This is an archive of an inactive wiki and cannot be modified.

An integrated view to medieval illuminated manuscripts

Contact e-mail: aisaac #


General purpose and services to the end user

The purpose of this application is to provide the user with access to 2 collections of illuminated manuscripts from different libraries, accessible online at and .The descriptions of these two collection follow different metadata schemes, and contain values from different controlled vocabularies for subject indexing. The user should however be able to search for items from the two collections using his preferred point of view, either using vocabulary from collection 1 or vocabulary from collection 2.

Functionality examples

The main feature of the application is collection browsing. In the original collection websites, a user can access documents using a hierarchical view of the vocabulary used for to describe the collection. The integrated view should allow the same, but with any of the original vocabularies giving access to items from both collections.

The application provides also search based on free text queries: documents can also be retrieved based on free-text querying of the different fields used to describe the documents (creator, place, subject, etc.). For the subject indexing, if a text query matches one controlled vocabulary concept, the documents indexed against this concept will be returned.

Application architecture

The application is accessible via a web browser. It practically contains 3 elements undistributed (though a more distributed approach is feasible):

The images coming from the two collections are not centralized. They are accessed from a distance on their respective servers (HTTP queries)

Special strategies involved in the processing of user actions

Browsing uses hierarchical links in vocabularies: if a concept matching a query have subconcepts, the documents indexed against these subconcepts should be returned. The same will hold for mapping links between the two vocabularies: depending on the link (equivalence or specialisation) between the query concept and another one, documents indexed by this concept shall be returned.

Integration between vocabulary-linked functions and other application functions

One access mode is free-text searching, which in case of subject-search requires matching a text query with one of the labels of concepts from controlled vocabularies

Additional references



Iconclass and Mandragore analysis vocabulary (Iconclass is described in EucIconclass)

General characteristics (size, coverage) of the vocabulary

16000 subjects, tailored to describe what can be found in the Mandragore manuscripts (persons, events, places, abstract concepts, objects, etc.). The vocabulary has been developed and is owned by the Manuscript Department of the French National Library.

Language(s) in which the vocabulary is provided

The only language is French.

Vocabulary extract

The vocabulary comes in the form of a hierarchy:

      .zoologie (généralités)
        bernard.l'ermite  [crustacé]
        crabe  [crustacé]
        crevette  [crustacé]
        écrevisse  [crustacé]

Here is the information accessible for a specific descriptor:

Descripteur :   cochon  [mammifère ongulé]
Formes rejetées : porc, truie
Classement thématique : .mammifères

Structure explanation

Images are indexed using a set of descriptors (actually, almost all the elements from the global vocabulary) that form a flat list: there are no (formal) links between descriptors. However, some structure is given by topic classes which organize the descriptors according to general domains (as shown in the two first levels of the hierarchy in 2.4). A topic class is defined by a label and an optional list of subclasses. A descriptor is specified by a label itself (“cochon” in 2.4), optional rejected forms (“porc”) an optional definition (“mamifère ongulé”), and a reference to one or more topic classes (“.mammifères”). A note can sometimes be found.

Additional references

The hierarchical view of the vocabulary can be accessed at

Vocabulary mappings

Mapped vocabularies

Iconclass (1) and Mandragore description vocabulary (2)

Extracts of Mappings

- “25F72 molluscs” (1) is equivalent to “mollusques” (2)
- "25F711 insects" (1) is more specific than "autres invertébrés (vers,arachnides,insectes...)" (2)
- "25F(+441)  herd, group of animals" (1)  is equivalent to “troupeau” (2)
- “11U4  Mary and John the Baptist together with (e.g. kneeling before) the judging Christ, 'Deesis' ~ Last Judgement” (1) is equivalent to the combination “s.marie”, “s.jean.baptiste”, “christ” and “jugement.dernier” (2)

Types of mapping used

We can have equivalence or specialization links between concepts from both vocabularies. On concept can be linked to several other ones. A concept can be linked to a combination of concepts from another vocabulary if allowed by this vocabulary (e.g. in the Iconclass case, a subject and an applied key)