This paper focuses on the conversion between the open standard XML Localisation Interchange File Format (XLIFF) and the Resource Description Framework (RDF). XLIFF is a localisation standard supported by proprietary and free and open source software (FOSS) localisation tools, while the latter is a standard model, basic ingredient in Semantic Web. We developed a converter based on Saxon XSLT Processor which translates XLIFF to RDF.
Generally speaking, standards incorporate a solid body of knowledge and provide a unified framework. In addition, when metadata is standardised, resources can be identified, catalogued, and processed faster and more efficiently. Although standards as such are a benefit for information management, in the last years we have seen too many standards evolving in information science. In our opinion, the existence of too many standards in tandem with their inflexible structure (of some standards) adds complexity and leads to lack of interoperability; interoperability between Web resources is crucial for communication between application components.
This paper focuses on XLIFF1 and RDF2 and the conversion based on Saxon from the former to the latter. Our work is motivated by the insight that Web resources should be multilingual and XLIFF as a localisation standard is capable to help localise ontologies and thus create multilingual linked data. A wider target range of users and applications will then be reached. The automatic conversion from XLIFF into RDF can be used as an API both by localisation tools and Semantic Web applications.
In section 2 we describe some related work about combining multilinguality with
Semantic Web. In sections 3 and 4 some examples of XLIFF and RDF are provided.
Section 5 discusses the XLIFF-RDF interoperability and then we conclude the paper.
In 2004 [
The Multilingual Semantic Web workshop started in 2010 and continues with annual workshops; the same holds for the XLIFF International Symposium. Some research projects: the Multilingual Web3, Flarenet4, META-NET5, and Monnet6 see the symbiotic relationship between multilingual resources and Semantic Web.
As far as the conversion between XLIFF and other standards is concerned, the
Okapi Framework provides XLIFF conversion utilities, e.g. to Translation Memory
eXchange (TMX). [
Our main motivation for XLIFF2RDF conversion is the concept of ‘ontology
localization’, a term coined by [
XLIFF is an open localisation standard supported by proprietary and FOSS localisation tools. It is under the auspices of OASIS and is understood by many 3 http://www.multilingualweb.eu/en, 12/09/11 4 http://www.flarenet.eu/, 12/09/11 5 http://www.meta-net.eu/, 12/09/11 6 http://www.monnet-project.eu/Monnet/Monnet/English?init=true, 12/09/11 7 http://lexinfo.net/, 12/09/11 actors: software providers, localisation service providers, and localisation tools providers. Semantic localisation metadata is very important in a localisation workflow to distinguish between the responsibilities of each stakeholder (project manager, engineer, translator, proofreader), between translatable and non-translatable content, annotate (in the case of translatable content) the status of the strings and so on. Particularly in software localisation, coordinates of menus dialogue boxes, version control, count of screenshots belong to the most important metadata. The following example contains an XLIFF file with three translation units (TUs). TU elements include a <source>, <target> and associated elements.
1. <?xml version="1.0" encoding="UTF-8" ?> 2. <xliff version="1.2" xmlns="urn:oasis:names:tc:xliff:document:1.2"> 3. <file original="minimal_XLIFF.html" source-language="en-us" target-language="de-de" datatype="html"> 4. <body> 5. <trans-unit id="#1"> 6. <source>book</source> 7. <target>Buch</target> 8. </trans-unit> 9. <trans-unit id="#2"> 10. <source>book publisher</source> 11. <target>Buchverlag</target> 12. </trans-unit> 13. <trans-unit id="#3"> 14. <source>This book is good!</source> 15. <target>Dieses Buch ist gut!</target> 16. </trans-unit> 17. </body> 18. </file> 19. </xliff>
Example 1. XLIFF file with three translation units. Line 1: XML declaration, Line 2: XML schema, Line 3: file metadata, Lines 5-16: file data (three TUs). 4 RDF RDF is family of W3C specifications which describe Web resources. Here is a brief explanation of Resource, Property, and Property value by means of the XLIFF Ex.1: • A Resource is anything that can have a URI, e.g. minimal_XLIFF.html; • A Property is a Resource that has a name, such as trans-unit, source; • A Property value is the value of a Property, such as This book is good! The example 1 can be represented in an RDF graph as follows:
Accordingly, every XLIFF file can be represented in an RDF graph. The circles are the resources, the labels on the arrows are the properties, and the content of the rectangles are the property values. idX is a placeholder for a resource representing the body.
Building a bridge for interoperability between RDF and other standards is something common: WSDL-RDF, RDF-Topic Maps, OWL-RDF, and others. However, these standards, which RDF can be converted from and into, also come from the Semantic Web world and not from the localisation scene.
As far as the representation of multilingual information in RDF is concerned, RDF
used the RFC 3066 standard (published in 2001) for language tags for literals in
natural languages. The revision RFC3066bis included productive use of language,
country and script codes. [
The greatest contribution of XLIFF is the nature of its content, i.e. the capture of translation pairs, rather than the formalisation vehicle of the knowledge, be it XML or RDF. We do not intend to reify XLIFF, but to make XLIFF portable to RDF. The reasons why an XLIFF2RDF mapping and conversion are useful follow: i. Any file format which can be converted into XLIFF can be then converted to RDF; ii. RDF ontology labels can be translated using XLIFF; iii. Web resources can be described by XLIFF metadata.
A practical implementation of standards’ interoperability between XLIFF and
RDF(S) is distinguished between two parts: mapping XLIFF elements and attributes
to RDF and automatically converting from XLIFF into RDF. The mapping of three
XLIFF files has been described in [
Example 2. Sample of the XSLT
It should be mentioned that there is discrepancy between interoperability between data based on standards and interoperability between standards. Conversion between standards plays a small part within the wider scope of interoperability which includes, among others, supporting relevant standards and conforming with specifications.
The development of a conversion tool to translate from XLIFF into RDF automates and thus accelerates the process. We used NetBeans IDE to create a GUI of the conversion tool (see Screenshot 1). For our conversion utilities we used the Saxon home edition 9.3 version8. The home edition is an open source product available under the Mozilla Public License. It provides implementations of XSLT 2.0, XQuery 1.0, and XPath 2.0 and is available for both Java and .NET. The user can input one or more XLIFF file(s) to the tool, convert them to RDF and preview them.
In this paper we discussed the interoperability between the localisation standard XLIFF and RDF. We showed ongoing initiatives, projects, and tools combining multilinguality with Semantic Web. We developed a converter from XLIFF to RDF by using and adapting the Java API of the XSLT processor Saxon. We wrote some sample XLIFF files and adopted a modular transitional file provided in the XLIFF latest specifications in order to create corresponding XSLTs.
In our opinion, localisation is often regarded only as a business strategy to increase return on investment and not as a research field which can both enrich and gain from the Semantic Web and Linked Data. Localisation standards and particularly XLIFF has received little attention although it covers many actors’ needs.
In Semantic Web context, it is an arbitrary decision in which natural language the ontology labels are provided, and thus many researchers see the need for multilingual ontologies; challenges, like cross-lingual mapping and translation follow the existence 8 http://saxon.sourceforge.net/, 12/09/11 9 http://code.google.com/p/xliff-rdf/, 28/03/11 of multilingual ontologies. Our conversion tool is a contribution to build a bridge between localisation and Semantic Web resources, so that localisation tools can localise ontologies and Semantic Web resources are populated with localisationrelated metadata. After the XLIFF2RDF conversion, metadata can be reused in the Semantic Web to represent multilingual ontologies. The XLIFF2RDF conversion tool is hosted on Google code hosting website. There other users can freely get a local copy of the tool; thus replication of the tool is allowed. The conversion tool fulfills its basic requirements, i.e. XLIFF files are represented in RDF. Not only minimal XLIFF examples with one TU, but with more TUs and also with file processing metadata, alternative translations, etc. can be successfully converted. Five use cases have been successfully tested, however more quantitative and qualitative examples are planned to be converted. We plan to extend the conversion API for other standards. At first place, we plan to translate from XLIFF into OWL. Also interoperability between other localisation and internationalisation standards is also among future prospects. In terms of quality assurance, existing validation tools will be part of our tool. Acknowledgment. We gratefully acknowledge the support of the Deutsche Forschungsgemeinschaft (DFG) through the Collaborative Research Center SFB/TR 8 Spatial Cognition - Subproject I5-DiaSpace.