Int J Digit Libr (2004) 4: 8–12 / Digital Object Identifier (DOI) 10.1007/s00799-003-0069-1
COLLATE – Acollaboratory supporting research on historic
Europeanfilms
Ulrich Thiel1, Holger Brocks1, Ingo Frommholz1, Andrea Dirsch-Weigand1, Ju¨rgen Keiper2, Adelheit Stein1,
Erich J. Neuhold1
1Fraunhofer IPSI, Darmstadt, Germany
e-mail: {thiel, brocks, frommholz, dirsch, stein, neuhold}@ipsi.fraunhofer.de
2Deutsches Filminstitut – DIF, Frankfurt am Main, Germany
e-mail: keiper@deutsches-filminstitut.de
Published online: 23 July 2004 –  Springer-Verlag 2004
Abstract. In the COLLATE project, we aim to design
and implement a Web-based collaboratory for archives,
scientists, and end users working with digitized cultural
material. Our example domain is the historic film docu-
mentation comprising digitized material about European
films of the early 20th century. Designed as a content-
and context-based knowledge working environment for
distributed user groups, the COLLATE system supports
both individual work and collaboration of domain experts
who are analyzing, evaluating, indexing, and annotat-
ing material in the data repository. The system provides
appropriate task-based interfaces for indexing and anno-
tating. As a multifunctional means of in-depth analysis,
annotations can be made individually but also collabora-
tively, for example in the form of annotation of annota-
tions. Combining results from manual and automatic in-
dexing procedures, elaborate content- and context-based
information retrieval mechanisms can be applied.
Keywords: Cultural heritage – Collaboratories – Digital
libraries – Annotations – Context-based retrieval
1 Introduction
VariousWeb-based collaboratories [8] with advanced dig-
ital library functions have been employed since the early
1990s, mainly in the natural sciences, but in the arts
and humanities we mostly find only systems with limited
functionality. One reason might be that work processes in
the social sciences are different from the procedures es-
tablished in the natural sciences and engineering and re-
quire appropriate system designs. However, the process of
compiling arguments, counterarguments, examples, def-
initions, and references to historical source material –
which is the prevailing method in the humanities – may
derive particular benefit from an electronic environment
that improves the capacity and reach of the individual
knowledge worker (cf., e.g., [6, 11]).
Another aspect of collaboratories is their ability to en-
able virtual teams to work together almost as if they were
at the same location. Although many (informal) contacts
between cultural archives constitute specific professional
communities, they so far lack effective technology support
for collaborative knowledge working. The World Wide
Web can serve both as communication platform for such
communities and as gateway for document-centered work
in such digital libraries and archives (e.g., [2]).
The EU-funded project COLLATE – Collaboratory
for Annotation, Indexing and Retrieval of Digitized His-
torical ArchiveMaterial1 – started in the fall of 2000 (IST-
1999-20882) andwill run for 3years [3, 7].An international
team of content providers, film domain experts, and tech-
nology providers works together to develop a new type of
collaboratory in the domain of cultural heritage. The im-
plemented system offers access to a digital repository of
historic text archivematerial documenting film censorship
practices for several thousand European films from the
1920s and1930s. For a subset of significantfilms it provides
enriched context documentation including selected press
articles, film advertising material, digitized photos, and
some film fragments. Major film archives from Germany,
Austria, and the Czech Republic provide the sources and
work as pilot users with the COLLATE system.
2 COLLATE’s goals and approach
Designed as a content- and context-based knowledge
working environment for distributed user groups, the
COLLATE system supports both individual work and
collaboration of domain experts who are analyzing, eval-
uating, indexing, and annotating the material in the data
repository. The example application focuses on historic
film documentation, but the developed tools are designed
to be generic and as such adaptable to other content do-
1 http://www.collate.de/

U. Thiel et al.: COLLATE – A collaboratory supporting research on historic European films 9
mains and application types. This is achieved by model-
based modules; exchanging, for instance, the annotation
types DTD allows us to support different collaboration
structures. Other domain ontologies and document de-
scription types can be plugged in as well.
The system provides appropriate task-based inter-
faces for indexing/annotation and collaborative activities
such as preparing a joint multimedia publication or as-
sembling and creating material for a (virtual) exhibition,
contributing unpublished parts of the film scientists’ work
in the form of extended annotations and commentaries.
Appropriate knowledge management tools, e.g., indexing
aids and domain-specific controlled vocabularies, have
been developed jointly by system developers and film do-
main experts, thus exploiting the benefits of a participa-
tory design. Using the tools for manual cataloging and
indexing, users create a growing body of metadata with
a special focus on subject indexing and content-based an-
notations of documents. Automatic indexing of textual
and pictorial parts of a document can be invoked to re-
ceive suggestions for index terms from the system. In add-
ition, users can rely on the support from automatic layout
analysis for scanned documents, which allows for the an-
notation of individual segments. Annotations are a cen-
tral concept in COLLATE. As a multifunctional means
of in-depth analysis, annotations can be made individu-
ally but also collaboratively, for example in the form of
annotation of annotations, collaborative evaluation, and
comparison of documents.
The system exploits user-generated metadata and
annotations by advanced XML-based content manage-
ment and retrieval methods. The final version of the on-
line collaboratory will integrate cutting-edge document
preprocessing and management facilities, e.g., XML-
based document handling and semiautomatic segmenta-
tion, categorization, and indexing of digitized text docu-
ments and pictorial material. Combining results from the
manual and automatic indexing procedures, elaborate
content- and context-based information retrieval mech-
anisms can be applied.
3 COLLATE system features
Collaboration support in the COLLATE working envi-
ronment makes use of some contemporary groupware
products and additional system functions based on an
explicit model of collaborative indexing and annota-
tion. Through interrelated free-text annotations users
can enter into a (direct or indirect) discourse on the in-
terpretation of documents and document passages, e.g.,
adding information, interpretations, arguments, etc. Pos-
sible relations between annotations can be predefined
or inferred by the system in order to represent the dis-
course structure. Additionally, explicit communication
about the interpretation of contents and the interrela-
tion of annotations are supported by a built-in discussion
forum and in the final system version by an intelligent
dialog/collaboration manager.
The COLLATE collaboratory is a multifunctional
software package integrating a large variety of function-
alities that are realized by cooperating software modules.
It comprises several databases and document represen-
tation schemata. XML is used as the uniform internal
representation language for documents in the repository
and the associated metadata as well as for the implemen-
tation of the communication protocol among its system
modules. An XML-based content manager is responsible
for the integration of knowledge processing methodology
and retrieval functionality in the system.
The main modules of the COLLATE system architec-
ture are (see [5] for further details):
– Three document preprocessingmodules for digital wa-
termarking of documents (copyright and integrity wa-
termarks), intelligent, automatic document structure
analysis and classification, and automatic, concept-
based picture indexing and retrieval.
– A distributed multimedia data repository comprising
digitized text material, pictorial material such as pho-
tos and posters, and digital video fragments.
– Tools for the representation and management of the
metadata, the XML-based content manager incorpo-
rating an ontology manager and a retrieval engine,
which are implemented as SOAP-basedWeb services.
– A collaborative task manager for complex individ-
ual and collaborative tasks such as indexing, anno-
tation, comparison, interlinking, and information re-
trieval, including tools for online communication and
collaborative discourse between the domain experts
and other system users.
– The user interface of COLLATE comprises several
workspaces for various tasks performed by distributed
user groups and user types allowing for different ac-
cess rights and offered interface functions. The final
system versionwill be generated semiautomatically by
exploiting knowledge from the underlying task model
and the user-specific dialog history.
4 Collaboration in the COLLATE environment
Digital libraries offer new opportunities for collaboration
and communication that were unfeasible in traditional
libraries [13]. Our goal is to develop a cultural collabo-
ratory, supporting interpretative work on mostly textual
material. Our starting point for collaborative work com-
prises already existing data in the form of binary image
representations of the digitized source documents. There-
fore, we do not focus on cooperative data acquisition but
rather on collaborative content-based indexing.
In COLLATE, we go beyond the mere replication
of traditional domain-specific workflows by providing
a comprehensive model of the various COLLATE domain
objects and their potential interrelations. Our notion

10 U. Thiel et al.: COLLATE – A collaboratory supporting research on historic European films
of task-guided collaboration includes the recognition of
structures as well as relations between different types of
annotations. By taking the users’ roles, tasks, and goals
into account we aim to provide comprehensive support for
the various levels of indexing.
The COLLATE system supports asynchronous collab-
oration in indexing for nontechnical users. In our under-
standing, the domain objects (scanned documents, meta-
data) represent the main focus of collaborative work, i.e.,
collaboration is performed through annotating the digi-
tized artifacts or their associated metadata objects.
The interrelations between the various domain objects
can either be unspecified or modeled in a more explicit
way by defining specific types of admissible relations. In
addition, certain communicative acts on the metalevel
(e.g., request for clarification) are part of the COLLATE
collaboration model.
In addition, the users’ individual tasks and goals have
to be taken into account for modeling a collaborative sys-
tem. Content-based indexing of a specific document, in
this sense, can be considered as a global task that can
be decomposed into partial tasks. In the COLLATE con-
text, the result of these partial tasks, which are to be
performed by various users, is the value-added informa-
tion in the form of metadata objects associated with the
original document. But these partial tasks are only rarely
performed in isolation. On the contrary, in most cases
a specific annotation will be part of a thematic thread,
e.g., some newsgrouplike discussion about a certain topic.
Digital signatures are employed to ensure authenticity of
the individual contributions as well as the chronological
order of the annotations.
But to properly represent scientific discourses, espe-
cially in the arts and humanities, annotations have to con-
sist of more than unstructured and uncontrolled text that
comments on another domain object (binary image file,
XML metadata object). For this reason we have devised
a comprehensive model of discourse structure relations
between (a) binary image versions of the original docu-
ment and annotations and (b) discussion threads realized
as annotations on annotations.
Our document-centered discourse model is loosely
based on the theory of discourse structure relations. Even
though they have been developed for monologs in the
linguistic context of text coherence, we think that dis-
course structure relations can also be adapted to describe
admissible relations between various data and metadata
objects in the COLLATE context, especially annotations
or comments. In particular, we employ a specific subset of
relations ranging from factual to more interpersonal lev-
els, i.e., focusing on certain qualities of the participants
of a discourse. Figure 1 illustrates the discourse structure
relations as used in the COLLATE system (originating
from text linguistics, cf. [9, 10]).
A detailed discussion of the discourse structure rela-
tions can be found in [4]. The seamless transition from
factual to interpersonal discourse structure relations de-
picted in Fig. 1 also corresponds to the illocutionary as-
pects of an annotation, i.e., the specific communicative
intention its author had in mind at the time of creation
(e.g., from stating factual information to active participa-
tion in a discussion thread).
Figure 2 displays a fictious example discourse about
the partial ban of the movie Kuhle Wampe by Berthold
Brecht.
Even though discourse structure relations proved ad-
equate for modeling the interrelations between annota-
tions, it turned out that there were some relevant prag-
matic aspects of collaborative indexing work that had
not yet been covered. In the next section we describe
how discourse structure relations can be complemented
Fig. 1. Discourse structure relations
Fig. 2. Example discourse

U. Thiel et al.: COLLATE – A collaboratory supporting research on historic European films 11
by communicative acts [12–14] to introduce metacommu-
nication, i.e., explicit communication about domain ob-
jects, in a seamless way.
Conceptually, discourse structure relations and com-
municative acts can be considered as complementary:
communicative acts focus on illocutionary aspects of
a specific dialog situation (cf., e.g., [1, 14]), whereas dis-
course structure relations describe characteristic rela-
tionships between assertive acts, e.g., annotations or
comments.
But on closer inspection it becomes evident that some
communicative acts might invoke certain types of dis-
course structure relations between the corresponding an-
notations. In our view, the set of discourse structure rela-
tions adopted for COLLATE can be considered as specific
instances of comments, i.e., they are treated as assertive
communicative acts.
From this perspective, we can regard explicit collab-
oration in the context of the COLLATE project as the
combination of specified relation types between annota-
tions, i.e., discourse structure relations, that are comple-
mented by a certain set of admissible acts for metacom-
munication (at the dialog level) referring to the various
Fig. 3. The COLLATE system
types of COLLATE domain objects (e.g., annotations,
cataloging information).
At its current stage of development, the COLLATE
prototype system already supports various tasks like cat-
aloging, indexing (structured, free), and nested annota-
tions using discourse structure relations (Fig. 3). Users
can select a whole document, a document page, or an
area of a document for cataloging, indexing, and anno-
tation. Furthermore, the prototype offers simple search
options based on filmographic information as well as on
cataloging, keywords, and annotations.
5 Context-based retrieval of documents
Appropriate search and retrieval functionality represents
a fundamental requirement for enabling users to access
a cultural digital library in a reasonable way. To allow for
advanced content- and context-based search, the docu-
ments in the digital collection must be indexed by content
and subject matter.
In context-based retrieval, not only the document it-
self, but also the actual context of this document is con-

12 U. Thiel et al.: COLLATE – A collaboratory supporting research on historic European films
sidered. In the case of COLLATE, this means that we
are dealing with the discourse context. Various domain
objects (digitized documents, metadata represented in
XML, annotations) are given unique URIs and interre-
lated by RDF statements. These statements are typed
according to the discourse structure relation they repre-
sent. With this information, we then know the specific
type of an annotation with respect to its context, e.g., an
elaboration or an example.
Taking a second look at Fig. 3 one can see that only
the inclusion of the associated annotations would yield it
as relevant for a query like “political background”. But it
also becomes evident that the annotations within a cer-
tain discourse context cannot be treated in isolation, e.g.,
the second counterargument weakens the statement it
comments upon in this context.
The introduction of discourse structure relations al-
lows for novel retrieval options with respect to the dis-
course context. They can be used to create a ranking of
relevant documents according to user queries. Depending
on the specific type of its connecting relation, an anno-
tation can possibly raise or lower the overall relevance
weight of its discourse context. An algorithm for this kind
of context-based retrieval is presented in [5].
6 Conclusion
The COLLATE system represents a new type of col-
laboratory supporting content- and concept-based work
with digital document sources. Innovative task-based in-
terfaces support professional domain experts in their in-
dividual and collaborative scholarly work, i.e., analyz-
ing, interpreting, indexing, and annotating sources. The
metadata provided in this way are managed by an ad-
vanced XML-based content manager and an intelligent
content- and context-based retrieval system.
References
1. Allen J, Ferguson G, Stent A (2001) An architecture for more
realistic conversational systems. In: Proceedings of Intelligent
User Interfaces (IUI ’01). ACM Press, New York
2. Bentley R, Horstmann T, Sikkel K, Trevor J (1995) Support-
ing collaborative information sharing with the World-Wide
Web: the BSCW shared workspace system. In: Proceedings
of the 4th international WWW conference, Boston, December
1995, pp 63–74
3. Brocks H, Thiel U, Stein A, Dirsch-Weigand A (2001) Cus-
tomizable retrieval functions based on user tasks in the
cultural heritage domain. In: Constantopoulos P, Sølvberg I
(eds) Proceedings of the European conference on research
and advanced technology for digital libraries (ECDL 2001),
Darmstadt, Germany, September 2001. Lecture notes in com-
puter science, vol 2163. Springer, Berlin Heidelberg New York,
pp 37–48
4. Brocks H, Stein A, Thiel U, Frommholz I, Dirsch-Weigand A
(2002) How to incorporate collaborative discourse in cultural
digital libraries. In: Proceedings of the ECAI
2002 workshop on semantic authoring, annotation and
knowledge markup (SAAKM02), Lyon, France, July
2002
5. Frommholz I, Brocks H, Thiel U, Neuhold E, Iannone L,
Semeraro G, Berardi M, Ceci M (2003) Document-centered
collaboration for scholars in the humanities – the COLLATE
system. In: Proceedings of the European conference on re-
search and advanced technology for digital libraries (ECDL
2003), Trondheim, Norway, August 2003. Lecture notes in
computer science, vol 2769. Springer, Berlin Heidelberg New
York, pp 434–445
6. Kahan J, Koivunen MR, Prud’Hommeaux E, Swick RR
(2001) Annotea: An open rdf infrastructure for shared web
annotations. In: Proceedings of the WWW10 international
conference, Hong Kong, May 2001, pp 623–632
7. Keiper J, Brocks H, Dirsch-Weigand A, Stein A, Thiel U
(2001) COLLATE – a web-based collaboratory for content-
based access to and work with digitized cultural material.
In: Bearman D, Garzotti F (eds) Proceedings of the Interna-
tional Cultural Heritage Informatics meeting (ICHIM ’01),
Milan, Italy, September 2001, Politecnico di Milano, Milano,
pp 405–511
8. Kouzes RT, Myers JD, Wulf WA (1996) Doing science on the
Internet. IEEE Comput 29(8):40–46
9. Maier E, Hovy EH (1993) Organising discourse structure re-
lations using metafunctions. In: Horacek H, Zock M (eds)
New concepts in natural language processing, Pinter, London,
pp 69–86
10. Mann WC, Thompson SA (1987) Rhetorical structure theory:
a theory of text organization. In: Polanyi L (ed) The structure
of discourse, Ablex, Norwood, NJ, pp 85–96
11. Nichols DM, Pemberton D, Dalhoumi S, Larouk O, Belisle C,
Twindale MB DEBORA: Developing an interface to support
collaboration in a digital library. In: Borbinha JL, Baker T
(eds) Proceedings of the 4th European conference on research
and advanced technology for digital libraries (ECDL 2000),
Lisbon, Portugal, September 2000. Lecture notes in com-
puter science, vol 1923. Springer, Berlin Heidelberg New York,
pp 239–248
12. Searle JR (1979) A taxonomy of illocutionary acts. In: Searle
JR (ed) Expression and meaning: studies in the theory of
speech acts, Cambridge University Press, pp 1–29
13. Sitter S, Stein A (1992) Modeling the illocutionary aspects
of information-seeking dialogues. Inf Process Manage 28(2):
165–180
14. Stein A, Gulla JA, Thiel U (1999) User-tailored planning of
mixed initiative information-seeking dialogues. User Model
User-Adapted Interact 9(1–2):133–166