=Paper=
{{Paper
|id=Vol-1275/paper2
|storemode=property
|title=Collaborative Semantic Tables
|pdfUrl=https://ceur-ws.org/Vol-1275/swcs2014_submission_2.pdf
|volume=Vol-1275
|dblpUrl=https://dblp.org/rec/conf/semweb/GoyMPS14
}}
==Collaborative Semantic Tables==
https://ceur-ws.org/Vol-1275/swcs2014_submission_2.pdf
Collaborative Semantic Tables
Anna Goy, Diego Magro, Giovanna Petrone, Marino Segnan
Dipartimento di Informatica, Università di Torino
{goy,magro,giovanna,marino}@di.unito.it
Abstract. The scenario defined by current Web architectures and paradigms
(such as Cloud Computing), poses challenges and opportunities to users. On the
one hand, they have to manage huge amounts of digital resources handled by
different applications in spite of their possibly related content; on the other
hand, they are enabled to share knowledge and participate to content creation.
The interaction of these two aspects provided a great impulse to collaborative
resource management: in this paper we present T++, an environment that ex-
ploits semantic knowledge about digital resources in order to face these chal-
lenges, by providing an integrated and smart management of heterogeneous in-
formation objects.
Keywords: Collaborative Workspaces · Ontology-based Content Management
· Personal Information Management · Linked Data · Semantic Technologies.
1 Introduction
The current ICT scenario, and in particular Web architectures and paradigms, are
posing new challenges to Personal Information Management [4]. Many aspects of
human-computer interaction have been influenced; in this paper, we focus on the most
relevant for our approach: (a) Users have to deal with a huge number of heterogene-
ous resources, stored in different places, encoded in different formats, handled by
different applications as belonging to different types (images, emails, bookmarks,
documents, ...), despite their possibly related content. (b) Web 2.0 and, more recently,
Cloud Computing, and in particular the Software-as-a-Service paradigm, have en-
hanced the possibility of user participation in content creation on the Web, as well as
the possibility of resources and knowledge sharing. The interaction of these two as-
pects provided a great impulse to user collaboration in managing shared resources.
In this paper we present Semantic Table Plus Plus (Sem T++), an environment
aimed at supporting users in collaborative resource management. In particular, in Sem
T++, two types of semantic knowledge are modeled: (1) knowledge about information
objects, i.e., information resources as such, and (2) knowledge about their content.
The goal of this paper is to show how collaborative annotations, based on the formal
semantic representation of these two types of knowledge, support users in the organi-
zation, retrieval and usage of shared digital resources.
The rest of the paper is organized as follows: in Section 2 we discuss the motiva-
tions of our approach and the main related work; in Section 3, we briefly present T++,
and in Section 4 we describe its semantic enhancement, by presenting its architecture
and the semantic model underlying it, and showing how the system supports users in
collaboratively handling semantic descriptions of digital resources. Section 5 con-
cludes the paper by discussing open issues and future developments.
2 Related Work
A survey and a discussion of existing Web-based applications supporting collabora-
tion, including groupware and project management tools or suites, can be found in
[10], a previous paper introducing T++. As far as the approach presented in this paper
is concerned, one of the most relevant research areas is represented by Kaptelinin and
Czerwinski's book Beyond the Desktop Metaphor [12], which contains an interesting
presentation of the problems of the so-called desktop metaphor, and of the approaches
trying to replace it. Within this framework, one of the most interesting models dis-
cussed in the mentioned book is Haystack [13], a flexible and personalized system
enabling users to define and manage workspaces referred to specific tasks. Another
interesting family of approaches are those grounded into Activity-Based Computing
(e.g., [3], [19]), where the main concept around which the interaction is built is user
activity. Also [9] and [16] propose a system supporting lightweight informal interac-
tions, as well as larger and more structured collaborative projects, by relying on activ-
ity-based workspaces, handling collections of heterogeneous resources.
Strategies exploited to organize resources have been studied within the field of
Personal Information Management; in particular, multi-facets classification of re-
sources has been taken into account: resources can be tagged with meta-data
representing different aspects (facets), leading to the creation of folksonomies, bot-
tom-up classification models collaboratively and incrementally build by users [6].
Interesting improvements of such tagging systems have been designed by endowing
them with semantic capabilities (e.g., [1]), in particular in the perspective of know-
ledge workers [14].
Another important research thread, aiming at coupling desktop-based user interfac-
es and Semantic Web, is represented by the so-called Semantic Desktop [17]. In par-
ticular, the NEPOMUK project (nepomuk.semanticdesktop.org) defined an open-
source framework, based on a set of ontologies, for implementing semantic desktops,
focusing on the integration of existing applications, in order to support collaboration
among knowledge workers. Finally, [7] presents an interesting model connecting the
Semantic Desktop to the Web of Data.
3 The Starting Point: Table Plus Plus
The Table Plus Plus (T++) project, described in [10] and [11], proposes an interaction
model supporting users in collaboratively handling digital resources, based on the
metaphor of tables, populated by objects. T++ is characterized by the following main
features.
Tables as thematic contexts. In T++, users can define shared workspaces devoted
to the management of different activities. Such workspaces are called tables and sup-
port users in the separated, coherent and structured management of their activities.
Users can define new tables, at the preferred granularity level; for instance, a table
can be used to manage a work project, to handle children care, to plan a journey.
Workspace awareness. Workspace awareness is supported by three mechanisms:
(a) On each table, a presence panel shows the list of table participants, highlighting
who is currently sitting at the table; moreover, when a user is sitting at a table, she is
(by default) "invisible" at other tables (selective presence). (b) Standard awareness
techniques, such as icon highlighting, are used to notify users about table events (e.g.,
an object has been modified). (c) Notification messages, coming from outside T++ or
from other tables, are filtered on the basis of the topic context represented by the ac-
tive table (see [2] for a more detailed discussion of notification filtering).
Collaboration. An important aspect of T++ tables is that they are collaborative in
nature, since they represent a shared view on resources and people; "tables represent
common places where users can, synchronously or asynchronously, share informa-
tion, actively work together on a document, a to-do list, a set of bookmarks, and so
on" [10, p. 32]. The most peculiar aspect of T++ tables is the collaborative manage-
ment of table resources: table participants, in fact, can (a) modify objects, delete them,
or add new ones; (b) invite people to "sit at the table" (i.e., to become a table partici-
pant); (c) define meta-data, such as comments and annotations (see below).
Heterogeneous objects management and workspace–level annotations. Objects ly-
ing on tables can be resources of any type (documents, images, videos, to-do items,
bookmarks, email conversations, and so on), but T++ provides an abstract view over
such resources by handling them in a homogeneous way. Table objects, in fact, are
considered as content items (identified by a URI) and can be uniformly annotated (by
visibility labels, comments, and tags).
Within the T++ project, we developed a proof-of-concept prototype, consisting in a
cloud application (a Java Web App deployed on the Google App Engine) accessible
through a Web browser. The current version exploits Dropbox and Google Drive API
to store files corresponding to table objects and Google Mail to handle email conver-
sations. We are investigating the availability of open API provided by other common
file sharing and online editing tools in order to improve interoperability and to enable
users to configure the preferred tools to be exploited for object sharing and editing.
In [11] we also reported the results of a user evaluation of T++ in which we asked
users to perform a sequence of pre-defined collaborative tasks (communication, re-
source sharing, and shared resources retrieval) using standard collaboration tools and
using T++. The results showed that, with T++, performing the required tasks is faster
and user satisfaction is higher.
4 The Semantic Enhancement of T++
4.1 Architecture
On the basis of T++, we designed an enhanced version, Sem T++, in which semantic
knowledge plays a major role in supporting resource management on tables. Fig. 1
shows the relevant components of Sem T++ architecture:
Semantic Knowledge Manager: it manages the semantic descriptions of table ob-
jects (stored in the Semantic KB) and invokes the Reasoner, when required. More-
over, it handles the connection with GeoNames (www.geonames.org), as described
in Section 4.2. In the current prototype, it uses the OWL API library (owla-
pi.sourceforge.net) to interact with the ontologies and the Semantic KB, and the
GeoNames Search Web Service (www.geonames.org/export) to query GeoNames.
It interacts with the following knowledge bases (written in OWL:
www.w3.org/TR/owl-features):
─ Table Ontology: it represents the (static) system semantic knowledge concerning
information objects.
─ Geographic Ontology: it represents the (static) system semantic knowledge con-
cerning geographic entities and features.
─ Semantic KB: it contains all the facts about the individuals involved in the se-
mantic representation of table objects.
Reasoner: it provides the system with "new" object features which can be exploited
to support the user in table object management. The current proof-of-concept pro-
totype uses Fact++ (owl.cs.manchester.ac.uk/tools/fact).
Object Manager: it manages the "used objects" (i.e., objects on a table or included
in objects on a table) and the references to elements used in the interaction with the
user (e.g., available object types and object properties, corresponding to Table On-
tology classes and relations). The Object Manager plays a mediation role between
the Table Manager (and thus, indirectly, the UI) and the components which
represents the system "intelligence", i.e. the Semantic Knowledge Manager and the
Smart Object Analyzer.
Smart Object Analyzer: it is a service that provides the Object Manager with the
analysis of table objects, in order to discover information about them; for example,
it looks for parts included in the analyzed object (e.g., images, links, etc.). In the
current prototype, it exploits a Python Parser Service, able to analyze HTML doc-
uments1.
1
In the current prototype we focused on HTML documents since they are very common,
quite easy to parse, and their semantic characterization introduces interesting aspects (e.g.,
the need of distinguishing between information content and content of the HTML file).
Fig. 1. Sem T++ architecture (relevant components).
4.2 Semantic Model
The core of our proposal is the Table Ontology, which models knowledge about in-
formation resources. It is grounded in the Knowledge Module of O-CREAM-v2 [15],
a core reference ontology for the Customer Relationship Management domain devel-
oped within the framework provided by the foundational ontology DOLCE (Descrip-
tive Ontology for Linguistic and Cognitive Engineering) [5] and some other ontolo-
gies extending it, among which the Ontology of Information Objects (OIO) [8]. The
Table Ontology enables us to describe digital resources lying on tables as information
objects, with properties and relations; for instance: a table object can have parts (e.g.,
images within a document), which are in turn information objects; it can be written in
English; it can be stored in a PDF file, or it can be an HTML page; it has a content,
which usually has a main topic and refers to a set of entities (i.e., it has several objects
of discourse). Given such a representation, reasoning techniques can be applied, in
order to infer interesting and useful knowledge; for example, if a document contains
an image of Monte Bianco, probably the document talks about Monte Bianco.
The most relevant class in the Table Ontology is InformationElement: all table ob-
jects are instances of it. Moreover, we introduced some specific subclasses (e.g., Doc-
ument, Image, Video, Audio, EmailThread, etc.), to provide a more precise characteri-
zation of the different types of objects that can lay on a table. In order to characterize
such classes, we relied on: (1) a language taxonomy defined in O-CREAM-v2,
representing natural, formal, computer, visual languages; (b) a set of properties (some
of them inherited from O-CREAM-v2). A complete account of such properties is out
of the scope of this paper; in the following we just mention the most important ones:
DOLCE : part(x, y, t) to represent relations such as the one between a document
and an image or a hyperlink included in it.
specifiedIn(x, y, t) to represent relations such as the one between a document and
the language it is written in (e.g., Italian).
hasTopic(x, y, t) to represent the relation between an information element (e.g., a
document) and its main topic.
hasObjectOfDiscourse(x, y, t) to represent what a resource "talks about" (it is a
subproperty of OIO : about).
identifies(x, y, t) to represent, for instance, the relation between a hyperlink and
the resource it points to.
An important fragment of the proposed semantic model refers to particular properties
which model candidate relationships. The idea is that the system, endowed with suit-
able axioms, can infer candidate features, mainly from included objects; for example,
the Reasoner can infer that Monte Bianco is a candidate object of discourse of a doc-
ument from the fact that the document itself includes an image of it. When the Rea-
soner infers such candidate relationships, the system asks the user for a confirmation:
if (and only if) the user confirms, for instance, that Monte Bianco is actually an object
of discourse of the document (doc), then a new relation hasObjectOfDiscourse(doc,
MonteBianco, t) is added to the knowledge base. Analogous axioms are available for
the hasTopic relation, to support the inference of hasCandidateTopic relationships.
Besides knowledge about information objects, all tables in Sem T++ are equipped
with geographic knowledge. We decided to model geographic knowledge for two
reasons: (a) Together with time (currently not modeled in Sem T++), space represents
a cross-domain feature, which is at least partially represented by geographic
knowledge2. (b) Geographic knowledge represents for us a testbed, i.e., an example of
knowledge that characterizes the "content" of table objects: in the future, tables could
be equipped with semantic knowledge about specific domains, by relying on the same
mechanism we used for geography.
In order to provide tables with geographic knowledge, we exploited GeoNames, a
huge, open geographical database containing over 10 million geographical entities.
Moreover, we equipped the system with a Geographic Ontology and a Vocabulary
Mapping, defining correspondences between Geography Ontology classes/properties
and GeoNames feature classes/codes. GeoNames entities, in fact, are categorized into
feature classes (e.g., A, corresponding to the generic concept of country or region)
and further subcategorized into more specific feature codes (e.g., A.ADM1, corres-
ponding to the concept of primary administrative division of a country). For each
topic3 mentioned on a table, the Semantic Knowledge Manager searches for corres-
ponding GeoNames entities. If one or more results are found, the system currently
asks the user to select the correct one, if any4; then a new individual is created in the
Semantic KB, as instance of the Geographic Ontology class identified through Geo-
Names feature class/code (thanks to the Vocabulary Mapping). For example, an in-
stance of the class Mountain is created for the topic Mont Avic (a mountain in Val
d'Aosta, a northwestern Italian region). Moreover, a new mapping between such an
instance and the corresponding GeoNames entity is created in the Semantic KB (see
2
Geographic knowledge here means commonsense competence such as the ability to georefe-
rence places, or knowing that Monte Bianco is a mountain, and not scientific knowledge.
3
In the current prototype, we consider only topics, although also objects of discourse could be
taken into account.
4
Mechanisms to support a (partially) automatic disambiguation are under study.
Instances Mapping in Fig. 1), thus making all information available in GeoNames
(e.g., its location on a map) also available on Sem T++ tables.
The described model enables table participants to specify and combine different se-
lection parameters in order to find objects on a table. For example, to get all email
threads talking about Monte Bianco trekking (i.e., having it as main topic), the user
can specify the following parameters: topics={Monte_Bianco_trekking},
types={emailThread}. Each parameter value corresponds to a user selection; object
types are references to Table Ontology classes (emailThread in this example); values
for other properties, e.g. Monte_Bianco_trekking, are references to individuals in the
Semantic KB. Moreover, the user could provide more general queries, such as asking
for all resources talking about mountains, thanks to the facts that topics (e.g., Monte-
Bianco) are represented as instances of classes in the Geographic Ontology (e.g.,
Mountain). User queries are handled by the Semantic Knowledge Manager, which
accesses the Semantic KB and invokes the Reasoner, in order to provide the objects
matching the parameters.
4.3 Collaboratively Handling Semantic Descriptions of Digital Resources
One of the most challenging aspects of the presented semantic model is the creation
and update of semantic representation of table objects. In fact, when a new object is
created, or when an existing one is modified (e.g., when a table participant includes a
new image or link in it), the corresponding semantic representation must be created or
updated. In the following we will see how the collaborative management of semantic
descriptions of table objects represents a step to face such a challenge, and, in particu-
lar, how this makes T++ tables actual shared semantic spaces.
Consider the new object case (the update case works in an analogous way): table
participants can create new objects from scratch (e.g., when they start writing a new
document), or they can add an existing resource (e.g., a bookmark pointing to a Web
site) to the table. In both cases, the system builds a new semantic representation in
four steps, in which different components play their role:
1. Smart Object Analyzer: some properties (e.g., mereological composition, types of
the parts, formats) are automatically determined by the Smart Object Analyzer.
2. Semantic Knowledge Manager/Reasoner: other properties (e.g., candidate topics
see Section 4.2) are inferred by the Reasoner.
3. User: some of the inferred properties need a "user confirmation" (e.g., candidate
topics); moreover, users can add properties (typically objects of discourse).
4. Semantic Knowledge Manager/GeoNames: the selected topic is linked to the cor-
responding geographic entity (through GeoNames), if any.
In order to describe the sketched process into more details, consider a usage scenario
in which Maria participates in a table concerning the activities of a small ONG for
environment safeguard, Our Planet, together with some other volunteers. Maria has to
write an article for an online local newspaper, discussing the situation of a local old
mule track in Champorcher (a small municipality in Val d'Aosta). Maria creates a new
table object (an HTML document), writes some text in it, adds a picture of the envi-
ronments and a hyperlink to a resolution by the Municipality of Champorcher con-
cerning an enlargement project for the mule track. When Maria clicks the
"save&update" button, the creation of the object semantic representation is triggered.
Fig. 2. Semantic representation of Maria's article.
In step 1 the Smart Object Analyzer discovers that the object is encoded in
HTML/UTF-8, written in Italian, and that it contains two parts: an image and a hyper-
link. In step 2 the Reasoner infers candidate topics and candidate objects of dis-
course. In particular, Maria can: select suggested candidates; select items within the
list of values already present on the table and associated to other objects; add new
topics/objects of discourse from scratch (step 3). The final semantic representation of
her article could be the one shown in Fig.2. In step 4, for the new topic (Champorch-
er) a corresponding GeoNames entity is searched and found, thus a new instance
(linked to GeoNames through a new entry in Instances Mapping) is created in the
Semantic KB.
Fig. 3 (left-hand side) shows a mockup of Sem T++ user interface; on the left, table
participants can access the main functionalities available on the table in focus, i.e.
access to table objects and three basic services: Chat (for synchronous communication
among table participants), Blackboard (for asynchronous communication), and shared
Calendar. Other functionalities, e.g., settings, are available through the "menu" in the
left-bottom corner. On the right, there is the list of tables and the list of participants at
the table in focus (which is highlighted): participants not currently sitting at the table
are in light grey. The central area shows the panel triggered by a click on the "objects"
label on the left, listing table objects. Objects can be visualized by clicking on their
names. Moreover, for each object, the visibility label is reported ("pub" means public,
i.e., visible to all table participants; objects could also be private, i.e., visible only to
their creator), together with an access to comments (if available), to the semantic de-
scription of the object ("sem prop" column), and a delete button.
When Mario, another volunteer, sits at the Our Planet table, he is notified about
the new object created by Maria (new/modified object names are highlighted). Mario
looks at its semantic description, by clicking on the button in the "sem prop" column:
a new panel, shown in Fig. 3 (right-hand side), is displayed. From this panel, it is
possible to change the object type (an example scenario is described below) and to
add/delete semantic properties, with the exception of the "contains" property, which
refers to DOLCE : part(x, y, t): object parts (e.g., images included in a document) can
be modified only by editing the object.
Fig. 3. Sem T++ user interface (mockup).
Mario finds that one of the objects of discourse (Ayasse river), up to him, is mislead-
ing: he thus decides to remove it and to add a new object of discourse (Ourty alpine
pasture). When Mario clicks the "save&update" button, the semantic description is
updated. Moreover, the table suggests Mario to write a comment in order to explain
his decision to modify it (see Fig. 4, right-hand side); such a comment is attached to
the modified object, but handled by the table, and is thus immediately accessible to
other participants, as they sit at the Our Planet table.
A very similar case occurs when a user adds to a table an object by classifying it in
a way which is then modified by another table user. For example, Maria could have
added an image which, actually, is the scan of a document, and Mario could decide to
change the object type from Image to Document. Again, Sem T++ suggests him to use
the table–level annotation mechanism in order to share the reasons of the change.
This very simplified usage scenario shows the support provided by Sem T++ to the
collaborative handling of semantic descriptions of table objects. Sem T++ has been
designed having in mind a very "democratic" model of collaboration, in which people
sitting at a table have the same privileges and thus all of them can modify objects, and
in particular their semantic representations. However, Sem T++ also provides effec-
tive and easy–to–use table–level annotation mechanisms which represent a significa-
tive support to such a collaborative activity, together with the other standard mechan-
isms enabling discussion and communication among table participants, i.e., the
Blackboard and the Chat (see [11] for details).
The coordination of collaborative activities taking place on Sem T++ tables de-
serves a final comment. As far as possible concurrent resource editing (e.g., document
modification) is concerned, Sem T++ relies on existing applications, which typically
handle issues related to collaborative editing (such as Google Drive). As regards ob-
ject management (adding, deleting, modifying table objects) and collaborative seman-
tic annotation, the default configuration is the already mentioned simple policy allow-
ing every table participant to freely modify, add, or remove all objects and annota-
tions. Obviously, there are specific contexts (e.g., when tables are used to handle
business activities, such as project management, for example) requiring more struc-
tured policies, modeling different user roles, privileges and specific workflows. We
are working on the definition of mechanisms enabling users to configure such policies
on each table.
5 Conclusions and Future Work
In this paper we presented Sem T++, an environment supporting users in collabora-
tive resource management, by showing how formal semantic knowledge about infor-
mation objects and their content can support an integrated, user-friendly management
of heterogeneous shared resources. Sem T++ is a work in progress and many aspects,
concerning the collaboration model and the user interface have not been discussed
here. In particular, we are designing configuration mechanisms enabling users to de-
fine specific policies to coordinate collaborative activities, and specifically collabora-
tive annotation management [18], on each table. Moreover, we are going to complete
the integration of semantic modules into T++ prototype, in order to plan a user evalu-
ation aimed at testing how the semantic model presented in this paper supports users
in organizing and retrieving information objects on tables, but also the effort required
to create and update semantic descriptions of table objects.
References
1. Abel, F., Henze, N., Krause, D., Kriesell, M.: Semantic Enhancement of Social Tagging
Systems. In: Devedžić, V., Gašević, D. (eds.) Web 2.0 & Semantic Web, pp.25–56.
Springer, Heidelberg (2010)
2. Ardissono, L., Bosio, G., Goy, A., Petrone, G.: Context-Aware Notification Management
in an Integrated Collaborative Environment. In: UMAP 2009 workshop on Adaptation and
Personalization for Web2.0, pp.23–39. CEUR-WS (2010)
3. Bardram, J. E.: From Desktop Task Management to Ubiquitous Activity-Based
Computing. In: Kaptelinin, V., Czerwinski, M. (eds.) Beyond the Desktop Metaphor,
pp.223-260. MIT Press, Cambridge, MA (2007)
4. Barreau, D.K., Nardi, B.: Finding and reminding: File organization from the desktop.
ACM SIGCHI Bulletin 27 (3), 39-43 (1995)
5. Borgo, S., Masolo, C.: Foundational choices in dolce. In: Staab, S., Studer, R. (eds)
Handbook on Ontologies, Second Edition, pp. 361–381. Springer, Heidelberg (2009)
6. Breslin, J. G., Passant, A., Decker, S.: The Social Semantic Web. Springer, Heidelberg
(2009)
7. Drăgan, L., Delbru, R., Groza, T., Handschuh, S., Decker, S.: Linking Semantic Desktop
Data to the Web of Data. In: Aroyo, L., Welty, C., Alani, H., Taylor, J., Bernstein, A.,
Kagal, L., Noy, N., Blomqvist, E. (eds.) The Semantic Web – ISWC 2011. LNCS, vol.
7032, pp 33–48. Springer, Heidelberg (2011)
8. Gangemi, A., Borgo, S., Catenacci, C., Lehmann, J.: Task Taxonomies for Knowledge
Content. Metokis Deliverable D07 (2005)
9. Geyer, W., Vogel, J., Cheng, L., Muller M. J.: Supporting Activity-Centric Collaboration
through Peer-to-Peer Shared Objects. In: Group'03, pp.115–124. ACM Press, New York,
NY (2003)
10. Goy, A., Petrone, G., Segnan, M.: Oh, no! Not Another Web-based Desktop!. In: 2nd Int.
Conf. on Advanced Collaborative Networks, Systems and Applications, pp. 29–34. XPS
Press, Wilmington, DE (2012)
11. Goy, A., Magro, D., Petrone, G., Segnan, M.: A Cloud-Based Environment for
Collaborative Resources Management. Int. J. Cloud Applications and Computing 4(4), in
press (2014)
12. Kaptelinin, V., Czerwinski, M. (eds.) Beyond the Desktop Metaphor. MIT Press,
Cambridge, MA (2007)
13. Karger, D.R.: Haystack: Per-User Information Environments Based on Semistructured
Data. In: Kaptelinin, V., Czerwinski, M. (eds.) Beyond the Desktop Metaphor, pp. 49–100.
MIT Press, Cambridge, MA (2007)
14. Kim H., Breslin J.G., Decker S., Choi J., Kim H.: Personal Knowledge Management for
knowledge workers using social semantic technologies. Int. J. of Intelligent Information
and Database Systems 3(1), 28–43 (2009)
15. Magro, D., Goy, A.: A Core Reference Ontology for the Customer Relationship Domain.
Applied Ontology 7(1), 1–48 (2012)
16. Muller, M. J., Geyer, W., Brownholtz, B., Wilcox, E., Millen, D. R. . One-Hundred Days
in an Activity-Centric Collaboration Environment based on Shared Objects. In: CHI'04,
pp.375–382. ACM Press, New York, NY (2004) Goy, A., Petrone, G., Segnan, M.: Oh,
no! Not Another Web-based Desktop!. In: 2nd Int. Conf. on Advanced Collaborative
Networks, Systems and Applications, pp. 29–34. XPS Press, Wilmington, DE (2012)
17. Sauermann, L., Bernardi, A., Dengel, A.: Overview and Outlook on the Semantic Desktop.
In: 1st Ws on The Semantic Desktop at ISWC 2005, vol. 175. CEUR-WS (2005)
18. Su, A.Y., Yang, S.J., Hwang, W.Y., Zhang, J.: A web 2.0-based collaborative annotation
system for enhancing knowledge sharing in collaborative learning environments.
Computers and Education 55(2), 752–766 (2010)
19. Voida, S., Mynatt, E. D., Edwards, W. K.: Re-framing the Desktop Interface Around the
Activities of Knowledge Work. In: UIST'08, pp. 211-220. ACM Press, New York, NY
(2008)