Semantic Web Applications for Danish News Media

Astrid Ildor

Department of Design and Communication, University of Southern Denmark,

Universitetsparken 1, 6000 Kolding, Denmark

Keywords: Semantic Web, Linked Data, Web 3.0, News Media, Journalism, RDF, Semantic Annotation, URI, Web

Ontologies, Application Development.

Abstract: Most news media possess a publish-and-forget mindset: Once a news article is published, the information it

contains devalue in the messy place of the unstructured Web and great potentials of combining and reusing

data is missed. News media has long constituted an area of interest for Semantic Web researchers, but few

studies merge technical knowledge with editorial insights. To fill the gap in literature, this study combines

technical analysis with interviews and Participatory Design studies with eight Danish news journalists and

digital editors. The exploration reveals three areas within the journalistic work process with significant

potential of improvement: Journalists’ challenge of finding the right person to comment on a specific topic,

issues of finding previously published articles, and the need for generating infoboxes. Each area is examined

as a type of Semantic Web application. It is demonstrated how profound annotation of persons, places,

organisations, and key terms mentioned in a body of articles is required for each application. Trustworthiness

is another major challenge as this cannot yet be fully achieved within the concept of Semantic Web.

1 INTRODUCTION

News media publish trustworthy information on an

hourly basis; however, the vast majority of

information is not being archived or annotated for the

purpose of reuse. Great potentials of combining

different datasets or autogenerating new information

based on accessible data cannot be fulfilled as lots of

data – including the content of most news articles – is

not machine-readable. To attain machine-readability,

Semantic Web (SW) reuses the Web’s global

indexing and naming scheme, meaning that every

semantic concept can be annotated a unique identifier

(Domingue et al., 2011) – such data is known as

linked data. Through formal computational

ontologies and ontology-aware technologies,

relationships between concepts published on the SW

can be processed and understood within specific

domains. In this context, SW contains potential as a

distributed reasoning machine that not only can

execute extremely precise searches but also analyse

existing data and create new knowledge (Goddard, &

Byrne, 2010). This can potentially improve the

process of any news journalist and entail innovative

storytelling and knowledge mediation.

However, before media organisations can be expected

to make large investments in archives of linked data,

journalists and other stakeholders must be convinced

that there are costly problems associated with their

current suite of processes and technologies, and that

SW applications can help solve these. On the other

hand, news media has long constituted an area of

interest for SW researchers, but remarkable little

research include insights of editors and journalists. To

fill the gap in literature, this study attempts to answer

the following research question: How can the work

process of news journalists and the user experience

of news journalism be improved in the context of SW?

The problem background is discussed in Section

2 on the basis of a small literature review carried out

as part of the study. The importance of linked data and

Resource Description Framework (RDF) is then

commented in Section 3, before Section 4 sets out the

method of the qualitative study. Qualitative analysis

reveals three areas within Danish news journalism

with significant potential of improvement in the

context of SW. These areas are discussed as three

types of SW applications for the news media industry

in Section 4.1–4.3. Finally, Section 5 sums up the

study in a short discussion and conclusion.

Ildor, A.

Semantic Web Applications for Danish News Media.

DOI: 10.5220/0010124102690276

In Proceedings of the 16th International Conference on Web Information Systems and Technologies (WEBIST 2020), pages 269-276

ISBN: 978-989-758-478-7

269

Figure 1: Semantic Web applications for news media.

2 PROBLEM BACKGROUND

In February 2008, semantic annotation tools reached

a point of development where Berners-Lee and the

W3C declared SW open for business (Miller, 2008).

Coinciding, multiple news media organisations

launched initiatives for annotating their archives of

news articles in order to generate SW applications

(see Figure 1): In 2008, Thomson Reuters launched a

Web service capable of extracting entities and

relationships in text documents and annotating these

with linked data URIs (Hendler et al., 2011). In

cooperation with Rattle Research, the BBC has

developed a similar service to empower applications

such as BBC Wildlife Finder which repurposes data

from Wikipedia, WWF, and the IUCN’s Red List of

Threated Species and combines it with natural world

footage from the BBC archive (Raimond, Scott,

Oliver, Sinclair, & Smethurst, n.d.). Finally, in 2015,

New York Times (N.Y.T.) launched a semi-

automated annotation tool trained to apply semantic

N.Y.T.-resources to plain text (N.Y.T. Labs, 2015).

N.Y.T. has annotated their archive of articles from

1981 to today and on this basis generated multiple

SW applications. A handful other smaller news media

The five examples identified as research tools for

journalists are: Wiki2News and Wiki-excerpt (Rudnik

et al., 2019). Social Semantic Journalism (Heravi &

McGinnis, n.d.). The last two examples are thought-up

scenarios where Semantically-Interlinking Online

Communities (SIOC) are applied to combine data from

organisations and researchers have launched or

described similar SW applications – in total 24 SW

applications or descriptions of thought-up

applications for the news media industry have been

identified as part of a literature review for this study.

Analysis of these demonstrates that none of the

applications seem to be based on systematic empirical

insights of news journalists or publishers.

A majority of the applications are concerned with

content search and new ways of presenting already

published information. Only five

of the 24

applications are identified as research tools to

support journalists’ work processes. In comparison,

all of the applications proposed in this study can be

characterised as research tools which in some way

support the work process of Danish news journalists.

This comparison indicates, that it might be beneficial

to change the objective of SW application

development for the news media industry: Instead of

focusing on how already published information can

be presented, this study demonstrates that potentials

of SW are more likely to be unfolded within news

journalists’ work processes.

different sources about a politician relevant for a

specific news article (Raimond et al., n.d.). In the other

example SIOC is applied to research the term ‘bog-

snorkelling’ across Facebook, Technorati, Flickr, and

YouTube (Meek, 2008).

WEBIST 2020 - 16th International Conference on Web Information Systems and Technologies

270

3 THE CONCEPT OF LINKED

DATA

In order for the Web of linked documents to evolve

into a Web of linked data, Berners-Lee introduced the

Linked Data Principles as a best practice for

publishing structured data on the Web:

1. Use URIs as names for things

2. Use HTTP URIs so that people can look up

those names

3. When someone looks up an URI, provide

useful information, using the standards RDF

and SPARQL

4. Include links to other URIs, so that they can

discover more things (Berners-Lee, 2006)

The first principle advocates using URIs to

identify, not just Web documents, but also real-world

objects and abstract concepts. Unicode in

combination with URI extends support for identifying

any type of resource regardless of its text and

scripting language (Alam et al., 2015).

The second principle advocates the use of HTTP

URIs to identify objects and abstract concepts just

like HTTP is the universal access mechanism for the

traditional Web (Berners-Lee, 2006).

The third principle advocates that HTTP clients

should be able to look up any URI and retrieve a

description of the resource. The agreement of HTML

as the dominant document format has been crucial for

the Web’s ability to scale (Hendler et al., 2011).

Similarly, it is important to agree on a standardised

content format for URI descriptions. According to the

third principle, this format should be Resource

Description Framework (RDF) which is readable for

both humans and machines (Berners-Lee, 2006). It is

common practice to use different URIs to identify

real-world object and the document that describes it.

The fourth principle advocates the use of links to

connect not only Web documents, but any concept

described on the Web (Berners-Lee, 2006). Such

links in a linked data context are called RDF links.

3.1 Resource Description Framework

RDF is a data model for publishing statements on the

Web. Each statement is represented as a triple

consisting of a subject, a predicate, and an object.

The subject is the URI identifying the described

resource. The object can either be a simple literal

value or the URI of another resource that is somehow

related to the subject. The predicate is also identified

by an URI and describes the relationship between

subject and object. RDF provides a data model for

describing resources, but it does not provide any

domain-specific terms for describing classes of

things, and how they might relate. This function is

served by lightweight RDFS ontologies, thus

predicate-URIs come from standard vocabularies or

ontologies. The strength of RDF lies in the flexibility

of integration. RDF graphs can quite easily be merged

by sharing particular resources, or claiming two

resources to be the same, although their identifier

might be different (Domingue et al., 2011).

Resource Description Framework in Attributes

(RDFa) is a serialisation format that embeds RDF

triples in the HTML document, meaning that existing

content within the HTML code can be annotated with

RDFa (Hendler et al., 2011).

RDF and the Linked Data Principles allow for

publishing and accessing simple facts but do not

support more complex queries (Hendler, Heath, &

Bizer, 2011). To retrieve this type of information,

SPARQL Protocol and RDF Query Language

(SPARQL) can be applied. The query language is

designed for evaluating queries against RDF datasets

and to ask meaning-driven questions to databases of

structured data on the Web (Wood, Marsha, Luke, &

Hausenblas, 2013).

3.2 Research on Proof and Trust

In most SW applications, it is fundamental to know

where presented information comes from and how

resulting conclusions have been constructed.

Technologies for including automatic proof checking

are however not yet standardised, and research lacks

to provide answers to several questions.

In SW applications, technologies of unifying logic

operate on top of the ontology to make new inferences

(Pandey & Sanjay, 2010). However, standards to

ensure transparency on how applied ontologies and

reasoning mechanisms are constructed are still

missing and require more research. Without

transparency, biased or manipulated ontologies can

provide answers which cannot be distinguished as

true or false.

Berners-Lee (2006) introduced the concept of

Proof to describe for software agents (or human

users) why they should believe a retrieved result.

Hendler et al. (2011) argues that Proof can be

achieved, if Linked Data Principles are applied to the

dataset itself as metadata, including information

about authorship, currency, license etc. (Hendler et

al., 2011). One mechanism for publishing this type of

metadata is Semantic Sitemaps which are an

extension of the well-established Sitemaps protocol.

In practice, however, semantic metadata are

processed by different loosely coupled systems which

Semantic Web Applications for Danish News Media

271

makes tracking, propagating, and querying difficult

(Jacques et al., 2012).

Finally, the concept of Trust is suggested to be

attained through digital signatures which are

envisioned to check if data really comes from the

claimed and trusted source (Berners-Lee, 2006). This,

however, has not yet progressed far beyond a vision.

As discussed above, current research envisions

how trustworthiness can be achieved, and the

framework for different systems and technologies

have been theoretically outlined. However, practical

solutions have not yet been developed, thus it is not

possible to state that the concept of SW is fully

trustworthy.

4 METHOD AND ANALYSIS

The study attempts to identify issues within the

existing practice of Danish news journalism which

might potentially be improved by SW technologies.

This is examined qualitatively through a series of

interviews with six news journalists and two digital

editors from four of the largest Danish news media

organisations. The same group of journalists and

editors have been invited to engage in multiple

Participatory Design (PD) activities. As a subgenre of

qualitative research, PD aims not only at describing

the social world, but also at contributing to the

improvement of it by inclusion of stakeholders in the

development of new services (Brandt, Binder, &

Sanders, 2013). In this study, activities and

techniques such as The Future Workshop, Scenarios,

The Magic If, and Prototyping (Brandt et al., 2013)

has been applied to explore: How SW technologies

can improve the journalistic practice? And: How the

user experience of Danish news journalism can be

improved in the context of SW?

Transcripts of the qualitative interviews and PD

studies have been categorised inductively after a

grounded theory approach and analysed in the context

of SW. Qualitative analysis reveals three areas within

Danish news journalism with significant potential of

improvement in the context of SW.

4.1 Semantic Archive of Sources and

Contact Details

The first area concerns the challenge of finding the

right person to comment on a specific topic.

Journalists often research in related news articles to

find relevant sources. Such searches are however

difficult to perform as persons are not semantically

related to key terms or organisations. This part of the

journalistic work process can be improved by

implementing a semantic database of sources and

contact details allowing a media’s journalists and

editors to quickly identify relevant sources and their

contacts in relation to specific search terms. The

analysis finds that information about all persons

described in previously published articles need to be

annotated to form a database of already used sources.

This application is referred to as Semantic archive of

sources and contact details. Figure 2 illustrates the

simplest graph needed to build a functional archive of

sources and contact details, where journalists can

easily evaluate a source’s area of expertise

(worksFor, knowsAbout, areaOfExpertise) and

reliability in a specific context (jobTitle,

qualifications). The graph also meets the third Linked

Data Principle about triples that should be included in

a resource’s RDF/XML description. All resources

(represented by circles in Figure 2) must be organised

in a formalised ontology with class hierarchies and

domain-/range- relations. It is recommended to apply

existing

standard vocabularies as resource- and

Figure 2: Minimum graph structure for semantic description of sources and contact details.

WEBIST 2020 - 16th International Conference on Web Information Systems and Technologies

272

property- URIs can then be reused together with the

underlying ontology of the standard vocabulary

(Berners-Lee, 2006). In the example illustrated in

Figure 2, the standard vocabulary Schema.org

applied. Most standard vocabularies include

resources phrased only in English which might entail

a language issue when used for annotating news

articles in Danish. This language differentiation

presents a weakness of semantic annotation as it can

be hard to integrate and might cause inconsistency or

missing links. Thus, it is recommended to create new

resources in Danish in a media’s controlled

vocabulary and link these to equivalent English

resources – if possible, in existing standard voca-

bularies – via owl:sameAs-relations. This ensures that

any Danish resource becomes part of a global graph

and allows standardised software to be applied.

Semantic annotation can be implemented in each

news article using RDFa. This is however a very

time-consuming process and requires basic

knowledge of HTML and the RDFa syntax which

journalists are not expected to possess. Thus, it is

necessary to develop and implement systems that can

automatically or semi-automatically write the

annotation. In 2008, Thomson Reuters launched the

linked data entity extractor Calais which is a Web

service capable of annotating documents with URIs

of places, people, and organisations mentioned in

unstructured text such as news articles (Hendler,

Heath, & Bizer, 2011). Today, Calais is integrated in

the latest version of Drupal which is the content

management system (CMS) used by multiple Danish

news media organisations.

To reach a critical number of annotated concepts,

the established SW community

emphasises large-

scale cooperation and user contributions. Same

approach needs to be applied within news media

organisations: Every journalist must contribute and

add pieces of semantic information when she comes

across new or updated data. Thus, this type of

application is not suitable for smaller organisations.

In fact, it is recommended that multiple Danish news

media organisations join forces and build a database

of sources and contact details together.

Once semantic annotation is applied, it is possible

to query sources and their contact details related to

specific organisations, persons, or topics using

SPARQL-queries. Journalists are not trained in

constructing such queries, and it is recommended to

Schema.org was launched in 2011 by Google, Yahoo, and

Bing as a standard for semantic mark-up of web pages

(Bradley, 2013). Since then, Schema.org has grown to

become one of the most popular standard vocabularies.

develop a search panel and design a user interface to

guide the construction and formatting of these. Figure

3 illustrates how a search panel can be constructed.

The search panel suggests three fixed search

parameters (the properties: areaOfExpertise,

worksFor, and qualifications) to be matched with

different search terms (values). It is possible to type

each search term and choose between suggested

predefined terms from a media’s controlled

vocabulary. These appear from a drop-down list as

the user starts typing. The user interface ensures that

the combination of properties and values are matched

and formatted correctly following the SPARQL

syntax. The technical implementation of this

formatting can be performed using the library

SPARQL Lib. This type of application raises

questions about privacy and GDPR regulations which

should be examined and discussed separately.

Figure 3: Search panel for sources and contact details.

The Linked Open Data Cloud (https://lod-cloud.net/#)

keeps track of how many datasets have been published

in the linked data format. The organisation also

contributes to conferences and advisory boards.

Semantic Web Applications for Danish News Media

273

4.2 Internal Semantic News Article

The second area of interest concerns issues of finding

previously published articles related to a specific

concept. When news break, journalists often search

for context, and sources in related news articles

published by their own media or by others. Journalists

and editors however experience inconsistency and

limitations in standard search engines such as Google,

e.g. it can be difficult to search for articles published

a long time ago. This process can be empowered by

semantic annotation: If all persons, organisations,

places, and key terms described in a media’s archive

of articles are semantically annotated, it is possible to

perform thorough and complex search queries. This

application is referred to as Internal Semantic News

Article Search. This type of application requires

concepts to be annotated as RDF triples the same way

as for the semantic archive of sources and contact

details. This annotation is however much more

comprehensive and requires use of multiple

vocabularies to describe relevant relations between

different domains. Furthermore, it is important to

distinguish real-world objects from the HTML

document (the news article) that describe these, and it

is recommended to annotate not only concepts and

domains but also metadata about the Web document

and its relation to the content (Hendler et al., 2011).

The rNews ontology

can be applied to annotate

information about author, publication data, and

thumbnail-URL. As an example, this allows users of

the semantic news article search application to search

for concepts described by a specific journalist or

written within a specific timeframe.

In order for journalists to apply semantic RDFa

annotation as they write and publish news articles, a

robust annotation system – preferably integrated into

the CMS – is needed. However, even with an

annotation system such as N.Y.T.’s Editor

, the

process of semantically describing an entire news

article requires more time, than journalists spend on

tagging articles today. Also, current technology still

requires manual editing to secure quality mark-up,

and it should be considered whether this is a task for

journalists, editors, or maybe even dedicated mark-up

specialist to do. These issues are highly relevant and

The vocabulary is designed by the International Press

Telecommunications Council (IPTC) to ensure

consistency in how news media annotate metadata.

In 2015, New York Times launched Editor which a semi-

automated tool for annotating news articles with

semantic information. The tool comprises a simple text

worth a deeper study with journalists and other

stakeholders.

Once all concepts are semantically annotated, it is

possible to develop an application to perform

complex and precise searches within a media’s

database of linked data. In fact, such an application

already exists: N.Y.T. has created a semantic search

application

for their archive of articles – consisting

of news articles from 1981 N.Y.T. Developer (n.d.).

The application is designed as an API for public use

over the HTTP and is used as a case study in this

exploration.

Apart from supporting the research phase, some

participants suggest that an internal semantic news

article search application can be used for additional

purposes: The function can be used to insert links to

related articles in a specific news article. This is

something Danish journalists currently do manually

because existing technologies are not good enough.

An automation of this process saves time and enables

new links – which are not present to the journalist –

to be added. Participants also suggest that internal

semantic news article search can be used to analyse

how minorities are represented and used as sources in

the overall news coverage. This type of analysis can

be applied as a tool for the media to shed light on

biases or specific discourses, but it can also provide

research for new articles worth reading for the public.

The sections above describe how a semantic

search application can be implemented for a media’s

own archive of articles. Journalists however often

research in other media’s archives too, and it would

increase usability remarkably, if not only one Danish

news media organisations, but all news organisations

world-wide decided to semantically annotate their

articles and provide open API search tools.

4.3 Semantic Infobox: Summary

The third area of interest targets improvement of user

experiences and concerns the issue of adding

encyclopaedic information in a short amount of time.

The analysis finds that Danish news articles do not

deploy online opportunities of providing additional

information, and that Danish news journalists are not

willing to spend more time than absolutely necessary

writing additional, encyclopaedic information

editor, supported by a set of networked microservices

that are trained to apply specialised N.Y.T. resources to

text documents.

https://developer.nytimes.com/docs/semantic-api-prod-

uct/1/overview

WEBIST 2020 - 16th International Conference on Web Information Systems and Technologies

274

Figure 4: Interface for adding autogenerated summaries.

about persons or concepts mentioned in news articles.

On the other hand, it can be concluded that such

additional information contributes to transparency

and ultimately supports a media’s trustworthiness.

To ensure the quality of information, it is

recommended to focus on one type of infobox:

Autogenerated summaries – and as a starting point

design the application to rely exclusively on

information from a media’s own database of

annotated information. This application is referred to

as Semantic Infobox: Summary.

The analysis finds that autogenerated summaries

must include information about a person’s name, job

title, workplace, educational background, and

seniority. This way integrated summaries can

strengthen a media’s trustworthiness as they

document why professors or other authorities are

chosen as expert sources. Additionally, summaries

should include personal information such as date of

birth, family relations, and information about

previous jobs and memberships. This allows readers

of the news article to easily recap information about

the person. Ultimately, news journalists can also use

the application as a research tool. For this type of

application, it is important that all string values are

phrased in Danish, and that all resources describing

anything else than names are linked to resources

phrased in Danish via owl:sameAs-relations.

Similarly, all properties need to be linked to

equivalent properties in Danish. This allows the

autogenerated summary to be displayed with both

properties and values phrased in Danish.

An application for autogenerated summaries

differs from the two previous solutions as it not only

support research but contributes with information

displayed directly as part of the news article. This

highly increases requirements for reliable

information as one incorrect information contained in

an infobox might affect the trustworthiness of all

other content on the media’s platform. Possibilities

for fact-checking can be provided as metadata such as

when the summary was last updated, and what

sources the data is collected from. This kind of

transparency and traceability can be implemented

using Semantic Sitemaps which allows metadata to be

included in the RDF/XML descriptions.

Figure 4 demonstrates how a user interface –

integrated in the CMS – for implementing and fact-

checking autogenerated summaries can be designed.

The journalist simply highlights a person’s name and

uses the middel panel to create a summary. A display

then appears showing all relevant information about

that person. For each line of information, the journalist

is provided with options to edit data or inspect the data

source. This way, the journalist can review when a

person’s current job title was last updated or delete

outdated information. It is also possible for the

journalist to add new information using predefined

properties and predefined or typed values. Technically,

summaries can be queried and formatted using

SPARQL-queries and SPARQL Lib the same way as

results are retrieved and displayed for the semantic

archive of sources and contact detail application.

Autogenerated summaries as described in the

sections above rely exclusively on linked data from a

media’s own archive of news articles. This guarantees

journalists and users that information contained in the

summary at some stage has been fact-checked and

edited by a journalist, but it also means that the

application can only generate summaries for persons

Semantic Web Applications for Danish News Media

275

previously mentioned by the media. In order to truly

support news journalists, and to truly take advantage

of the concept of SW the application should be

extended to rely also on external sources of linked

data. It is recommended to develop some kind of

certification to guarantee that the datasets live up to

current GDPR regulations and are maintained and up

to date, e.g. guaranteeing that the dataset is reviewed

at least once a month, and that a person registered

with name and contact details is responsible for this.

Establishing national or even international standards

for the quality and maintenance of linked data allows

media organisations – and other organisations – to

share and reuse encyclopaedic data from each other.

The analysis demonstrates that news journalists

across Danish news media organisations work in very

similar ways and with extended focus on research and

fact-checking. These conventions might ease the

process of defining a set of standards for summary

data, but legal and practical implementation of such

certification requires further research.

5 DISCUSSION AND

CONCLUSIONS

The core intention of SW is to make content machine-

readable in order to improve findability and enable

knowledge- and context-based information to be

generated. This way semantic annotation can be seen

as pure preparation for AI agents. However, this study

demonstrates that AI in the context of Danish news

media still seems distant. Remarkably little research

explores aspects of reliability and objectivity within

the field of SW even though major challenges on how

to secure trustworthy information need to be solved.

As demonstrated in this study, these challenges need

to be solved not only locally, but as standards or

certifications agreed upon by all trustworthy linked

data providers. Another area which needs attention in

this context is the concept of neutral ontologies

(Uschold and Gruninger, 2004). The concept of

ontologies presupposes that the entire world can be

objectively categorised. However, several current

debates illustrate how this – in reality – is often

negotiable or political dependent. Thus, those in

charge of annotating information and constructing

underlying ontologies can easily influence the way

we understand our surroundings which has proved to

be extremely powerful. In contrast to the power of

current media, this impact is much less visible and

harder to trace and should be an area of extreme

interest for researchers in the field of SW and in the

field of journalism. In the context of Danish news

media, this paper has, in a practical manner, explored

potentials of SW technologies exposing three areas

where these technologies can be used to make

significant improvement. Before this can be realised,

basic research on implementation of annotation tools

and standards for trustworthiness is still needed.

REFERENCES

Alam, B., Birbeck, M., McCarron, S., Herman, I. (2015).

Syntax and processing rules for embedding RDF

through attributes.

Berners-Lee, T. (2006). Linked Data. Retrieved from:

http://www.w3.org/Designissues/LinkedData.html

Brandt, E., Binder, T., Sanders, E.B.N. (2013). Tools and

techniques: Ways to engage telling, making and

enacting. In Routledge International Handbook of

Participatory Design (p.145–182). Routledge.

Domingue, J., Fensel, D., Hendler, J. (2011). Handbook of

Semantic Web Technologies. Berlin: Springer, vol. 1.

Goddard, L., Byrne, G. (2010). Linked Data tools: Semantic

Web for the masses. In First Monday, 15(11).

Hendler, J., Heath, T., Bizer, C. (2011). Synthesis Lectures

on the Semantic Web: Theory and Technology. Linked

Data evolving the Web into a Global Data Space.

Morgan & Claypool Publishers.

Heravi, B., McGinnis, J. (n.d.). A Framework for Social

Semantic Journalism.

Jacques, Y., Anibaldi, S., Celli, F., Subirats, I., Stellato, A.,

Keizer, J. (2012). Proof and Trust in the OpenAGRIS

Implementation. Int’l Conf. on Dublin Core and

Metadata Applications.

Miller, P. (2008). Sir Tim Berners-Lee: Semantic Web is

open for business. Retrieved from:

http://blogs.zdnet.com/semantic-web/

N.Y.T. Developer (n.d.). Semantic API. Retrieved from:

https://developer.nytimes.com/docs/semantic-api-

product/1/overview

N.Y.T Labs (2015). Editor (2015). Retrieved from:

https://nytlabs.com/projects/editor.html

Pandey, R., Sanjay, D. (2010). Interoperability between

Semantic Web Layers: A Communicating Agent

Approach. In International Journal of Computer

Application, 12(3).

Raimond, Y., Scott, T., Oliver, S., Sinclair, P., & Smethurst,

M. (n.d.). Use of Semantic Web technologies on the

BBC Web Sites.

Rudnik, C., Ehrhart, T., Ferret, O., Teyssou, D., Troncy, R.,

Tannier, X. (2019). Searching News Articles Using an

Event Knowledge Graph Leveraged by Wikidata. 2019

IW3C2.

Uschold, M., & Gruninger, M. (2004). Ontologies and

semantics for seamless connectivity. In SIGMOD,

33(4), pp. 58-64.

Wood, D., Marsha, Z., Luke, R., Hausenblas, M. (2013).

Structured Data on the Web. Manning Publications.

WEBIST 2020 - 16th International Conference on Web Information Systems and Technologies

276