漢文電子佛典的緣起與遠景
(數位化文獻/多價文獻模式)
惠敏法師
一九九九年六月二十七日
呂美智、林佩琪、李志強摘要整理
參考資料：
"Applications of A New Document Model for Digitalization of East Asian Classical Documents", Howie Lan(藍效農) Instructional Technology Program Uninversity of California, Berkeley
Digital Library Project: University of California at Berkeley http://elib.cs.berkeley.edu/
"Multivalent Documents: Anytime, Anywhere, Any Type, Every Way User-Improvable Digital Documents and Systems", Thomas A. Phelps, A dissertation for Ph. D, University of California, Berkeley (1998)
"Multivalent Documents: A New Model for Digital Documents", Robert Wilensky Thomas A. Phelps, Technical Report, CSD-98-999 (1998)
"Multivalent Documents: Inducing Structure and Behaviors in Online Digital Documents", Thomas A. Phelps & Robert Wilensky, Proceedings of Hawaii International Conference on System Sciences '96 (Best Paper Award, Digital Documents Track)
運用新文獻模式資料，整理後推廣到數位化圖書館計劃(digitallibrary project)，介紹有關國外的技術、搜集、行政。
多價文獻 ( 多訊息層次文獻模式 )
新、舊文獻不同的特性：
1.印刷文獻:Monolithic (單塊)
2.數位文獻:Multivalent(多價)
Multivalent View of Document(多價文獻模式)提供給下列人士參考：
l.決策者
2.專業管理者
3.技術執行者
Traditional View of Document(文獻的傳統觀念)
Physical Media(物理性質媒介)
l.Clay tablets(泥版)
2.Oracle bones(甲骨文)
3.Bamboo strips(竹簡)
4.Paper(紙本)
Traditional View of Document(文獻的傳統觀念)的缺點:
l.Equivalents of books(與書同義)
2.Basic Units are Stand-along and Disconnected (基本單元是獨立且不互相連結的)
3.Time consuming to search and find informations across basic units(互參、對照、查詢資料費時)
4.For human to "read" only( 只提供 給 人" 讀 " 的功能)
The Multivalent Document (數位化文獻是「多價文獻 」 )
可提供多元性的價值與功能
information entity
content & functional factor
layers & behaviors
Layer:(同形式情報層)
Behaviors:(運用於layer的功能性處理因子)
incrementally and openly
"having many values, meanings, or appeals", relating to the multifarious benefits to the user of the new document model.
an atomic analogy.Atoms bond with one another by sharing their outermost,or valence, electrons, forming molecules with useful emergent properties.
behaviors and layers unite to form single conceptual documents with powerful emergent capability from simple parts, but with more extensive sharing and interaction among components.
Layer Examples:
Scanned text image Layer ( 掃描 原版 圖擋層次 )
Encoded text data Layer (編碼版層次)
Font reconstructed Layer (字庫再現版層次)
Character geometries Layer
Meta data Layer
Future content Layer
Examples of Within-layer Behaviors:
in scanned text image layer﹕
Local enlargement
Copy image part
Paste image part
in encoded text layer﹕
Character finding
String sorting
Character counting
in reconstructed display layer﹕
Show Character with font
Set Typeface
Horizontal text display
Vertical text display
Examples of Cross-layer Behaviors:
Synonym lens同義字透鏡
Translation翻譯
Multivalent View of Document(多價文獻的優點)
基本單位﹕字、筆劃、圖素(pixel)
重組(rearranged)、連結(connected) 簡易快速
Processor + Human Mind( 同時可提供 給 機器與人" 讀 " 的功能)
若被傳統的文獻概念所限，則會造成以下的影響
不會連結不同的資料層次
<例如>“Dead”page images( 認為 文件圖檔 單只是圖檔 )連結部分資料層次，但僅限於單一的軟體
<例如>OCR software's post-process(光學字元辨視軟體的後置作業)
Multivalent Document Characteristics:
多價文獻三特色
1. Incremental Extendable(逐加延伸)
在以下幾個方面，具有逐加延伸的特性
1.Data
2.Layers
3.Behaviors
2. Structurally Distributed(結構性分散管理)
可利用結構性分散管理
Layers can be at different locations in network(各層級可分散放置於網路上不同的位置)
例如：
1. 大正藏1~55(編碼檔)：CBETA
2. 大正藏56~100(編碼檔)：SAT
3. 高麗藏(圖檔)：海印寺，韓國
4. 手抄古本(圖檔)：京大、東大
5. 房山石經(圖檔)：北京圖書館
3. Internally Complementary(內部的相互補足)
相輔相成
1. Layers or behaviors may not be perfect. (各層和表現方面無法完美)
2. Some layers may be implemented less accurate than the others.(部份的層級展現上可能較不精確)
3. The multivalent document framework best complementary from different components of the content to contribute to the document as a whole.(多價文獻可由不同的物件加以補足而臻至圓滿)