아시아 미술정보의 공유와 확장

_아시아에서의 문화 데이터 공유력 활성화

아시아 아트 아카이브(이하 총칭 AAA:Asia Art Archive)는 지난 3년간 20,000여 점의 디지털 자료를 모아 온라인에서 볼 수 있게 만들었다. 이 디지털 콘텐츠는 AAA에서 자체적으로 개발한 디지털 자산관리 시스템(digital assets managment system)에 의해 운영되는 AAA 웹사이트에서 ‘Collection Online(온라인 컬렉션)’을 통해 제공되고 있다. 이 시스템을 2009년 처음 개발했을 당시  우리의 근본적인  목표는 귀중한 자료들을 사용자들이 쉽게 접근할 수 있도록 온라인상에서 디지털 콘텐츠를 보여주는 것이 었다. 우리의 웹사이트가 2012년 새롭게 론칭된 지  얼마 지나지 않아 방문객 수는  눈에 띌 만큼 배로 증가하였고, 대부분이 우리의 디지털 콘텐츠를 이용하기 위해 접속해 들어왔다. 비록 그 콘텐츠들이 온라인에서 무료로 볼 수 있게 제공되고 있긴 하지만, 사실 저작권 소유자의 승인 확보 없이는 사용자들이 자료들을 다운로드하거나 재사용할 수 없는 부분이 있다. 그리고 이러한 기준이 콘텐츠의 재사용과 그 이상의 지식 생산을 가로막는다.

내가 처음 Linked Open Data(LOD, 링크드 오픈 데이터)에 대해 알게 된 것은 2011년 말이었으며, 우리가 디지털 방식을 내부적으로  전체 도입했을 때, 나는 2년 내에 AAA 또한 이 프로젝트에 참여할 수 있을 것이란 가능성만을 보았다. 우리는 게티 Getty 테이트 Tate, 영국 국립도서관 British Library 와 같은 큰 규모의 문화기관에서의 디지털 방식 개발에 대한 소식들을 조사해 왔다. 그리고 LOD를 사용해 데이터를 더욱 공유하기 쉽도록 만들 수 있다는 걸 알게 되었다; 예들 들면, 사람과 기계(컴퓨터)가 해독할 수 있는 구조를 메타데이터에 적용시킨 후, 무료로 다운로드할 수 있는 디지털 콘텐츠와 짝지어 연결해주는 방법 같은 것이다.

Linked Open (Cultural) Data  링크드 오픈 (문화) 데이터에 대한 역사적 설명은 Lotte Belice Baltussen 로테 벨리스 발투젠 이 2013년 박물관과 웹 심포지엄에서 발제한 글 ‘오픈 컬처 데이터: 보텀업 방식의 GLAM(갤러리, 도서관, 미술관, 박물관) 데이터 공개’(Open Culture Data: Opening GLAM Data Bottom-up)’ 를 참고해 주기를 바란다. 특히 Jon Voss 존 보스 가 2012년도 박물관과 웹 심포지엄에서 발제한 ‘문화유산 데이터를 웹상에서 혁신적으로 공개하라 (Radically Open Cultural Heritage Data on the Web)’는 도서관, 아카이브, 박물관(LAM) 분야에서의 LOD 적용 방안을 개발하기 위한 연구인 LODLAM(Linked Open Data in Library, Archives and Museums)에 대해 주목할 만한 성과를 보여주는 좋은 글이다. 

LOD에 대한 움직임은 2010년 부터 유럽과 영어권(Euro-Anglo)국가의 기관들로부터 시작되었다. 이 아이디어는 시맨틱 웹 미트업 (Semantic Web meetup)의 거점이 된 곳들의 지리적 분포에서도 알 수 있듯이 유럽과 미국을 중심으로 급속하게 확산되어 왔으며, 그 영향력이 아시아로 확산되고 있다는 것이 나의 생각이다. 아시아 내에서 메타데이터 혹은 콘텐츠의 공개를 준비하고 있는 기관이 더욱 늘어남에 따라,  우리 또한  아시아만의 콘셉트(concepts), 사건(incidents), 경험(practices)들을 의식적으로 LOD의 구조에 적용시키기 위해 준비를 해야만 한다.

사람과 기계(컴퓨터)가 읽을 수 있는 구조적인 메타데이터 생성하기

최소한, 기계(컴퓨터)가 읽을 수 있는 구조적인 메타데이터는 근본적으로 ‘잘 정리된 데이터’(메튜 링컨(Matthew Lincoln)의 블로그 글에서 설명된 것처럼)를 말한다. 데이터가 잘 정리되어 있을 때  컴퓨터가 읽을 수 있는 값들로 마크업(mark-up: 문서 정보를 전달, 처리하기 위해 문서 데이터에 추가된 텍스트나 기호들을 말한다) 되고 인코딩(encoding: 각종 정보를 재현시킬 수 있는 가장 적합한 형태로 코드화, 암호화 시키는 것을 말한다) 될 가능성이 높아진다. 아직은 내가 컴퓨터가 해독할 수 있는  마크업과 인코딩의 서비스에 대한 필수 요소 혹은 어려움을 설명할 수 있을 만큼 기술적인 부분에 대한 이해력이 충분하진 않다. 하지만, 도서관 사서로서, LOD 시스템을 개발과 관련된 모든 사람들에게 아주 중요한 두 가지 사항을 강조하고 싶다: 주요 데이터(key data)를  번역하는 것과 데이터 스키마(schema: 데이터베이스에 존재하는 자료의 구조 및 내용 그리고 이러한 자료들에 대한 논리적, 물리적 특성에 대한 정보를 표현하는 데이터베이스의 논리적인 구조)에 다 개국 어의 필드(field: 데이터 레코드 중 명확히 정의된 범위, 항목, 영역, 속성)를 포함시키는 것이다.

구조화된 메타데이터를  확보하기 위한 시스템을 구축할 때 우리는 아시아인의 다민족적, 다언어적인 특징들을 포용해야 한다. 가능하다면, 제목, 출판사, 제작자/창작자와 같은 주요 데이터와 전거 표목(authority terms: 전거는 색인 작업 시 특정 개념을 표현하기 위해 일관되게 사용되는 용어로 데이터 목록의 상위에서 배열기준과 접근점이 되는 단어나 구, 기호를 말한다)은 영어로 번역되어 (유럽과 영어권(Euro-Anglo) 국가를 기반으로 한 연구에만 치우치게 된 부분에 대해 양해를 구합니다.) 자국어와 함께 별도의 항목으로 작성되어야 한다. 또한 아시아권 언어와 함께 지정된 타언어 데이터 필드가 있다면 태그(tag: 전산 문서에서 사용하는 편집용 코드, 용어) 작성이 명확하게 되어있음을 확인하여, 데이터들이 정확하게 통합될 수 있도록 해야 한다. 

사람이 읽을 수 있는 (시맨틱(Semantic): 의미) 메타데이터는 구조화의 여부와는 상관없이 HTML 웹페이지와 같이 사용자(end users)를 위한 인터페이스상에서 나타나는 정보이다. 하지만 이것이 구조화되어 있다면, 사용자가 정보를 이해하는데 도움이 될 것이며, 더욱 효과적인 검색을 실행하고, 컴퓨터를 통해 더욱 강력하게 자료를 수집할 수 있을 것이다.

시소러스(Thesaurus: 정보검색이나 정렬에 사용되는 용어나 단어의 집합)는 인간과 기계 양쪽의 매핑이 모두 가능한 구조화된 시맨틱 데이터의 유형 중 하나이다. 게티(Getty)의 예술, 건축 시소러스 (AAT: Art and Architecture Thesaurus)는 LAM 분야에서 폭넓게 쓰이고 있지만, 잘 알려진 전거 리스트에서 아시아인과 관련된 개념과 사건들은 이해력이 떨어진다. AAA는 AAA 컬렉션에 기반하여 아시아인에 관점에 맞춘  키워드를  AAT와 같이 쓸 수 있도록 편집해오고 있다. 아시아인의 관점에 맞춘 시소러스를 최대한 다양한 지역 특색 언어로 번역해 개발한다면, 이는 아시아의 콘텐츠들을 연결하는데 큰 도움이 될 것이며, LOD가 더욱더 국제적으로 알려졌을 때 메타데이터는 좀 더 쉽게 찾아질 수 있을 것이다.

디지털 콘텐츠를 공유할 수 있게 만드는 것

저작권이 있는 온라인 콘텐츠의 제한적인 이용 상황에 대해 인지하고 있었기에, AAA는 AAA가 생산한 콘텐츠에 대해 크리에이티브 커먼즈 라이선스(CC: Creative Commons licences)를 적용해 왔다. 그 결과 대부분의 AAA 프로그램 기록들은 현재 온라인에서 자유롭게 이용 가능하다. 최근 우리는 AAA 컬렉션의 자료 기증자들에게도 크리에이티브 커먼즈 라이선스를 승인해 줄 것을 제안하고 있다. 크리에이티브 커먼즈 라이선스는 콘텐츠 이용에 대해 매번 승인을 받아야만 하는 어려움 없이 융통성 있게 콘텐츠가 이용될 수 있게 해준다. 이러한 환경은 (아시아의) 크리에이티브 콘텐츠를 참고하는 사례를 증가시킬 것 이며, 더 나아가 지식생산의 확장 또한 신장시킬 것이다.

이는 미술 분야의 모든 사람들이 콘텐츠 공유에 참여할 수 있는 가능성을  열어준다. 만일 자신이 작가이거나 미술 단체의 멤버, 혹은 미술 분야의 대소사에 대한 기록을 생산하는 관리 기관에 속한 사람이라면, 크리에이티브 커먼즈 라이선스의 확산에 동참해주길 간곡히 부탁한다. 우리 AAA와 같은 LAM 기관에서 당신 개인 혹은 당신의 기관에 관한 기록을 사용할 수 있도록 승인을 요청한다면,  CC 라이선스에 대해 동의한다는 의사를 밝혀주기 바란다. 함께 노력한다면, 우리는 아시아의 미술 기록들을 세계의 미술 현장에서 더욱더 가시화 시킬 수 있다. 

마지막으로 중요한 것을 말하자면, 우리는 크리에이티브 커먼즈 제로(Creative Commons Zero(CC0): 어떤 조건의 제약 없이 콘텐츠를 사용할 수 있는 퍼블릭 도메인에 부여되는 라이선스)의 막강한 이용량에 대해 알고 있으며, AAA의 메타데이터에 이 라이선스를 등록할 예정이다. CC0 라이선스에 동의된 메타데이터는 그 어떤 수집자들(유로피아나와 같은 선도적인 문화 수집기관)도 자유롭게 사용할 수 있으며, 그 결과 우리 기관의 메타데이터가 이전엔 생각지도 못 했던 사람들과의 접속이 이루어지고, 우리 웹사이트의 콘텐츠를 이용하려는 사용자들이 이어 생겨날 것이다.

- 리디아 나이(Lydia NGAI) 오스트리아 뉴사우스웨일스대 정보경영학 석사 졸. 오스트레일리아 대학도서관 근무, 홍콩의 미술과 디자인 역사에 관한 조사원 활동, 홍콩 AAA 디지털 아카이브와 라이브러리 컬렉션 운영 담당(2008- ), 현 홍콩 AAA 도서관장

* 본 콘텐츠는 필자와의 협의를 통해 편집본을 『서울아트가이드』 지면에 게재하였음을 알려드립니다.

* AAA 온라인 컬렉션 (AAA’s Collection Online) - http://www.aaa.org.hk/Collection/CollectionOnline

* 오픈 컬처 데이터: 보텀업 방식의 GLAM 데이터 공개’ (Open Culture Data: Opening GLAM Data Bottom-up) -

  http://mw2013.museumsandtheweb.com/paper/open-culture-data-opening-glam-data-bottom-up

* 문화유산 데이터를 웹상에서 혁신적으로 공개하라 (Radically Open Cultural Heritage Data on the Web)-

  http://www.museumsandtheweb.com/mw2012/papers/radically_open_cultural_heritage_data_on_the_w

* 시맨틱 웹 미트업(Semantic Web meetup) – http://semweb.meetup.com

* ‘잘 정리된 데이터’ 에 관한 매튜 링컨의 블로그 글(Matthew Lincoln’s blog on ‘tidy data’)-

  http://matthewlincoln.net/2015/09/23/tidy-art-historical-data.html 

* 크리에이티브 커먼즈 라이선스 (Creative Commons (CC) licences) -

  http://creativecommons.org/licenses/by/4.0/legalcode

Enabling Shareabilities of Cultural Data in Asia

The Asia Art Archive(AAA) has accumulated almost 20,000 digital objects, viewable online over the past 3 years. The digital contents are delivered through ‘Collection Online’ in the AAA website which is managed by an in-house developed digital assets management system. When we initially developed the system in 2009, a principal objective was to display digital content for online viewing, so as to enable our users to access the valuable content conveniently. We had remarkably doubled the number of online users shortly after the website relaunch in 2012, mostly attracted to our digital contents. However, even though the contents are free for viewing online, they are not downloadable nor reusable by users without further obtaining permission from copyrights holders. And this practice discourages reuse and further creation of knowledge. 

I first learned about Linked Open Data(LOD) around the end of 2011, yet, I can only see the possibilities for AAA to participate in it about 2 years, when we fully enabled the digital operation internally. We had been monitoring digital progress news of big cultural organisations, such as the Getty, the Smithsonian, Tate, and the British Library etc. We recognised that making data more shareable can be archived by LOD; i.e. by applying human- and machine-readable structure to our metadata, paired with freely downloadable digital contents.

For an historical account of Linked Open (cultural) Data, please refer to the article by Lotte Belice Baltussen et al., ‘Open Culture Data: Opening GLAM Data Bottom-up’, delivered in the 2013 Museums and the Web Symposium. And, Jon Voss’s article 'Radically Open Cultural Heritage Data on the Web' in Museums and the Web Symposium 2012 is also a good reference; it provides an impressive study into LODLAM, an operation cultivates the adaptation of LOD in the Library, Archive and Museum(LAM) sector. 

LOD movement was led by the Euro-Anglo institutions since 2010. The idea has been rapidly spreading fundamentally in Europe and America as reflected through the geographical distribution of the Semantic Web meetup communities, and I reckon that LOD is extending its reach to Asia. While more institutions in Asia are preparing to open up metadata or contents, we should also prepare to cultivate Asian specific concepts, incidents, and practices consciously into the LOD structure.

Providing human- and machine-readable structured metadata 

At the bare minimum, structured machine-readable Metadata are fundamentally ‘tidy data’(as explained in Matthew Lincoln’s blog). When data is tidy it has the potential to be further marked-up and encoded into machine-readable values. My technical understandings are not strong enough yet to explain neither the essentials nor the difficulties in enabling machine-readable markups and encoding services. However, as a librarian, I would recommend two essential aspects for anyone involved in LOD system planning: the translation of key data, and the inclusion of multiple language fields in the data schema.

We must embrace the multi-ethnic and multilingual characters in the Asian region when building the systems to capture structured metadata. If possible, key data such as titles, publishers or producers/creators, and authority terms should be translated into English(forgive my inclination towards Euro-Anglo-based studies) in separate fields next to those in local languages. We must also make sure the data-fields are tagged clearly if designated to Asian languages, so as to assure data be aggregated correctly. 

Human-readable(Semantic) Metadata is information appears on interfaces for end users, such as HTML web-pages, no matter it is structured or not. However, if it is structured, it helps user to assimilate information, conduct effective searches, and be aggregated(via machines) more powerfully. Thesaurus is one type of structured semantic data that works for both human and machine mapping. The Getty’s Art and Architecture Thesaurus(AAT) is widely used in the LAM sector, however, Asian-related concepts and events are incomprehensive in the well-received authority list. Intended to complement to AAT, AAA has been compiling an Asian-focused keywords list based on the strength of AAA collections. Developing an Asian-focused art thesaurus, preferably with translation in various local languages, would be useful for linking Asian contents, and assure metadata be discovered easier when LOD germinates further globally.

Making digital contents shareable 

Understanding the restrictive use of online contents in copyrights, AAA has started acquiring Creative Commons(CC) licences to AAA-produced contents. This resulted in most of AAA-programme documentations is now freely available online. Recently, we offer material-donors of AAA collections to comply with CC licences as well. CC licences allow flexible use of contents without the trouble of getting permission in every incident of content usage. It drastically encourages the referencing of(Asian) creative contents and therefore will encourage further creation of knowledge. 

It is possible for everyone in the art communities to contribute to content-sharing. If you are an(Asian) artists, or member of art groups, or among the management of organisations produce documentation for art events, I plead you to help the diffusion of Creative Common(CC) licences. When LAM organisations, such as AAA, contact you for permission to use documentations involving you or your organisation, please consider giving your CC-licence-complying consents. Together, we can make Asian art documentations more visible in the international art scene.  

Last but not least, AAA is also aware of the powerful usage of Creative Commons Zero(CC0) and is planning to acquire on AAA’s metadata. CCO-complying metadata could be freely used by any aggregator(including the leading cultural aggregators such as Europeana) and will bring our metadata in touch with users we never thought of contacting before, and divert users to our website to access the contents.

* AAA’s Collection Online - http://www.aaa.org.hk/Collection/CollectionOnline

* Open Culture Data: Opening GLAM Data Bottom-up - 

  http://mw2013.museumsandtheweb.com/paper/open-culture-data-opening-glam-data-bottom-up/ 

* Radically Open Cultural Heritage Data on the Web - 

  http://www.museumsandtheweb.com/mw2012/papers/radically_open_cultural_heritage_data_on_the_w

* Semantic Web meetup – http://semweb.meetup.com

* Matthew Lincoln’s blog on ‘tidy data’ - http://matthewlincoln.net/2015/09/23/tidy-art-historical-data.html 

* Creative Commons(CC) licences - http://creativecommons.org/licenses/by/4.0/legalcode