=Paper=
{{Paper
|id=Vol-369/paper-23
|storemode=property
|title=Meaning Of A Tag: A collaborative approach to bridge the gap between tagging and Linked Data
|pdfUrl=https://ceur-ws.org/Vol-369/paper22.pdf
|volume=Vol-369
|dblpUrl=https://dblp.org/rec/conf/www/PassantL08
}}
==Meaning Of A Tag: A collaborative approach to bridge the gap between tagging and Linked Data==
https://ceur-ws.org/Vol-369/paper22.pdf
Meaning Of A Tag: A Collaborative Approach to Bridge the
Gap Between Tagging and Linked Data
Alexandre Passant Philippe Laublet
Electricité de France R&D LaLIC, Université Paris-Sorbonne
1, Avenue du Général de Gaulle 28, rue Serpente
92141 Clamart, France 75006 Paris, France
alexandre.passant@edf.fr philippe.laublet@paris4.sorbonne.fr
&
LaLIC, Université Paris-Sorbonne
28, rue Serpente
75006 Paris, France
alexandre.passant@paris4.sorbonne.fr
ABSTRACT using URIs that uniquely identify resources used to annotate
This paper introduces MOAT, a lightweight Semantic Web data. Nevertherless, this process is generally a task harder
framework that provides a collaborative way to let Web 2.0 to overcome than free-tagging.
content producers give meanings to their tags in a machine-
readable way. To achieve this goal, this approach relies on This paper introduces MOAT - http://moat-project.org -,
Linked Data principles, using URIs from existing resources a framework based on Semantic Web principles designed to
to define these meanings. That way, users can create inter- bridge this gap between free-tagging and semantic annota-
linked RDF data and let their content enter the Semantic tion. Its goal is to provide a simple and collaborative way to
Web, while solving some limits of free-tagging at the same annotate content thanks to existing URIs with as little ef-
time. fort as possible and by keeping free-tagging habits. We will
first briefly introduce the well-known limits of free-tagging
Keywords and why, while a human can solve them, computers are not
Semantic Web, Web 2.0, Tagging, MOAT, Linked Data, Ar- able to do it easily. We will then introduce how we rep-
chitecture of Participation, SIOC resent the meaning of a tag and the way we model it in
a machine-understandable way. We will then present the
MOAT project, and start with related work about tagging
1. INTRODUCTION vocabularies on the Semantic Web. We will describe the
Among the various tools and principles that Web 2.0 intro-
MOAT ontology, used to model these relationships between
duced such as blogging, collaborative knowledge manage-
tags and their meanings, using URIs of existing Semantic
ment with wikis or on-line social networking, tagging is one
Web resources. Next, we will see how our framework helps
of the most interesting phenomena. While one of the main
users to take advantage of these ideas thanks to simple tools
ideas of Web 2.0 is to let users play an important role in the
and collaborative principles that ease the task of semantic
process of creating content, tagging goes a step further by
annotation. Finally, we will overview how this approach is
letting them control the way they organise it. By adding
related to the Linked Data movement.
simple keywords, or tags, to their data or to data they
browse on-line, they can decide themselves which meta-data
must be related to any content. These tags can refer to var- 2. TAGS AND THEIR MEANINGS
ious and really different levels of annotation, as Golder and 2.1 Limits of Tagging for Software Agents
Huberman [4] identified, from content meta-data (topic(s) of
While tagging became popular thanks to Web 2.0 services
a blog post) to quality meta-data (opinion about a webpage)
such as del.icio.us and Flickr, or blogging and dedicated
or even self reference. Yet, whatever the way people use it,
agregation websites like Technorati, it raises various issues
tagging raises various issues regarding information retrieval.
from an information retrieval point of view. These limits
The Semantic Web and especially semantic annotation [7] of-
mainly consists in the ambiguity and heterogeneity of tags,
fers better perspectives regarding information retrieval, by
as well as the flat organisation of folksonomies. Ambiguity
and heterogeneity may produce too much noise or silence
while the lack of relationship between tags makes difficult to
find related content from a given entry point. Thus, Mathes
argues that ”a folksonomy represents simultaneously some of
the best and worst in the organization of information” [11].
These limits cannot be easily overcome since tags, from a
Copyright is held by the author/owner(s). machine point of view, do not carry any semantics about
LDOW2008, April 22, 2008, Beijing, China. what they represent, while a human can interpret such se-
visiting online
have a completely different meaning on another one, e.g.
France ? betting ? .
del.icio.us query del.icio.us query
3. THE MOAT ONTOLOGY
3.1 Tagging and the Semantic Web
paris tagged http://www.parissportifs.com/
Various work has been done regarding tagging and the Se-
mantic Web, in which we can distinguish work related to
tagged tagged modeling tags thanks to Semantic Web technologies and
work related to mining folksonomies from ontologies [6] [15]
http://paris.fr http://fr.wikipedia.org/wiki/Paris_Hilton or linking ontologies and folksonomies [14]. Here, we will
focus on the first aspect. The Tag Ontology [13] provides
a model that introduces Tag and Tagging classes in order
to represent tags and tagging actions. Its Tag class inher-
its from skos:Concept and the model relies on FOAF[3] for
Figure 1: User perception and tag search modeling the user aspect which makes the ontology com-
pliant with existing standards. Gruber [5] defined a simi-
lar model but extends this concept by notifying the source
mantic when tagging or reading some content. For example, (i.e. the webspace) where the action takes place whereas it
when tagging a blog post ”paris”, a user has in mind an does not consider its temporal aspect contrary to the pre-
existing concept from the real-world, which can be a city, vious approach. Yet, his ideas have not been implemented.
a person, or anything else. Yet, from a computer point of The SCOT ontology [10] focuses on a way to share tags
view, there is no way to make a difference since the only by modeling tagclouds and also provides various properties
thing it will consider is a text string. Thus, when retrieving to link tags together (e.g.: synonymy, case-variation ...). Fi-
data, the user himself will have to manually deal with tags nally, while they do not define any semantic modeling, Flickr
ambiguity to find, from a resulting dataset, in which post machine tags2 allow people to embed RDF-like assertions
this tag has been used in a similar way that he had in mind within their tags.
(see Fig.1).
3.2 Classes and Properties of the Ontology
The MOAT ontology features a Tag class that extends the
2.2 Defining and Representing the Meaning of one defined in the Tag Ontology. Indeed, in our case, a Tag
a Tag Thanks to the Semantic Web instance must have a single label (thus the ontology uses
In order to represent the meaning of a tag, we first consider an OWL restriction) and its URI must respect a certain
the meaning it can be assigned to in a particular tagging pattern that is defined by a MOAT server, as we will see
context (e.g. in that post context, ”paris” means a french later. That way, it offers common URIs for tags that can be
city). Thus, we extend the tripartite model of tagging and shared across communities and social media sites.
folksnomies [12], by adding a local meaning for each tagging
action: In order to represent tag meanings, the ontology can be di-
vided in two parts. The first one, dedicated to global mean-
T agging(U ser, Resource, T ag, M eaning) (1)
ings (2), introduces a moat:hasMeaning relationship and a
moat:Meaning class, that are used to link a Tag instance
From this definition, we define the global meanings of a tag, to all its meanings. Each moat:Meaning instance features a
i.e the list of all different meanings a tag can be assigned to. unique property called moat:meaningURI in order to link to
To keep a social aspect within this definition, each meaning the URI of an existing Semantic Web resource that repre-
is related to the set of users that used it: sents the given meaning. Moreover it features at least one
foaf:maker link (using once again a cardinality contraint in
M eanings(T ag) = {(M eaning, {U ser})} (2) the ontology) in order to keep a trace of the user(s) that
defined that URI as a meaning for the tag. The follow-
ing snippet of code represents the global meanings assigned
In order to represent these meanings in a machine-readable to the tag ”paris”. Here, three different meanings have been
way, which can help to solve some of the issues raised be- represented, while one of them is shared by two users. While
fore, we think that Semantic Web, and especially URIs of the following example shows how MOAT can be used to rep-
existing resources can play an important role. Since they resent tag ambiguity, the ontology can also be used to deal
provide unique identifiers for resources of the real-world, we with hetereogenity since two different tags can share com-
believe that they are one of the most efficient way to de- mon meanings (using meaningURI), which can be helpful in
fine it, either globally or in a tagging context. Thus, in multi-lingual systems where different tags refer to the same
the two previous definitions, meanings are defined thanks resource (e.g. paris and parigi).
to URIs of existing resources, which can be part of any
knowledge base, as GeoNames1 or DBpedia [1], but also
internal corporate datasets. For example, the meaning or
the tag ”paris” can be, in a given blog post context, the
URI while it can
2
http://tech.groups.yahoo.com/group/yws-
1
http://geonames.org flickr/message/2736
tags:RestrictedTagging
Thus, the framework consists of the association of (1) a
http://tags.moat-
project.org/tag/paris MOAT server that can deliver the list of all global meanings
rdf:type tags:associatedTag
for a given tag in a community and that can be updated
tags:name by users themselves and (2) clients that provide interfaces
http://example.org/ to define and choose the meaningful URI for a given tag in
tagging/1 paris
a tagging action and then produce RDF data describing it.
moat:tagMeaning Both interact with each other using HTTP, and exchange
tags:taggedBy data thanks to the previously defined ontology. To benefit
http://sws.geonames.org/
tags:taggedResource 2988507/ from this architecture, users simply install a client on their
favourite blogging tool while subscribing to a tag server as
http://example.org/alex
they could have done with Annotea [9]. Then, users create
http://example.org/ their content and tag it as usual. As soon as the content is
post/1
saved, the client automatically queries the server to get the
list of all of the URIs associated to the given tag(s). The
user then choose which one he wants to assign to each tag in
Figure 2: Tagging and the local meaning of a tag this context - or define new ones if nothing relevant is found
- and then saves its content, which is instantanely exported
as interlinked RDF data (see Fig.3).
@prefix moat: .
@prefix foaf: .
@prefix dbpedia: . 4.2 Interaction Between Parties
MOAT clients can interact with a server both in reading and
a moat:Tag ;
writing data, respectively to retrieve meanings for a tag or
moat:name "paris" ;
to add new ones. Both actions are performed over HTTP in
moat:hasMeaning [
a RESTful way with normalized API calls3 .
a moat:Meaning ;
moat:meaningURI ;
In order to retrieve the set of meanings for a tag, clients can
foaf:maker
use various URLs to query the server. First, since the server
] ;
uses content-negociation principles, clients can simply re-
moat:hasMeaning [
quest the tag URI to get its related RDF description, as soon
a moat:Meaning ;
as they send the correct accept header to the server. More-
moat:meaningURI ;
over, it means that the tag itself carry some semantics, since
foaf:maker ;
its deferencable URI gives information about all the mean-
foaf:maker
ings it can be related to. As explained before, the tag URI
] ;
must respect a given pattern so that the server understand
moat:hasMeaning [
it, which is: tag_uri = SERVER_BASE+urlencode(tag_label).
a moat:Meaning ;
Yet, in order to ease the task of writing MOAT clients to
moat:meaningURI dbpedia:Paris_Hilton ;
developers, clients can directly request the RDF file URL,
foaf:maker
or even a JSON description of the content, available at
] .
tag_uri/json, so that they can write clients without to
deal with RDF. The description sent to the client is cre-
ated upon request thanks to a SPARQL query sent to the
The second part of the ontology defines a model for the local
triple-store which is used as a back-end of the server. That
meaning (1) of a tag. Here, we rely on the RestrictedTagging
way, a MOAT server acts as a layer between a triple-store
class from the Tag Ontology, which identifies a tagging rela-
and various clients.
tionship between a post, a user, and a single and only tag.
Thus, we introduced a moat:tagMeaning property to link a
To update the server data, i.e. add new meanings - since a
RestrictedTagging instance to the meaningful URI in this
MOAT server is initially empty when a community install it
context, as show on Fig.2.
- , clients simply send the URL of a file containing Tag in-
stances and related Meanings (that is automatically created
4. FRAMEWORK ARCHITECTURE by the client itself thanks to user actions). Depending on the
4.1 Global Architecture choice of the community, this action may require an aditional
While the MOAT ontology provides a model to define re- API key to allow this operation to be performed. When
lationships between tags and their meanings, this is not getting the file URL, the MOAT server imports it in the
enough to let users easily define these meanings, either glob- back-end store, thus merging the new meanings with already
ally or locally. In order to achieve this goal, we designed a existing ones for the given tag and adding new foaf:maker
client-architecture that (1) lets users define which URIs they statements, so that decentralized Meaning instances related
want to assign as meanings to their tags and (2) let them to a particular tag are then combined. Hence, we immedi-
choose a URI from an existing set in a given tagging context. ately benefit from the collaborative aspect of this process,
Furthermore, this process relies on an architecture of partic- since as soon as one user defines a new meaning for a tag,
ipation, since meanings are shared across a community and the whole community can reuse it.
can evolve among time thanks to users themselves, that can
3
add new meaning URIs. http://moat-project.org/server
User creates content and tag it Client queries the MOAT server
Server returns the set
for global meaning URIs
User chooses local meaning URI
tags:associatedTag
User saves the content
http://example.org/ http://tags.moat-
tagging/1 project.org/tag/paris
moat:tagMeaning
Content enters
tags:taggedResource http://sws.geonames.org/
the Semantic Web 2988507/
tags:taggedBy
http://example.org/
post/1
http://example.org/alex
Figure 3: Global architecture
4.3 Implementations
The MOAT server is currently available as a PHP5 applica-
tion4 that must be plugged on the top of a triple-store. It
uses a Connector class to provide an interface between the
server and a RDF storage system and currently features con-
nectors for ARC25 and for the 3store API6 , using SPARUL
LOAD queries to add new data.
A first client implementation was developed as a Drupal
module7 . It provides an interface with any MOAT server,
export of the tagging object as a RDF file, and uses SIOC [2]
to describe other meta-data about the tagged item. In order
to help users find new URIs for their tags, the module uses
the Sindice [16] Widget8 , as shown on Fig. 4. Moreover,
already used URIs are displayed as links so that user can
browse them. A client implementation was recently added
in Openlink Virtuoso9 and another one may be available
soon for Wordpress through the SparqlPress10 add-on. Figure 4: Drupal module interface
5. MOAT AND THE LINKED DATA WEB queries like ”Find all blog posts tagged with french cities”.
While MOAT can be mainly seen as a way to solve some Finally, it can also be used to suggest related content by
issues of free tagging by giving meaning to tags in RDF, it looking at all resources linked to a meaning URI and find
must also be considered from a Linked Data point of view. posts linked to one of this resource (Fig. 5). This example
By providing a way to link any Web 2.0 content to existing shows how, from a simple free-tagging scheme, posts could
URIs, it may help to discover content related to these URIs be finally related thanks to the way they are linked to URIs
thanks to lookup services such as Sindice. Furthermore, it and existing links between these URIs. Moreover, in order
can also be helpful to provide new tag-based search engines to provide a direct link between the tagged content and a
using Semantic Web principles that could answer advanced meaning URI, one can directly rely on the sioc:topic prop-
4 erty, thus letting SIOC further enter the Linked Data Web.
http://moat-project.org/server
5
http://arc.semsol.org
6
http://threestore.sf.net 6. CONCLUSIONS
7
http://moat-project.org/clients In this paper, we introduced MOAT, an ontology and a col-
8
http://sindice.com/dev/widget laborative framework which goal is to let users bridge the
9
http://www.openlinksw.com/virtuoso/ gap between free-tagging and semantically-annotated con-
10
http://wiki.foaf-project.org/SparqlPress tent in a simple way. This framework relies on an architec-
http://tags.moat- Semantic Web and Information Systems, 3(2), 2007.
project.org/tag/paris
[6] H. Halpin, V. Robu, and H. Shepard. The dynamics
tags:associatedTag
and semantics of collaborative tagging. In Proceedings
http://example.org/ tags:name
tagging/1 of the 1st Semantic Authoring and Annotation
paris Workshop (SAAW06), November 2006.
moat:tagMeaning
[7] S. Handschuh and S. Staab. Annotation for the
tags:taggedBy
Semantic Web. Number 96 in Frontiers in Artificial
tags:taggedResource http://sws.geonames.org/
Intelligence and Applications. IOS Press, Amsterdam,
2988507/ 2003.
http://example.org/alex
[8] C. Hayes and P. Avesani. Using tags and clustering to
sioc:topic
identify topic-relevant blogs. In Proceedings of the 1st
geonames:parentFeature
http://example.org/
post/1
International Conference on Weblogs and Social Media
(ICWSM 07), Boulder, Colorado, March 2007.
http://sws.geonames.org/ [9] J. Kahan and M.-R. Koivunen. Annotea: an open
3017382/
http://example.org/julie RDF infrastructure for shared Web annotations. In
Proceedings of the tenth international conference on
http://tags.moat-
World Wide Web, pages 623–632, New York, 2001.
moat:tagMeaning
project.org/tag/france tags:taggedBy ACM Press.
tags:associatedTag [10] H. Kim, J. Breslin, S. Yang, and H. Kim. int.ere.st:
Building a Tag Sharing Service with the SCOT
tags:name http://example.org/
tagging/2 sioc:topic Ontology. In Proceedings of the AAAI 2008 Spring
Symposium on Social Information Processing,
france
tags:taggedResource
Stanford University, California, 2008.
[11] A. Mathes. Folksonomies - Cooperative Classification
http://example.org/
and Communication Through Shared Metadata.
post/2 http://www.adammathes.com/academic/computer-
mediated-communication/folksonomies.html,
December 2004.
Figure 5: Tagging and the Linked Data Web [12] P. Mika. Ontologies are us: A unified model of social
networks and semantics. In Proceedings of the 4th
International Semantic Web Conference, ISWC 2005,
ture of participation and allows people to interlink content volume 3729 of Lecture Notes in Computer Science,
with any URI from existing resources, thus letting tagging pages 522–536, Galway, Ireland, November 2005.
and related services as blogs enter the Linked Data Web. [13] R. Newman. Tag ontology design.
Future works regarding MOAT will mainly focus on imple- http://www.holygoat.co.uk/projects/tags/, March
mentations for other platforms, as well as integrating social 2005.
networking aspects when retrieving a list of URIs from a [14] A. Passant. Using Ontologies to Strengthen
server for a given tag, so that a user could be suggested Folksonomies and Enrich Information Retrieval in
in priority to use meanings that have been assigned by one Weblogs. In Proceedings of International Conference
of his friend. Moreover, while MOAT does not currently on Weblogs and Social Media, Boulder, Colorado,
focuses on semi-automatic meaning definition for tags [17] March 2007.
or clustering [8], we think it could provide a nice machine- [15] L. Specia and E. Motta. Integrating folksonomies with
readable representation model for such work, and that it the semantic web. In Proceedings of the European
could also be combined with such approaches in the future. Semantic Web Conference (ESWC2007), LNCS, pages
624–639, July 2007.
7. REFERENCES [16] G. Tummarello, R. Delbru, and E. Oren. Sindice.com:
[1] S. Auer, C. Bizer, G. Kobilarov, J. Lehmann, Weaving the open linked data. In 6th International
R. Cyganiak, and Z. Ives. DBpedia: A Nucleus for a Semantic Web Conference, pages 552–565, 2007.
Web of Open Data. 6th International Semantic Web [17] C. M. A. Yeung, N. Gibbins, and N. Shadbolt.
Conference, November 2007. Understanding the semantics of ambiguous tags in
[2] J. Breslin, A. Harth, U. Bojars, and S. Decker. folksonomies. In Proceedings of the International
Towards Semantically-Interlinked Online Workshop on Emergent Semantics and Ontology
Communities. 2nd European Semantic Web Evolution (ESOE2007) at ISWC/ASWC2007,
Conference, May 2005. November 2007.
[3] D. Brickley and L. Miller. FOAF Vocabulary
Specification 0.91, November 2007.
http://xmlns.com/foaf/0.1/.
[4] S. Golder and B. A. Huberman. Usage patterns of
collaborative tagging systems. Journal of Information
Science, 32(2):198–208, April 2006.
[5] T. Gruber. Ontology of folksonomy: A mash-up of
apples and oranges. International Journal on