Hyperlink semantics for standoff markup of read-only documents   Table of contents   Indexes   STEP/SGML Standards Working Together

  Melese  Bertrand 
  Toche  Olivier 
 

Access to cultural heritage through an on-line multimedia data service

 

Application to the archive folders of France's General Inventory of Monuments and Art Treasures

 

Abstract:

  This document presents the European Aquarelle project and the missions and the documentation system of the General Inventory. It then examines one of the first applications of this research project with Aquarelle project and the missions and the documentation system of the General Inventory. It then examines one of the first applications of this research project with SGML tagging of a digital version of Inventory archive folders dealing with France's monuments and art treasures.
 

Description of the Aquarelle Project

 
 

Why Aquarelle?

 
  Sharing cultural information is the heart of the Aquarelle vision. New expectations of visitors, the increasing economic importance of cultural tourism, information requirements from new categories of users (e.g.: art-traders, publishers, cultural mediators, police authorities, etc.), mean that documentation - in a broad sense - is becoming one the major productions of museums and cultural organisations.
 
 

Aquarelle Users

 The project's principal users will be industry professionals: museum curators, urban planners, commercial publishers and researchers. All users should be able to collect the information relevant to their need or answering their curiosity, wherever the information components are located. Aquarelle will provide the user with facilities to author and manage folders as well as tools for searching information and browsing in the folders available on the network of connected servers.
 
 

Technical Objectives

 The main technical objectives of the project are the following:
 
  • Develop a resource discovery system for the cultural heritage information available in archive and folder databases
  •  
  • Provide the technical facilities supporting information access through hypertext navigation as well as information retrieval by querying
  •  
  • Design and install intermediary Aquarelle servers aiding users in formulating, expanding, refining, translating and routing queries. These servers should have also the role of managing connections and access rights
  •  
  • Design and integrate an authoring environment supporting creation of multimedia-derived products (folders), and encouraging reuse of reference information and multimedia assets available on the network
  •  Given the above stated technical objectives as well as the previously described requirements and background to our work, we can classify here the technical components of the Aquarelle project into four main categories, namely:
     
    1. The archive servers
    2. The folder servers and the folder authoring environment
    3. The access server(s)
    4. The user terminal(s)
     
     

    General architecture of the target information system:

     Aquarelle will support:
     
  • The creation and dissemination of information folders and detailed catalogues
  •  
  • Broadcasting of queries
  •  
  • Multilingual information retrieval
  •  
  • Querying databases with key-words
  •  
  • Copyright protection
  •  
     

    Specifications and standards

      The technical and documentary specifications and standards selected are TCP/IP for internal and external networks, HTML for pages of text,SGML (Standard Generalized Markup Language (ISO 8879)) for digitised content folders and the Z39.50 request protocol for access to data bases, standards ISO 2788 and 5964 for drawing up monolingual and multilingual thesauri, and theCIMI (Consortium for the Computer Interchange of Museum Information) DTD and the Inventory DTD for applications respectively relating to museums/art galleries and monuments
     
     

    Calendar and project phases

     A feasibility study was carried-out with the assistance of the European Commission during the first half of 1995 to assess user requirements and to review existing cultural heritage information systems in use in Europe.
     The main phases are:
     
  • 1996: design and creation of an experimental prototype 1997 (second quarter): in-depth checking and evaluation of the prototype 1997 (second semester): implementation of the system 1998: onsite evaluation and demonstrations
  •  

    Exploitation plan

     During the three years of the project the economic, legal, industrial and technical environment will have changed. Partners of the Project will constantly refine their exploitation plan, taking into account these changes, to ensure a straightforward deployement of an operational service at the end of the R&D Project.
     

    Project partners

     AQUARELLE has been defined and achieved by a European consortium that brings together:
     
  • Cultural institutions such as the Ministry of Culture in France (General Inventory and museum/art gallery departments), the Bibliothèque Nationale de France, the Greek Ministry of Culture, l'Istituto centrale per il Catalogo e la Documentazione (Italy), the Royal Commission on the Historical Monuments of England, the Museum Documentation Association (UK), Benaki Museum (Greece), the Alinari agency and the multimedia publisher Giunti (Italy)
  •  
  • Information technology companies: BULL, Ergomatic Consultants, EUROCLID, GRIF (France), INTRASOFT (Greece), FINSIEL (Italy), System Simulation Ltd (UK)
  •  
  • Research bodies of the European economic interest grouping ERCIM, including INRIA (France), ILSP and Institute of Computer Science - Foundation for Research and Technology-Hellas (Greece), CNR-CNUCE (Italy), Rutherford Appleton Laboratory (UK)
  •   The project is co-ordinated byERCIM (European Research Consortium for Informatics and Mathematics) and backed by the European Commission within the framework of the "telematics applications, information engineering" programme.
     

    France's General Inventory of Monuments and Art Treasures

     
     

    Department missions and documentary system

      The General Inventory is part of the Ministry of Culture's research and documentation department whose task it is to make an inventory of France's monuments and art treasures, to study them and make them better known. Since it was set up in 1964 the Inventory has set out documentation on France's monument and art treasure heritage working from on-site surveys carried out by regional departments. Each inventoried work or group of works gives rise to a "monograph" or " Inventory content folder". All monographs or work content folders are archived in a content folder of format A4 paper. Inventory content folders are kept regionally in binding folders. They are copied in the form of microcards. In all there are some 12.000 units of Inventory content folder microcards, or approximately 1.200.000 content folder pages.
     Monographs or content folders have a pre-established recurring structure, always identical and including all or part of the following items:
     
  • an indexing data sheet grouping key word areas (designation of the work, localisation, materials and techniques, structure and legal status) free text areas (history of the work, description, observations) and fields of date or digital type. This data sheet enables the setting up the content folder's various indices, record cards and current headings. It is used again in the Inventory's national database, named MERIMEE for monuments, and PALISSY for art treasures
  •  
  • a summary document , possibly broken down into sections, chapters, paragraphs and able to include tables, diagrams or illustrations
  •  
  • the list of the consulted documentation (the work's archival or bibliographical sources)
  •  
  • maps for localisation , an old and recent cadastral plan
  •  
  • a series of illustrations , comprising:
  •  
  • a general plan of the site or overall plan to locate the various buildings
  •  
  • surveyor's plans (elevations, sections, etc.)
  •  
  • reproductions of old documents (engravings, postcards, etc.)
  •  
  • Recent photographic prints (aerial views, views of the overall site, external views, internal views, detailed views)
  •  All illustrations are captioned with titles, keys, numbers, and photographs credits. The vocabulary used is that of a thesaurus (possibly multilingual) providing definitions, synonyms, associated terms, generic terms, discarded terms, etc.
     Properties are given a map reference (Lambert projection) and can therefore be located anywhere on the map. Items are given a map reference corresponding to the co-ordinates of the building containing them.
     There are several types of content folders that are filed together according to a hierarchical system and grouped per index content folder:
     
  • Study area binding folder (for a given programmed geographical area), e.g.: COGNAC AND SURROUNDING AREA
  •  
  • content folder of large assembly, e.g.: TOWN
  •  
  • collective content folder (for a given « group of buildings »), e.g.: HOUSES
  •  
  • individual building content folder, e.g.: PARISH CHURCH
  •  
  • item content folder, e.g.: MAIN ALTAR
  •  
  • item sub-content folder, e.g.: ALTAR PIECE
  •  
     

    Consulting Inventory content folders

     Whereas Inventory content folders can at present be consulted only at the site of conservation or by means of microcards of which several sets exist, the dissemination and exploitation of databases is possible from remote sites, online, on the domestic public network Transpac, via Minitel or, more recently, via Internet by means of the Ministry of Culture's Web server. This is how, thanks to the SPIC project, ("public service of information on cultural heritage", project selected through the Ministry of Industry's call for tenders relative to "information super highways"), the MERIMEE database, which indexes all Inventory building content folders, as well as the full list of Inventory monuments in France, can now be consultedvia the Web (http://www.culture.fr/cgi-bin/mistral/merimee).
      An analysis of printed content folders shows that the items that generally make up an Inventory content folder are increasingly being processed electronically, even before being printed. This is the case for all documents produced by word processing for unstructured information, or by structured editing tools for structured information (indexing instructions, bibliography, captioning illustrations). This is becoming the case for photographs for which the department is launchingan ambitious image digitising programme , financed by the delegation for regional development planning (DATAR), and graphical and cartographic documents printed from drawing or geographical data systems software (S.I.G.).
      The problem remained of finding a tool= capable of uniting all these disparate items, of taking into account the contents of Inventory content folders as well as their structure, the links between them and their relation with the territory. Lastly, a crucial point, this tool had to be independent of commercial software in such a way as to guarantee standardised and long term archiving.
     
     

    Creation of digital content folders and test model on Cognac content folders

     ISO Standard 8879 SGML enables the definition of the logical structure of disparate document types, such as Inventory content folders, independently of the software and databases being used. For the purposes of dissemination, SGML documents can easily be converted to HTML.
      Within the framework of a research agreement made between the Ministry of Culture andINRIA (National Institute of Research for informatics and automation) , a reference model, has been designed and developed by the firm Euroclid, via the Web, from electronic multimedia content folders produced by the Inventory's regional department for the administration of cultural affairs in the Poitou-Charentes region. The General Inventory department wanted to check on the feasibility of transposition of its content folders to SGML and to test systems likely to be able to replace, in time, the current system of reproduction and consultation of archived content folders by micrographics.
     Some twenty content folders of different hierarchical levels (generality, collective and individual content folders) and of varying content (involving either buildings or items), relative to the two cantons of Cognac North and Cognac South in the Charente region were chosen for the test, captured and reassembled. Texts were extracted from several of the department's databases (MERIMEE ,PALISSY ,Illustrations ) or came fromWord files; the bibliography was transferred from UNIMARCPsilog acquisition software; maps and plans were produced by CAD (usingCorelDraw ); photographs were provided in multiple resolutions after digitisation onKODAK PHOTO CD .
      TheDTD of Inventory binder folders ( DTD CI) was defined by the firm Euroclid, after an analysis of the Inventory's data system. This DTD has enabled the formalising of a strict editing and organisation framework for Inventory content folders. This will be used as anexchange format for setting up computer tools for the production, management, operation and exchange of Inventory content folders which are provided by the firm Grif within the framework of the Aquarelle project.
     It is important to note that this DTD defines both the logical generic structure of content folders and the way in which a content folder must be tagged in order to be exchanged within the network and interpreted by various tools.
     This exercise has resulted in several advantages:
     
  • it helps to specify the structure of content folders with greater precision and identify certain shortcomings
  •  
  • it provides a content folder exchange format that is independent of the different tools used
  •  
  • it provides a worldwide accepted electronic format using the SGML standard
  •  A rough draught of the Inventory content folder tree, such as was selected in the demonstration model, is given below:
     
  • Inventory binding folder_study area "Cognac and surroundings"
  •  
  • generalities, observations, illustrations (political and physical map)
  •  
  • Inventory binding folder_communal "Cognac centre"
  •  
  • generalities and observations
  •  
  • illustrations (political map and figures), documentation
  •  
  • 2 collective_study area content folders (hotel/houses/farms and distillery/wine-making plant), 2 building assembly content folders (town and square)
  •  
  • 9 individual building content folders
  •  
     
  • Hôtel de voyageurs, Department store, Chabot Hotel, Manor house hotel, 2 Houses, Distillery
  •  
  • St Léger Priory with assembly or individual item content folders
  •  
  • item content folders (Painting of Blanchard, Pulpit, altar piece with statues)
  •  
  • Monument to Francis 1st
  •  
  • statue item content folder.
  •  
  • Inventory content folder_ "Periphery" study area
  •  
  • generalities, observations, illustrations (political and physical map), documentation
  •  
  • collective content folder_Town halls/Schools study area
  •  
  • Inventory binding folder_communal "ARS"
  •  
  • generalities, illustrations
  •  
  • 3 individual building content folders. (Church with item content dossiers (Altar piece and before altar, Baptismal fonts)), Château and house
  •  These combined, assembled and tagged documents were then placed on the experimental Aquarelle server and transferred to the O2 /Web object-oriented multimedia document management and consultation software from the company O2 Technologies. It can be consulted online at address http://aquarelle.inria.fr/Inventaire using an HTML browser.
     This experimental platform prefigures the consultation of all content folders of the General Inventory via Internet and not just the indexing items present in the MERIMEE and PALISSY databases.
     This enables:
     
  • direct access to any type of content folder or merely to certain types of content folder from any computer on the network
  •  
  • very flexible navigation between all content folders, whatever their type (generalities, collective, individual), object (monument or item) or content (structured or unstructured documents, images, maps and plans, graphical surveys)
  •  
  • the possibility of links via clickable image maps
  •  
  • direct access to specific parts of a content folder: survey type (overall plan), category of photograph (old views), bibliographical heading (archive documents), etc.
  •  Furthermore, this study offered the opportunity for serious thinking with regard to the possible use of the cartography to consult department documentation.
     

    DTD of Inventory content folders and the Aquarelle project

     The DTD drawn up was adopted for the prototype of the European Aquarelle project for multimedia telematic access to cultural heritage information in order to standardise heritage data types (Cf. supra). A new version of this DTD has been created by GRIF.
     This version identifies two levels of content folders:
     
  • Firstly, a series of recurring items enabling geographic areas to be described with increased precision: referred to as binding folders
  •  
  • Secondly, another structure for recurring elements describing buildings, items, assemblies or themes: referred to as content folders
  •  In addition, a distinction is made between two types of links:
     
  • Internal reference links; within a content folder
  •  
  • External reference links (Aquarelle link); from one content folder to another
  •  Furthermore, it is possible to set up links from any point in a content folder.
     An automatic index generation system ("full text") is integrated in the SGML documents. Every content folder is accompanied by a header enabling it to be described and is useful for its management by the system ("folder metadata").
     Grif has been assigned the task of creating a complete electronic content folder creation environment, based on software that the company currently publishes, of which Grif SGML Editor. The department will be testing a production line of electronic content folders on two new pilot sites and examine methods for the retrospective digitising of paper content folders already archived, either working from microcards (without content recognition) or from A4 format pages.
     Meanwhile, an in-house working group has been assigned the task of re-evaluating the model and of proposing improvements. Directives will be made to all regional Inventory departments to systematically backup all electronic documents that have been produced (texts, images, graphic documents) in order to prepare the move from conventional paper archiving to electronic archiving.

    Hyperlink semantics for standoff markup of read-only documents   Table of contents   Indexes   STEP/SGML Standards Working Together