background image
DigiCULT
.
Info
20
F
ollowing the seminar on the applica-
tion of the EAD DTD for cultural
heritage purposes in archives and librar-
ies, held two years ago in the National
Library of France, the Directorate of the
Archives of France is organising a European
Conference on 7 and 8 October 2004
on data-processing software for encoding
research aids for archives and authority data
into XML. The last European Conference
on the subject took place at the Public
Record Office in Kew in October 1999.
EAD, whose first public version (1998)
was revised in September 2002, is a docu-
ment type definition founded on the same
principle of multilevel description as the
General International Standard Archival
Description, ISAD(G). EAD contains ele-
ments that are equivalent to MARC fields
(an EAD inventory aid may be linked with
a MARC note and, conversely, an EAD
notice may be converted into MARC).
E
AD is now well established in North
America and in the rest of the world,
and it has attained technical stability while
continuing to keep abreast of technologi-
cal advances. Training courses have been
organised and handbooks and application
guidelines have been produced. Several
institutions using the EAD in France,
the UK, Germany, The Netherlands and
Spain have developed innovative solu-
tions for encoding, editing and display-
ing their searching tools . It is now time to
assess progress and to facilitate the exchange
of experience. Since 2001 another DTD,
EAC (Encoded Archival Context), has been
under development. Perfectly compatible
with the International Standard Archival
Authority Record for Corporate Bodies,
Persons, and Families ≠ ISAAR (CPF) ≠
and complementary to the UNIMARC/
Authorities format, the EAC combines bib-
liographic authority records and archival
authority records, which give information
about both the creator and the context of
creation of archival material.
V
ersion 1.0 of the EAC Tag Library
should shortly become available. Tests
have been carried out on the conversion
of authority data into EAC/XML for-
mat, in particular within the LEAF project
(Linking and Exploring Authority Files):
local authority data will be download-
ed from the local servers of the participat-
ing organisations to a central system which
links automatically the authorities belong-
ing to the same entity. The participants will
present their research results and ideas on
different issues:
∑ what is required for EAD implementa-
tion: thoughts on standardisation, training;
∑ EAD implementation: tools for creating
EAD documents, interoperability, com-
patibility with other standards;
∑ publishing EAD/XML documents: edit-
ing tools, documentary management
applications with documents stored on
a Web server in native XML;
∑ the EAC DTD and the first examples of
authority records encoded into XML.
INTRODUCTION
F
rench archives, like other archives serv-
ices around the world, own millions of
pages of documents containing handwritten
information which are difficult for mem-
bers of the public to access. Even if archi-
vists always built indexes and research tools,
the quantities are so huge (the basic count-
ing unit is a kilometre of shelving) that
there are still a lot of documents which are
both difficult and time-consuming to find,
both from the shelves and within the doc-
ument itself, if the reader is looking for a
specific page. At the same time, a growing
number of people, such as geneaologists,
are interested in these documents. How
can archives offer public access to millions
of pages of documents with handwritten
information, when no research tools exist?
Archives started to scan documents; how-
ever, even with digitised versions it is still
often necessary to leaf (virtually) through a
considerable number of images. Even if the
page required has been digitised, the time
needed to find it can still be enough to dis-
courage the user to the point of giving up.
T
o solve this problem, systems must be
defined to allow document retriev-
al by content. For this to be achieved it
is necessary to associate annotations to
the images of documents. Annotations for
archival documents can be geometric (a
position in the image represented by a rec-
tangle or a polygon) or textual (e.g. a date,
a place, a name, or a keyword). Of course,
textual annotations can be linked to geo-
metric ones, making it possible to select the
appropriate images automatically.
E
UROPEAN
C
ONFERENCE
ON
E
NCODING
R
ESEARCH
A
IDS
M
AKING
H
ANDWRITTEN
A
RCHIVES
D
OCUMENTS
ACCESSIBLE
TO
THE
P
UBLIC
WITH
A
D
OCUMENT
I
MAGE
A
NALYSIS
S
YSTEM
B
ERTRAND
C
O‹ASNON
(IRISA/INRIA),
J
EAN
C
AMILLERAPP
(IRISA/INSA),
I
VAN
L
EPLUMEY
(IRISA/INSA)