What is ITS 2.0?

• Internationalization Tag Set (ITS) 2.0
• W3C Recomendation defining metadata ("data categories") for multilingual content production
• Data categories in ITS 2.0: Translate, Localization Note, Terminology, Directionality, Language Information, Elements Within Text, Domain, Text Analysis, Locale Filter, Provenance, External Resource, Target Pointer, ID Value, Preserver Space, Localization Quality Issue, Localization Quality Rating, MT Confidence, Allowed Characters, Storage Size

ITS 2.0 data categories

• Defined in prose = in an abstract way. Examples:
• The Translate data category expresses information about whether the content of an element or attribute is intended for translation or not. The values of this data category are "yes" (translatable) or "no" (not translatable).
• The Text Analysis data category is used to annotate content with lexical or conceptual information for the purpose of contextual disambiguation. [... ] Example: A named entity recognizer provides the information that the string "Dublin" [...] denotes a town in Ireland.

ITS 2.0 data categories - annotation for certain types of content: HTML, XML

<p>The event is organized by the 
<span translate="no" 
its-ta-ident-ref=
"http://dbpedia.org/resource/World_Wide_Web_Consortium"

its-ta-class-ref="http://schema.org/Organization">W3C</span></p>

Growing interest in annotating further types of content with i18n information

• Localizing RDF ontologies; providing information about e.g. what parts should be localized. Serialization of annotations: RDF?
  • Continuation of work that started informally during ITS 2.0 draft time: ITS 2.0 ontology
• Annotating JSON with information what parts should be localized. Serialization of annotations: JSON?
• All of the above is being discussed in various fora & non standard yet

ITS 2.0 -
and the Web Annotation Model?

• Is there an interest by the Web Annotation community in ITS 2.0 data categories? A general question; the concrete approach to represent ITS 2.0 in relation to Web annotations is to be discussed :-)
• How is the relation of Web Annotations to an existing approach of annotating Web content with ITS 2.0 - NIF?
  • Natural Language Processing Interchange Format
  • NIF developed in EU project context (LIDER, FREME) and deployed in context of German project Digital curation technologies

NIF

• Linked data format for creating Natural Language Processing (NLP) workflows, involving tools, language resources and annotations.
• Example input sentence: "This event is organised by the W3C"
• See example NIF output created via EU project FREME API
  (the /e-entity/freme-ner/documents API endpoint)
• Annotation types provided by NIF: oriented on linguistic analysis
  • isString, Word, Phrase, ...
• Annotations are anchored in textual content via character offsets
  • beginIndex, endIndex

NIF output example (1/2)

 "@id" : "http://freme-project.eu/#char=0,31",
"@type" : [ "nif:String", "nif:Context", "nif:RFC5147String" ],

"beginIndex" : "0",

"endIndex" : "31",

"nif:isString" : "This event is organized by W3C."

NIF output example (2/2)

"@id" : "http://freme-project.eu/#char=27,30",

"nif:anchorOf" : "W3C",

"beginIndex" : "27",

"endIndex" : "30",

"referenceContext" : "http://freme-project.eu/#char=0,31",

"taClassRef" : "http://nerd.eurecom.fr/ontology#Organization",

"itsrdf:taConfidence" : 0.9884513430050544,

"taIdentRef" :
"http://dbpedia.org/resource/World_Wide_Web_Consortium"