The Semantics of Collaborative Tagging System
Milorad Tošić and Valentina Milićević
University of Niš,
Faculty of Electronic Engineering,
Intelligent Information Systems Lab – InfosysLab,
Ul. Aleksandra Medvedeva 14, PoBox 73, 18000 Nis,
Serbia and Montenegro
mbtosic@yahoo.com , valentina@elfak.ni.ac.yu
Abstract. In this paper, we adopt a system-oriented approach to the collabora-
tive tagging and define it as a set of interactions in the system of Web re-
sources. First, the system of Web resources is modeled as a set of interacting
agents and collaborative tagging is represented as concurrent initiation of inter-
actions between agents in the system. Also, we define concept of knowledge for
individual agents. Later we use concepts of interaction and knowledge to give
definition of a Link. Then, for a given Universal Set of Resources, we introduce
Tag Cloud System (TCS) and definition of (possibly fuzzy) collections of re-
sources. Finally, we introduce concept of Class, based on projection of collec-
tions of resources in the TCS, to lay down some of the groundwork towards
TCS-based type system.
Introduction
In its essence, Web is all about resource locators (URLs), resource identifiers
(URIs) and resource names (URNs) [1] distilling resource as one of the most funda-
mental concepts of the Web. Until recently, Web was considered only within its
original hyper-text framework: web pages are network retrievable text documents,
easy to render for human visual consumption, that may contain hyper-links to other
web pages. However, massive adoption of the Internet and particularly broadband
“last mile”, have changed the very nature of the Web that has now been declared
“Web as platform”. So, Web is not anymore for human eyes only but it is also Web of
data. Two different technological and philosophical methodologies are the most visi-
ble now days: Semantic Web [2] and Web 2.0 [3]. In spite of the impression that
some tension exists between these two communities, we consider Semantic Web and
Web 2.0 as two sides of the same coin addressing the same gap between how current
technology is applied and the new opportunities. The difference is in the philosophy –
general vs. simple: Semantic Web is based on a firm theoretical background and pur-
sues a rigorous, generic top-down approach. In the same time, Web 2.0 is extremely
flexible, based on ultimately simple, easy to use and easy to understand stuff, adopts
bottom-up approach and worships architecture of participation (services get better as
the number of users increases), collective intelligence and long tail model [4].

2 Milorad Tošić and Valentina Milićević
In [5], authors are concerned with knowledge acquisition for software develop-
ment, and accordingly they define tagging as chinking and indexing knowledge ac-
quisition dialogue using structures that are relevant to software development. How-
ever, the collaborative tagging is more traditionally considered within a framework of
strategies that can be used in order to classify and organize content [4,6,7]. The classi-
fication strategies are characterized by several distinguishing attributes: If each item
may be associated to exactly one category then the strategy is exclusive. If each cate-
gory belongs to a more general one until the root of the tree is attained then the strat-
egy is hierarchical. Strategy that is exclusive and hierarchical is called taxonomy.
One of the typical examples of the taxonomy is the hierarchical directory set up by
Yahoo Inc. as an impressive attempt to grow a kind of universal Web taxonomy. Tag-
ging system is a non-hierarchical and non-exclusive strategy where each item is being
assigned a list of keywords, called tags. All the tags are at the same level. The tagging
systems are further classified by means of who defines the set of words or phrases that
may be used as tags and who assigns tags to items. The set of tags may be defined by
a central authority, such as editor or a librarian, or may include any word composed of
letters. Tags may be assigned to items by the same central authority or by community.
For example, in the ACM Computing Classification System [8], the central authority
defines the set of keywords that may be used to classify a paper while author of the
paper assigns selected keywords to the paper1. Collaborative Tagging (Folksonomy)
is an ad hoc classification scheme that Web users invent as they surf to categorize the
data they find online. Consequently, it is anarchic (the choice for the keywords are
not restrained by any central authority but may be any string of alphanumeric charac-
ters) and democratic (the tagging is performed by a large ensamble of people, and not
by a central one) [7]. Social software – software that enables users to share informa-
tion and collaborate online – makes these tags available to other users, who can than
take advantage of all this tagging to search for the information they need [4]. This ap-
proach has become increasingly popular, and some Web sites (call them Web 2.0 or
not?) maintain tag cloud, a list of all tags used on the domain usually with a visual in-
dication of individual tag’s popularity. The collaborative filtering is a democratic
method of classification that does not require tags to be words only. The collaborative
filtering exploits user access patterns to link items to people who use it [7].
We can identify three orthogonal dimensions of the concept of scripting language:
1) Language characteristics that identify a programming language as a scripting lan-
guage (weak typing or even no typing at all, reflection and introspection, etc.); 2) Sys-
tem that is programmed by the scripting language (in the case of OS shell scripting
languages the system is set of OS commands, while in the case of MSVisualBasic the
system is composed of a set of registered ActiveX and/or COM components); and 3)
Application under development. In this paper, we are focused on theoretical founda-
tions for the second aspect, i.e. we envision Tag Cloud System as a system that will

1
To the best of our knowledge, there is no tagging system like ACM CCS where set of possible
keywords is defined by a central authority while readers assign the list of tags (or at least are
allowed to edit it) to the paper instead of author of the paper.