Preliminary Exploration of Using RDF

for Annotating DITA Topics on Agroforestry

Thomas Zschocke1 and Sissi Closs2,3

1
World Agroforestry Centre, PO Box 30677, Nairobi 00100, Kenya
[Link]@[Link]
2
Karlsruhe University of Applied Sciences, Postfach 2440, 76012 Karlsruhe, Germany
3
Comet Computer GmbH, Rückertstrasse 5, 80336 München, Germany
closs@[Link]

Abstract. Collections of topics composed with the Darwin Information

Typing Architecture (DITA) depend on annotations for improved search
and retrieval. Through a set of metadata capabilities embedded in the
DITA markup authors can define controlled values to identify and classify
the subject matter of the content as well as to express the hierarchy and
relationships between DITA elements and non-DITA resources. However,
while these mechanisms typically provide both structural and semantic
markup of DITA topics, it is difficult to manage, extend, and integrate
a growing volume of DITA-based content and make them available for
more intelligent Semantic Web services. Rather, the search and retrieval
of DITA topics can benefit if combined with annotations captured in the
Resource Description Framework (RDF). The paper addresses the issue
of making the semantics of DITA XML documents explicit by using
RDF for annotating existing documents. It reviews options for lifting
DITA XML data into RDF for ease of processing. The paper shows that
enriching DITA topics with semantic annotations helps to make DITA
content more comprehensible and accessible, and improves the semantic
interoperability among DITA topic instances. It concludes with general
observations and an outlook on future work on exploiting the mapping
and linking of DITA topics with RDF for improved sharing of data across
collections as well as Linked Data.

Keywords: DITA XML, RDF, semantic annotation, metadata,

AGROVOC.

1 Introduction
The Semantic Web envisions enhanced structure to a Web of data and infor-
mation of shared semantics and data integration through well-defined meaning
based on a "ladder of authority" as a sequence of specifications of Web standards
[28]. Two of the important components for developing the Semantic Web are the
eXtensible Markup Language (XML) and the Resource Description Framework
(RDF) [4]. XML allows authors to use tags (or labels) to annotate Web pages
or sections of text on a page, but without a fixed vocabulary. RDF represents a

J.M. Dodero, M. Palomo-Duarte, P. Karampiperis (Eds.): MTSR 2012, CCIS 343, pp. 276–288, 2012.
c Springer-Verlag Berlin Heidelberg 2012
 Preliminary Exploration of Using RDF for Annotating DITA Topics 277

minimalist knowledge representation for the Web in terms of making assertions

about objects with certain properties and values [9].
In general, RDF was designed to standardize the definition and use of meta-
data in terms of a basic object-attribute-value data model. Metadata is impor-
tant for both resource discovery to locate and access resources as well as for
resource management to facilitate the effective use of structured documents. In
the context of XML, metadata can be both captured within the XML markup
as well as or stored in XML-encoded documents [23]. RDF, instead, can express
meaning encoded in a set of triples, which in turn can be written using XML tags.
In addition, RDF provides more appropriate mechanisms for knowledge repre-
sentation on the Web by applying ontology representation languages [4]. When
expressed in the RDF/XML format, resources can be described in various alter-
native ways, which is especially relevant when the discovery and understanding
of these resources would benefit from additional interrelated descriptive infor-
mation. RDF enables the encoding, exchange, and reuse of structured metadata,
which can be defined as needed, and supports metadata interoperability.
Even though XML as a format for metadata can also enhance the interoper-
ability of metadata resources, RDF is more suitable for semantic interoperabil-
ity by better defining mappings between terms with the data and making them
understandable. Semantic interoperability is defined as the ability of different
agents, services, and applications to exchange data, information and knowledge
as a prerequisite to manage, share, and retrieve such information effectively in an
networked environment [36]. While XML only aims at the document and does not
provide a general interpretation of a document’s data, RDF’s object-attribute
structure provides the natural semantics because all objects are treated as inde-
pendent entities. They can be more easily mapped between two RDF transcrip-
tions without additional translation steps that would be needed in XML. The
advantage of RDF for semantic interoperability is that its object-attribute struc-
ture provides semantic units that can be used to constitute a domain model for
data interchange, which increases the level of potential reuse of components [4].
The Darwin Information Typing Architecture (DITA) is an XML markup
language designed for generating topic-oriented, information-typed content that
can be reused and single-sourced in a variety of ways [7], [22]. DITA supports
metadata in its basic building blocks, that is, topics and maps. The semantic
markup in DITA provides intelligence about its content. However, these com-
mon textual annotations about the content of the documents are limited in
scope and mostly intended for content creators. Instead, viewing DITA docu-
ments as "intelligent documents" [32] would be more beneficial by annotating
document content based on well-defined semantics from domain ontologies to
identify concepts and relations between concepts in documents. Semantic an-
notations provide enhanced information retrieval through the improved ability
to perform searches, and increased interoperability, for instance, by providing a
framework based on a common ontology to integrate information from different
sources in different formats that do not easily interact.
278 T. Zschocke and S. Closs

When considering a large store of DITA topics, it becomes expensive to search

the entire store sequentially to locate the content information needed. While
DITA embeds metadata to describe its content in the sense of semantic hinting,
it would be more beneficial to make this information explicit through semantic
lifting, that is, by extracting the knowledge in terms of semantic annotation
from the stored content [8]. By aggregating descriptions from both metadata
and the XML markup through RDF, new documents can be created to indicate
the component elements as well as the possibilities of combining and recombining
relations for the original or new resources, which would lead to more consistency
and enhanced interoperability. The advantage of utilizing markup languages such
as DITA and metadata description frameworks like RDF allows the development
of more robust mechanisms for linking many different document instances in
a particular collection into a complex associated web [29], [35]. This issue is
particularly important when considering that DITA documents are composed of
single topics that can be aggregated to larger entities and collected in an even
larger pool.
This paper explores the notion of semantic annotation as applied to DITA
topics. The first section presents a general overview of handling metadata in
DITA. Then, the paper introduces the notion of semantic annotation and se-
mantic lifting to RDF. In the fifth section it illustrates the lifting of semantic
information from DITA to RDF using a set of DITA topics on agroforestry. The
final section provides some concluding remarks and an outlook of future work.

2 Metadata Handling in DITA

In general, metadata contains information to describe, provide context, indicate

the quality, or to document other relevant characteristics of any document or
object in both digital and conventional formats [10]. In DITA, metadata is em-
bedded into the markup and can be applied through the <metadata> element
in both DITA topics (<prolog> elements) and DITA maps (<topicmeta> ele-
ment). DITA also provides support for learning metadata, <lcLom>, based on
a subset of the IEEE standard for learning object metadata (LOM) [11]. DITA
supports the three broad functional categories metadata, that is, descriptive,
structural, and administrative [19] (see Table 1). Elements inside of <metadata>
provide information about the content subject of a topic (descriptive metadata);
prolog elements outside of <metadata> provide lifecylce information for the con-
tent unit (administrative or provenance metadata, including rights management
metadata and preservation metadata).
DITA contains additional elements that correspond with Dublin Core, but
are not strictly handled as metadata. For instance, the DITA <desc> contains
the description of the current element. Similarly, the short description (<short-
desc>) element represents the purpose or theme of the topic. This is similar
to the [Link] element. Or, the DITA <title> element contains a head-
ing or label for the main part of a topic; it can also be used to provide a title
for a map. This is similar to the [Link] element. The @xml:lang metadata
 Preliminary Exploration of Using RDF for Annotating DITA Topics 279

Table 1. Examples of metadata types, DITA metadata elements, and equivalent Dublin
Core elements

Metadata Function DITA metadata Dublin Core

types elements elements
Descriptive metadata Resource discovery and identification
author [Link]
publisher [Link]
coverage [Link]
[Link]
Administrative metadata Resource management (Preservation)
Technical metadata Processing
Provenance metadata Lifecyle author [Link]
management critdates [Link]
resourceid [Link]fier
Rights management Use and copyright [Link]
metadata access control permissions
Structural metadata Storage and presentation of resources (composition)
critdates [Link]
subject scheme [Link]
map
classification domain
elements

attribute corresponds with the [Link] element to indicate the language of

the resource.
The [Link] element defines an unambiguous reference to a related re-
source within a given context with the following qualifiers: IsPartOf, IsFormatOf,
IsVersionOf, References, IsBasedOn, and Requires. The most flexible facility
to manage relationships in DITA are the classification elements implemented
through (1) a subject scheme and (2) the subject classification domain. The
specialized DITA map, <subjectScheme>, is used to create custom controlled
values and to manage metadata attribute values. A subject scheme can be en-
hanced to form a taxonomy using the same core elements and specify a binding
between a category and its subject that enumerates the values of a metadata at-
tribute. Similar to the [Link] qualifiers, the <hasNarrower>, <hasPart>,
<hasInstance>, <hasKind>, and <hasRelated> elements specify the kind of
relationship in a hierarchy between a container subject and its contained sub-
ject. The classification domain elements are used to identify the subject matter
of content that is referenced in a map. The subjects must be elements defined
in a subject scheme map.

3 Managing Semantic Annotations

In general, metadata can be associated with a resource by either locating the
embedded metadata within the markup of the resource or by coupling the asso-
280 T. Zschocke and S. Closs

ciated metadata in a separate file that is closely coupled with the resource [6].
While both types of associating metadata can be harvested, the advantage of as-
sociated metadata is that the metadata files can be managed without modifying
the content of resource files. Semantic annotation implies that descriptive data
is attached to some other data in the sense of associated metadata. It entails
the transformation of documents into computer-understandable descriptions of
resources by augmenting them with metadata that describe their meaning and
allow for new or extended ways to access information [13].
In a comparative study Schönberg & Freitag [25] analyzed the performance of
different graph-based RDF querying models based on technical documentation,
and contrasted the results with standard path-based technologies like XQuery on
XML serializations. Even though the RDF framework outperformed the XQuery
approached, one of their datasets based on a collection of DITA topics showed
that RDF frameworks could not effectively take advantage of the simplified struc-
tures. However, in a subsequent study on an improved framework for verifying
technical documentation, including DITA [26], the authors were able to demon-
strate that the information extraction of content fragments and the mapping of
metadata and structural data on a RDF graph of the document model resulted
in the anticipated performance results.
In a related study [18], an ontology-driven approach was used for the seman-
tic annotation, indexing and retrieval of document units. The authors applied
a novel semantic document model, which was modeled on DITA. By utilizing
DITA’s ability to divide content into small, self-contained topics for reuse in dif-
ferent deliverables, the authors used their model to create office-like document
units with unique IDs, semantic annotations based on concepts from annotation
ontologies as RDF graphs, and the ability for generating links across a pool of
documents. The results showed that their semantic document model achieved
the improved performance over simple syntactic matching.
Other corresponding studies on managing collections of technical documenta-
tion [31], [34] applied controlled vocabulary of faceted classifications for the se-
mantic specification of modular content elements rather than RDF/OWL triples.
However, while this novel approach helps in the structured decomposition of doc-
uments in constituent elements, it does not appear to show the same flexibility
for semantic annotations as can be achieved with RDF/OWL.

4 Lifting XML Markup to RDF

RDF represent statements about resources in the form of triples of subject-

predicate-object, which can be rendered as graphs. The literature documents
different approaches to lift XML markup to RDF. For instance, Patel-Schneider
& Siméon [20] address the issue that these two distinct models, along with spe-
cific processing steps for XML and for RDF, amount to additional efforts in
processing information both at the data level and at the semantic level, includ-
ing duplication of work due to the similarity of processing in both cases. The
authors developed a unified XQuery and XPath data model for both XML and
 Preliminary Exploration of Using RDF for Annotating DITA Topics 281

RDF that integrates XML processing before moving to semantic processing of

RDF in order to avoid these duplications.
Sperberg-McQueen & Miller [30] address the same problem of mapping XML
into RDF. Their approach associates the mappings with the vocabulary using the
xsd:annotation mechanism of XML Schema. RDF allows to make more explicit
some of the assumptions embedded in XML vocabularies, and, more importantly,
to make the data available for greater reuse and improved understanding by
larger user communities.
In a related approach [14], the proposed procedure applies an ontology to
specify relevant concepts, their relationships, and properties. The conversion
process begins by changing the document into an active resource, that is, the
URI of the document will serve as the subject of the first statement. Then, all the
elements of the XML document are traversed before mapping each element with
the corresponding property, class definition or text element in the describing
ontology. The goal of this process is to transform ambiguous XML data into
more useful RDF statements.
A similar approach of XML to RDF conversion [33] integrates the XML
Schema with the accompanying XML data, an OWL ontology, and the map-
ping document. The latter describes the link between an XML data and an
OWL ontology. The result of this procedure are RDF instances based on the
XML data, which are compliant with the applied OWL ontology.
In another example [21] bibliographic metadata was combined with non-
bibliographic data from different sources by lifting and mapping the former to
RDF triples based on a Dublin Core XML representation. This process begins
with a lossy mapping from a few pre-identified ontologies. Then, the triples are
supplemented by incorporating relevant data from external sources in order to
reveal the relationships among data elements. After that, those relationships
are considered that could benefit from RDF graph representations. Finally, the
collection is made available for further exploration by end-users. The goal was
to integrate digital library content with data from non-library sources, and to
provide semantically enhanced services to end-users to explore such collections
more effectively.
In the example of a system for semantic annotations and search in virtual
collections of cultural-heritage objects [27] the procedure begins by translating
thesauri into RDF/OWL in order to make the controlled vocabularies interoper-
able. In a second step, the metadata schema is aligned by mapping it to Dublin
Core. Then, the collection metadata is enriched through a basic transforma-
tion to RDF/OWL triples. Finally, the controlled vocabularies are aligned using
owl:sameAs and skos:exactMatch relations. The result of this harvesting process
leads to a graph representing a connected network of objects and thesauri entries
that serve as the background knowledge of the collection.
In a corresponding example of a document management architecture [17] an-
notations serve as external remarks that can be attached to a document or a
subsection of a document. As annotations reside outside, the document can be
annotated as a whole or in part without actually having to modify its content or
282 T. Zschocke and S. Closs

structure. At the same time the annotations are used to enhance query requests
to access all the annotations for a particular document as well as to access the
metadata property of specific annotations.

5 Case Example: Annotating DITA Topics

The following example has been produced to illustrate the application of se-
mantic annotation of DITA topics derived from material on agroforestry. The
following code snippet is a DITA concept on tree nurseries, one of the key prac-
tices in the work of the World Agroforestry Centre:

<?xml version="1.0" encoding="UTF-8"?>

<!DOCTYPE concept PUBLIC "-//OASIS//DTD DITA Concept//EN"
"[Link]">
<concept id="concept_nursery_res">
<title>Tree nurseries</title>
<shortdesc>Nurseries managed by research units.</shortdesc>
<prolog>
<author>James M. Roshetko</author>
<publisher>World Agroforestry Centre, Winrock International
</publisher>
<copyright>
<copyryear year="2010"/>
<copyrholder>World Agroforestry Centre</copyrholder>
</copyright>
<metadata>
<audience type="user" job="using" experiencelevel="novice"/>
<keywords>
<keyword keyref="plantnurseries">
plant nurseries
</keyword>
</keywords>
</metadata>
</prolog>
<conbody xml:lang="en">
<p>A tree nursery is a managed site, designed to produce
tree seedlings grown under favorable conditions until they
are ready for planting.</p>
</conbody>
</concept>

The example produces the definition of the concept "tree nursery" along with
some metadata descriptions contained in the <prolog> and the corresponding
<metadata> element. Noticing that many metadata elements in DITA map to
Dublin Core, the topic can be expressed in RDF as follows:
 Preliminary Exploration of Using RDF for Annotating DITA Topics 283

<?xml version="1.0" encoding="utf-8"?>

<rdf:RDF xmlns:rdf="[Link]
xmlns:dc="[Link]
<rdf:Description rdf:about="[Link]
concept_tree_nursery">
<dc:title xml:lang="en">Tree nurseries</dc:title>
</rdf:Description>
<ex:editor>
<rdf:Description ex:fullName="James M. Roshetko">
</rdf:Description>
</ex:editor>
</rdf:Description>
</rdf:RDF>
The concept "tree nurseries" can now be expressed in RDF as follows (using
the Turtle syntax for RDF):
ex:tree nurseries rdf:type skos:Concept;
skos:prefLabel "tree nurseries"@en.
Following the procedure of [14] the following interpretation of the meaning of
some of this data can be produced based on the tag names of the DITA markup
(using pseudo syntax):
Class prolog
Class copyryear
Property copyright
Property year
- rdfs:domain copyryear
Assuming that the URI to identify the document is
[Link] the following sam-
ple RDF triples can be produced:

subject predicate object

[Link] rdfx:describes anon_1
concept_tree_nursery
anon_1 rdf:type onto:prolog
anon_1 onto:copyright anon_2
anon_2 rdf:type onto:copyryear
anon_2 onto:year "2010"

The elements inside of <metadata> in the example above provide information

about the content and subject of the topic. One of the elements is a keyword
related to the concept retrieved from AGROVOC. However, the relationship is
not indicated. Instead, a <subjectScheme> is referenced in which the keyword
is listed within a collection of controlled values:
284 T. Zschocke and S. Closs

<?xml version="1.0" encoding="UTF-8"?>

<!DOCTYPE subjectScheme PUBLIC "-//OASIS//DTD DITA
Subject Scheme Map//EN" "[Link]">
<subjectScheme>
<subjectdef href="[Link]
data/c_5973" format="rdf+xml"
keys="plantnurseries" navtitle="Plant nurseries">
<hasNarrower keys="ornamentaltreenurseries" navtitle=
"Ornamental tree nurseries"/>
<hasNarrower keys="forestnurseries" navtitle="Forest
nurseries"/>
<hasNarrower keys="fruittreenurseries" navtitle="Fruit
tree nurseries"/>
<hasRelated keys="seedbeds" navtitle="Seedbads"/>
<hasRelated keys="plantingstock" navtitle="Planting
stock"/>
<hasRelated keys="fruittreenurseries" navtitle="Fruit
tree nurseries"/>
</subjectdef>
</subjectScheme>
The controlled vocabulary has been retrieved from AGROVOC. Notic-
ing the similarity in expressing the semantic relationship in DITA with the
SKOS, the DITA <subjectScheme> map can easily be transformed to the
skos:ConceptScheme class in SKOS/RDF:
ex:plant nurseries rdf:type skos:Concept;
skos:inScheme ex:agroforestryThesaurus.
ex:ornamental forest nurseries rdf:type skos:Concept;
skos:broader ex:plant nurseries;
skos:inScheme ex:animalThesaurus.
ex:Forest nurseries rdf:type skos:Concept;
skos:broader ex:plant nurseries;
skos:inScheme ex:agroforestryThesaurus.
These few examples provide a brief glimpse into the issue of expressing topics
and relationships authored in DITA by lifting DITA XML to RDF. The use
of controlled vocabularies on agroforestry in the semantic annotation in the
above examples retrieved from the AGROVOC thesaurus provides an additional
mechanism to enhance the discovery of these topics in both the original DITA
markup and in RDF. However, while these terms are embedded in the markup,
it would be more beneficial to utilize AGROVOC published as Linked Data
[1], [16]. The relevant terms can be linked by pulling content to the referencing
element using DITA’s addressing attributes @conref or @href, as indicated above,
especially when handling non-DITA XML resources. At the same time, DITA
could be exposed as Linked Data by pushing content using its @id attribute (for
topics or elements) or its URI-based (direct) addressing mechanism.
 Preliminary Exploration of Using RDF for Annotating DITA Topics 285

6 Conclusion

Studies have shown that DITA is cost-effective in integrating the management

of technical documentation and training materials for increased reusability [15].
Semantic annotations provide the flexibility to make more explicit the topic-
oriented structure of DITA and corresponding aggregates in larger stores of
DITA topics for improved searches by adding more valuable background knowl-
edge. Novel graph-based RDF querying models for managing large data stores
seem equally effective for DITA topics (see [26]). However, more testing is needed
with respect to the transformation of DITA XML files into RDF by exploring
and adapting existing generic XSLT transformations (e.g., [2]). We also intend
to continue developing a thesaurus on agroforestry to be integrated as con-
trolled vocabulary into DITA topics (see [37]), making it available as linked data
while benefiting from mappings of existing thesauri with AGROVOC [16]. Using
DITA’s topic-based structure, we intend to experiment with services based on
these components and semantic annotations to couple DITA topics with poten-
tial user actions as Linked Data. For instance, this could enhance DITA-based
online help [3] towards micro-applications that are joined together using auto-
mated assistance for more unified and rich user interaction (for a related example,
see [5]).

References
1. Baker, T., Keizer, J.: Linked Data for fighting global hunger: Experiences in setting
standards for agricultural information management. In: Wood, D. (ed.) Linking
Enterprise Data, pp. 177–201. Springer, Heidelberg (2010)
2. Breitling, F.: A standard transformation from XML to RDF via XSLT. Astron
Nachr. 330, 755–760 (2009)
3. Closs, S.: Single Source Publishing. Modularer Content für EPUB & Co (Single
source publishing. Modular content for EPUB & co). [Link], Frankfurt/M
(2011)
4. Decker, S., Meinik, S., Van Hermelen, F., Fensel, D., Klein, M., Broekstra, J.,
Erdmann, M., Horrocks, I.: Semantic Web: The Roles of XML and RDF. IEEE
Internet Comput. Mag. 4, 63–74 (2000)
5. Dix, A., Lepouras, G., Katifori, A., Vassilakisc, C., Catarci, T., Poggie, A., Ioan-
nidis, Y., Mora, M., Daradimos, I., Md. Akima, N., Humayoun, S.K., Terella, F.:
From the web of data to a world of action. Web Semant Sci. Serv. Agents World
Wide Web 8, 394–408 (2010)
6. Duval, E., Hodgins, W., Sutton, S., Weibel, S.L.: Metadata Principles and Prac-
ticalities. D-Lib Mag. 8, [Link]
(retrieved September 16, 2012)
7. Eberlein, K.J., Anderson, R.D., Joseph, G. (eds.): Darwin Information Typ-
ing Architecture (DITA) Version 1.2. Organization for the Advancement
of Structured Information Standards (OASIS), Burlington, MA (December
2010), [Link] (re-
trieved September 16, 2012)
286 T. Zschocke and S. Closs

8. Ferdinand, M., Zirpins, C., Trastour, D.: Lifting XML Schema to OWL. In: Koch,
N., Fraternali, P., Wirsing, M. (eds.) ICWE 2004. LNCS, vol. 3140, pp. 354–358.
Springer, Heidelberg (2004)
9. Gibbins, N., Shadbolt, N.: Resource Description Framework (RDF). In: Bates,
M.J., Maack, M.N. (eds.) Encyclopedia of Library and Information Science, 3rd
edn., vol. 5, pp. 4539–4547. CRC Press, Boca Raton (2010)
10. Greenberg, J.: Metadata and Digital Information. In: Bates, M.J., Maack, M.N.
(eds.) Encyclopedia of Library and Information Science, 3rd edn., vol. 6, pp. 3610–
3623. CRC Press, Boca Raton (2010)
11. IEEE Std 1484.12.1TM -2002:IEEE Standard for Learning Object Metadata. Insti-
tute of Electrical and Electronics Engineers (IEEE), New York, NY (June 2002)
12. ISO 15836:2009: Information and Documentation—The Dublin Core Metadata El-
ement Set (2nd ed.). International Organization for Standardization (ISO), Geneva
(February 15, 2009)
13. Kiryakov, A., Popov, B., Terziev, I., Manov, D., Ognyanoff, D.: Semantic annota-
tion, indexing, and retrieval. Web Semant. Sci. Serv. Agents World Wide Web 2,
49–79 (2004)
14. Klein, M.: Using RDF Schema to interpret XML documents meaningfully. In:
Handschuh, S., Staab, S. (eds.) Annotation for the Semantic Web, pp. 79–89. IOS
Press, Amsterdam (2003)
15. Levine, D.B.: Cost-benefit analysis of a bridge to integrate the management of
technical information for producing technical manuals and training courses. Insti-
tute for Defense Analyses, Alexandria (2010),
[Link]
getRecord&metadataPrefix=html&identifier=ADA542429
(retrieved September 16, 2012 )
16. Morshed, A., Caracciolo, C., Johannsen, G., Keizer, J.: Thesaurus alignment for
Linked Data publishing. In: Proceedings of the International Conference on Dublin
Core and Metadata Applications (DC 2011), pp. 37–46. Dublin Core Metadata
Initiative (DCMI), Dublin (2011),
[Link]
(retrieved September 16, 2012)
17. Meena, E., Kumar, A., Romary, L.: An extensible framework for efficient document
management using RDF and OWL. In: Ide, N., Romary, L. (eds.) Workshop on
NLP and XML (NLPXML-2004): RDF/RDFS and OWL in Language Technology
(NLPXML 2004), pp. 51–58. Association for Computational Linguistic, Strouds-
burg (2004), [Link]
(retrieved September 16, 2012 )
18. Nešić, S., Jazayeri, M., Crestani, F., Gaševic: Concept-based semantic annota-
tion, indexing and retrieval of office-like document units. Technical report, USI-
INF-TR-2010-1, Faculty of Informatics, Università della Svizzeria italiana (2010),
[Link] (retrieved Septem-
ber 16, 2012 )
19. NISO: Understanding Metadata. National Information Standards Organization
(NISO) Press, Bethesda, MD (2004), [Link] (retrieved Septem-
ber 16, 2012)
20. Patel-Schneider, P.F., Simeon, A.: The Yin/Yang Web: a unified model for XML
syntax and RDF semantics. IEEE Trans. Knowl. Data Eng. 15, 797–812 (2003)
 Preliminary Exploration of Using RDF for Annotating DITA Topics 287

21. Powell, J.E., Collins, L.M., Martinez, M.L.B.: Semantically enhancing col-
lections of library and non-library content. D-Lib Magazine 16 (2010),
[Link] (retrieved Septem-
ber 16, 2012)
22. Priestley, M., Hargis, G., Carpenter, S.: DITA: An XML-based Technical Documen-
tation Authoring and Publishing Architecture. Tech. Comm. 48, 352–367 (2001)
23. Salminen, A., Tompa, F.: Communicating with XML. Springer, Heidelberg (2011)
24. Sánchez-Alonso, S., Sicilia, M.-Á.: Using an AGROVOC-based ontology for the
description of learning resources on organic agriculture. In: Sicilia, M.-Á., Lytras,
M.D. (eds.) Metadata and Semantics, pp. 481–492. Springer, Heidelberg (2009)
25. Schönberg, C., Freitag, B.: Evaluating RDF querying frameworks for document
metadata. Technical report, MIP-0903, Fakultät für Informatik und Mathematik,
Universität Passau (2003),
[Link]
forschungsberichte/[Link]
(retrieved September 16, 2012 )
26. Schönberg, C., Weitl, F., Freitag, B.: Verifying the consistency of web-based tech-
nical documentation. J. Symbolic Comput. 46, 183–206 (2011)
27. Schreiber, G., Amin, A., Aroyo, L., van Assem, M., de Boer, V., Hardman, L.,
Hildebrand, M., Omelayenko, B., van Osenbruggen, J., Tordai, A., Wielemaker,
J., Wielinga, B.: Semantic annotation and search of cultural-heritage collections:
The MultimediaN E-Culture demonstrator. Web Semant. Sci. Serv. Agents World
Wide Web 6, 243–249 (2008)
28. Shadboldt, N., Berners-Lee, T., Hall, W.: The Semantic Web revisited. IEEE Intell.
Syst. 21, 96–101 (2006)
29. Shreve, G.M., Zeng, M.L.: Integrating resource metadata and domain markup in
an NSDL collection. In: Proceedings of the International Conference on Dublin
Core and Metadata Applications (DC 2003), pp. 223–229. Dublin Core Metadata
Initiative (DCMI), Dublin (2003),
[Link]
(retrieved September 16, 2012)
30. Sperberg-McQueen, C.M., Miller E.: On mapping from colloquial XML to RDF
using XSLT. In: Extreme Markup Languages 2004 (2004),
[Link]
2004/Sperberg-McQueen01/[Link]
(retrieved September 16, 2012)
31. Streich, R.: Techniques for managing collections of interrelated text modules.
Markup Languages: Theory and Practice 1, 77–94 (1999)
32. Uren, V., Cimiano, P., Iria, J., Handschuh, S., Vargas-Vera, M., Motta, E.,
Ciravegna, F.: Semantic annotation for knowledge management: Requirements and
a survey of the state of the art. Web Semant. Sci. Serv. Agents World Wide Web 4,
14–28 (2006)
33. Van Deursen, D., Poppe, C., Martens, G., Mannens, E., Walle, R.: XML to RDF
conversion: A generic approach. In: Nesi, P., Ng, K., Delgado, J. (eds.) Interna-
tional Conference on Automated Solutions for Cross Media Content and Multi-
Channel Distribution (AXMEDIS 2008), pp. 138–144. University of Florence, Flo-
rence (2008)
34. Wild, P.J., Giess, M.D., McMahon, C.A.: Describing engineering documents with
faceted approaches. J. Doc. 66, 420–445 (2009)
288 T. Zschocke and S. Closs

35. Zeng, M.L.: Domain-specific markup languages and descriptive metadata: Their
functions in scientific resource discovery. Revista Eletrônica de Biblioteconomia e
Ciência da Informação 15, 164–176 (2010),
[Link]
(retrieved September 16, 2012)
36. Zeng, M.L., Chan, L.M.: Semantic Interoperability. In: Bates, M.J., Maack, N.
(eds.) Encyclopedia of Library and Information Sciences, 3rd edn., vol. 6, pp. 4645–
4662. CRC Press, Boca Raton (2010)
37. Zschocke, T.: Resolving controlled vocabulary in DITA markup: A case example
in agroforestry. Program. 46, 321–340 (2012)