want to add some sports under "competition discipline" which are only played in their country, or some specific festivals and holidays that are not celebrated worldwide and thus wouldn’t pass the guidelines for adding a new term to the global Media Topics vocabulary.

For example, a news organisation specialising in animal stories would have a hard time using Media Topics because we only define one term for "animal" (medtop:20000500).

Using NewsML-G2 Knowledge Items, this organisation could create its own "scheme" that extends Media Topics using its own controlled vocabulary definitions.

This uses the SKOS "broadMatch" property which allows for broader and narrower mappings across "schemes".

As shown in the example below, the definitions of new concepts can also include mappings to Wikidata entities or other external systems.

In NewsML-G2, this is represented as:

In SKOS RDF/Turtle this would be represented as:

Which format you choose will depend on the software you are using to manage your controlled vocabularies.

Please do remember that if you have a suggestion for a term that would apply in multiple countries and passes our criteria for new terms, you can request that the NewsCodes Working Group adds it to the Media Topics vocabulary. All requests are considered.

If your company has its own taxonomy, you might want to create mappings from your terms to the IPTC Media Topics and vice versa, so your customers and partners can index their content in a standard taxonomy and your content will become available to a wider audience.

This is straightforward to do because Media Topics are defined in the W3C’s SKOS format (plus a few extension properties that we call IKOS IPTC Knowledge Organisation System). SKOS has well-defined relation properties: skos:closeMatch, skos:exactMatch, skos:broadMatch and skos:relatedMatch. You should choose the relation type that best suits the particular semantic mapping. That means that you should only use skos:exactMatch if the concepts are indeed exactly equivalent to each other, otherwise skos:closeMatch or skos:relatedMatch may be better.

In SKOS RDF/Turtle, this would be represented as:

The Media Topics vocabulary is by necessity very high level and avoids detail in many areas. The NewsCodes Working Group decided to create some criteria that can be used to decide whether a suggested term should be added to the vocabulary or not. The general guide is that the "granularity", or level of detail, should be roughly even across all branches of the vocabulary.

IPTC delegates discussed the best approach to a policy on level of granularity in Media Topics entries, as change requests are submitted for a variety of granularity levels. Media Topics were created with the intent of well-balanced granularity across the whole vocabulary.

When considering any proposed change, the NewsCodes development group will consider whether the change meets the follow criteria:

The NewsCodes Working Group uses these guidelines as a way of agreeing on a writing style for Media Topics NewsCodes. Different NewsCodes vocabularies may not conform to this guide, and Media Topics in other languages don’t conform to all of these recommendations.

Overall rules:

Rules for term labels:

Rules for term definitions:

The Media Topics have been translated from English into French, German, Spanish, Swedish, Arabic, Chinese, Danish, Portuguese and Brazilian Portuguese.

We welcome offers of help to create translations into other languages.

We have some scripts that can help to get you started with a translation project, and it is always worth getting in touch before you start because we usually have internal working versions of the MediaTopics vocabulary that are not yet released, and it is better to start with the latest working version so that your translations don’t become out of date.

If you are interested in providing a translation into your language, please contact the NewsCodes Working Group via the IPTC Contact Us form.

The first IPTC taxonomy for categorising news was the Subject Codes which has about 1,400 terms (more than 1,000 not counting sport competitions) in a hierarchy limited to three levels. In 2007 the basic requirements of a successor taxonomy - now the Media Topics - were defined:

The first version of the Media Topics taxonomy in 2010 met these requirements at a high level. Since that time, the set of Media Topics has grown and adapted, language has been changed to clarify meaning, make terms more global and update to modern usage and many language translations have been added.