Desi De signi gning ng the he Metadata Mod odel for or the he - PowerPoint PPT Presentation

ICADL 2019 Workshop: MAGIC- Information Commons for Manga, Anime and video Games Desi De signi gning ng the he Metadata Mod odel for or the he Aggr Aggrega gation ion of of MAG AG Metadata - the Development of Media-Art Database Dr. Tetsuya Mihara Faculty of Library, Information and Media Science, University of Tsukuba

Se Self Introduction: Tetsuya Mi Mihara • Research Interest: Metadata for Manga • Manga becomes representatives of Japanese popular culture • Over 12,000 titles, 16% of all books in Japan are published per year (2017, except e- publishing) • Keywords: Linked Open Data, semantic Web, digital archives, digital humanities • Joined the Media-Art Database project in 2013 • as a specialist about MAG metadata and information system

Ov Overview of of Manga ga Meta tada data ta Research Proj oject

Medi Media-Ar Art Da Data taba base (メディア芸術データベース, MADB) https://mediaarts-db.bunka.go.jp/ • The unique national database for Media-Arts produced in Japan • Media Art ( メディア芸術、メディアアート ) consists of 4 category of media • Manga • Animation • Video game • New media art • The law “Basic Act for the Promotion of Culture and the Arts” is enforced in 2001

The The Mission on of of MADB Proj oject • Collect data to cover information of Media Art produced in Japan Bibliographies about over 330 thousand Manga books and magazines, 9 • thousand Anime series , 35 thousand video game titles are recorded • Connect libraries and museums • There are not any institution to collect and cover all Media-Art in Japan • Database development started to gather cataloging records of libraries and museums which have Media-Art collections • These are insufficient to cover whole products • Collaborator: National Diet Library (MAG), 5 special libraries and museums (manga), Ritsumeikan Center of Game Studies (video game) etc…

Cha Challenge nges/ s/Di Difficul ulties s of of MADB ADB Pr Proj oject • Creation of data to trans • Difference of data creator each media • Media • Work, series • Institution • Author, creator • Mix of different grounds of data • Difference of the format of • Objects in the collection the description • From records or catalog • Title + readings • Object not in the collection • Date

Cr Creat eate L e Lin inked ed Dat Data o a of “M “Med edia ia-Ar Art” t”

Re Relationships of MAG AG • Media franchise • メディアミックス(Media Mix) • Manga/Anime/Video Game • ACG (Anime, Comic, Game) • Example: Pokémon, Gundam, etc… • Other cultural resources • Japanese sword (Nihon-tou): Token Ranbu(⼑剣乱舞) • Historic heritage: Assassinʼs Creed series

Work Wo • Work is core idea to link publications of different media Pokémon • User seems to use work User recognize information (especially title) for Work searching items ≠ • Difficulties • Things called “Work” would not be the same in different media • Identification Game Manga Anime User can handle Item

Wh What Do Do W We Do Do f for t the M MADB DB De Deve velo lopment 1. Design (meta)data model • Need integrated data model to describe structural relationships of each media • Abstract data model + data model for each media 2. Develop the data management system • Customize an existing proprietary system by IT vender 3. Develop how to create data instance/cataloging rule • Convert existing cataloging records from collaborators • Create new records • Transcribe from old printed catalogs • Generate abstract instance from specific description of records

St Structure of f the MADB B Abstract Metadata Model Authority Work Entity Core Part Agent Work Event Variation Across Categories Component Each Categories Related Instantiation Resource Game … Console Item

De Defini nition on of of Wor ork Ent ntities • Work Entity： consists of “Media-Arts” • Work ： is identified as a Media-Arts work • Variation ： represents different types of expression or appearance of a Work • Component ：is a part of Variation • Instantiation ： is an instance of work • Related Resource ：does not consists of Work, but related as material • Item ：is actual item or record of Instantiation and Related Resource which represents its exitance

Data Model: Transmedia Work (reading) Work (title) hasPart subtype Work Group (reading) hasPart (title) Work Work Work Work subtype subtype author subtype subtype New Media Art Videogame Agent Manga Anime Magazine Version Demonst- Series Title (Videogame) ration hasPart (Anime) Series Series Version Exibit (Manga) (Anime) (Videogame)

Data Model: Manga Magazine Print in Work Work Magazine Tankoubon (reading) Tankoubon (reading) subtype (title) Magazine Series Series Title (Manga) (series title) subtype (Manga) JA/EN/… JA/EN/… Print in Magazine Subtype (reading) monthly (title) External Magazine Code Episode ID xxxxx-x Epi.XX Episode YYYY (reading) (reading) Vol.XX (title) Volume number Volume Number Vol. X (Title) Tankobon publishing date Book publishing date Brand Magazine Appearance YYYY-MM-DD YYYY-MM-DD Volume Author Author hasPart ISBN: XXX- publisher YYYY-MM-DD X-XX- Agent External Author Agent Agent XXXXXX-X ID Agent Item Agent Item Agent

Data Model: Anime TV Broadcast 作品 Movie Video よみ サブタイプ テレビ放映 版（アニメ） 版名 版（アニメ） 版（アニメ） サブタイプ サブタイプ ⾔語 責任主体 作者 JA/EN/… ビデオシリーズ 劇場版 よみ 役名 キャスト 各話名 責任主体 各話 各話 各話 24分 よみ 役職名 よみ XX話 スタッフ タイトル名 よみ 責任主体 タイトル名 VHS タイトル名 ビデオ X巻 テレビ〇〇系 テレビ番組 パッケージ 劇場上映 YYYY-MM-DD 巻 公開年⽉⽇ YYYY-MM-DD YYYY-MM-DD 公開年⽉⽇ hh:mm 公開年⽉⽇ 発⾏者 責任主体 〇〇系配給 YYYY-MM-DD 開始年⽉⽇ hh:mm 放送履歴 放送局 所蔵情報 上映履歴 シアター〇〇 責任主体 YYYY-MM-DD 終了年⽉⽇ hh:mm

Data Model: Videogame 家庭⽤ゲーム 家庭⽤ゲーム 作品 （パターン2） （パターン１） よみ よみ タイトル タイトル バージョン バージョン スタッフ 役職名 スタッフ （ゲーム） 役職名 （ゲーム） サブタイプ 責任主体 責任主体 ⾔語 表現規制版 ⾔語 JA/EN/… ゲームコンソール JA/EN/… プラットフォーム よみ よみ タイトル タイトル サブ サブ スタッフ バージョン 役職名 スタッフ バージョン 役職名 サブタイプ 責任主体 ゲームコンソール 責任主体 ⾔語 プラットフォーム ⾔語 復刻版 JA/EN/… JA/EN/… よみ よみ パッケージ パッケージ タイトル サブタイプ タイトル サブタイプ 公開年⽉⽇ YYYY-MM-DD Blu-rayディスク 公開年⽉⽇ YYYY-MM-DD ROMカートリッジ 所蔵情報 責任主体 所蔵情報 責任主体

We Will Keep Struggling… • Identifier : need to obvious • Publish and open discussion rules for identifying entities • Metadata model • Label of entities: always difficult • Cataloging rules • Controlled vocabulary : who • Actual Data and how create? • Web APIs • Proves the meanings of the data • Wiki data which is crucial for the long-use • Data correction