A Framework for Experience Sharing Through Contextual Tagging

Grégory Bourguin and Arnaud Lewandowski

University of Lille Nord de France, ULCO, LISIC, France

Keywords: Experience Sharing, Contextual Tagging, Activity, Multi Viewpoint, Folksonomy, Ontology.

Abstract: The web is a knowledge-sharing place where many tools allow people to share their own experience about

the resources they use. This shared experience informs about how resources have been perceived and

involved in particular contexts. Such sharing is expected to help new users in building their own working

contexts. Most of these tools involve a tagging system. Tags can help in navigating through shared

knowledge, but tags also carry semantics that can help in understanding it. In this paper, we propose a

literature review showing that tag semantics can only be fully understood while considering the context it

comes from. Our assumption is that it is possible to better link tags to their creation context. We thus

propose the EVOXEL framework, which relies on an activity-based structure and basic mechanisms that

allow reaching this objective. We then discuss its capabilities, and provide first use cases we applied to test

them.

1 INTRODUCTION

The web is a knowledge-sharing place where many

tools allow people to share their own experience

about the resources they use. This sharing can take

different forms like a comment in a repository, a

blog post, a Youtube video showing a set of

resources involved in a particular performance, or a

resources collection in a knowledge management

system like PearlTrees or Elium (formerly

Knowledge Plazza). Shared experience informs

about how resources have been perceived and used

in particular contexts. It is expected to help new

users in building their own contexts for performing

their own activities. Sharing one’s context indeed

facilitates resource appropriation, and investigating

one’s universe (what they have created, used, etc.)

helps users to assess others, and then favors

inspiration by proposing new resources and contexts

of use (Singer, 2013).

Most of these tools involve a tagging system.

Many researchers have indeed shown that tags help

in organizing knowledge (Kersten, 2012). Moreover,

they also have shown that tags themselves reflect

their creator’s experience (Saab, 2010). As a result,

tags can help in navigating through shared

knowledge, but tags also carry semantics that can

help in understanding it. Some tagging systems

consider tags as simple labels, but others define

them as more complex structures. Different works

propose ontologies designed to better represent the

knowledge carried by tags (Lohmann, 2011). One

can notice that most of the information added to tags

is related to the context in which a tag has been

created (NB: we will use the term “created” for

designating the tag creation, as well as the action of

associating a tag with an entity). Indeed, a tag

reflects its creator’s knowledge, and knowledge can

only be fully understood while considering the

context it comes from (Ning, 2012). If previous

works altogether add interesting information about

tags, our assumption is that it would be possible to

better link them to their creation context.

Another point is that if the above-mentioned

researches have shown that ontologies can be used

to bring more contextual information into tagging

systems, we can notice that tags are usually not

considered themselves as ontological types. Tags

carry semantics, and semantics definition is the

purpose of ontologies. Some sharing tools support

tags hierarchies, but they do not consider inheritance

mechanisms. It is also not possible to create tags

using inference. We believe that tagging systems

could better benefit from the power of semantic web

technologies like semantic reasoners.

To fulfill these needs, we aim at developing a

new tag-oriented framework for supporting the

sharing of end-user’s experience about their

Bourguin G. and Lewandowski A.

A Framework for Experience Sharing Through Contextual Tagging.

DOI: 10.5220/0006509502050211

In Proceedings of the 9th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (KMIS 2017), pages 205-211

ISBN: 978-989-758-273-8

resources. In the first part of this paper, we propose

a literature review related to the semantic of tags. In

the next part, we introduce the EVOXEL framework

that is centered on the tag and activity concepts,

while taking advantage of ontological mechanisms

and tools. We will finally discuss this framework’s

capabilities by proposing selected use cases.

2 TAGGING AND CONTEXTUAL

EXPERIENCE

2.1 Folksonomies Need Context

Folksonomies focus on user experience sharing

through tagging. Our interest in folksonomies lies in

their own definition quoted by Knerr (2006):

“Folksonomy is the result of personal free tagging of

information and objects (anything with a URL) for

one’s own retrieval. […] The act of tagging is done

by the person consuming the information”. As

recalled by Cernae (2008), in folksonomies,

metadata is not created by experts, but it is

spontaneously generated by consumers.

Folksonomies answer to the real need for semantic

descriptions that are closer to the knowledge domain

of the users (Gayo, 2010). Even if folksonomies are

nowadays broadly and successfully used, research

has to face several problems that inform about the

semantics and knowledge carried by tags.

The most emblematic representation of

folksonomies is the tags cloud that allows to

navigate through the whole set of tags that have been

created by the whole set of users. In parallel, a

resource selection shows the whole set of tags

created for it. Researchers have shown for long that

this sole information is too simple for users to take

best benefit from tags. Golder (2006) indicates that

« information tagged by others is only useful to the

extent that the users in question make sense of the

content in the same way ». For being totally useful,

it is necessary for the reader to be able to share the

tag semantics with its creator. Due to its polysemy,

the sole tag label is insufficient. Different solutions

like SCOT (Kim, 2008) or MOAT (Passant, 2008)

have then been proposed to link tags to definitions

issued from external ontologies like DBpedia or

Wordnet. This approach is interesting while

allowing different users to refer to a shared

definition of the same term. However, this solution

is also not sufficient since, even if they share the

same “formal” definition, people can use the same

tag for different purpose.

Aware of the fact that tags reflect personal view

of the world by individual users, researchers share

the idea that knowing who provided the tag can help

people in determining its relevance for their own

goals (Van Setten, 2006). This finding resulted in

many different works like MUTO (Lohman, 2011)

proposing ontologies for creating augmented tags. In

particular, tags are explicitly related to their creator.

This information offers a first essential link between

each tag and its creation context. However, Saab

goes even further. He indicates that “a single

individual can effortlessly switch their perspectives

based on their identity and create tags for the same

phenomenon based in different, sometimes

conflicting, identities” (Saab, 2010). The author

explains how a person being a hunter can tag a

weapon resource as “essential”, and the same person

being a father can tag the same resource as

“prohibited”. In his analysis, Saab demonstrates that

knowing a tag’s creator is indeed important, but not

sufficient. In order to let the experience carried by a

tag be fully understood (and useful), it is necessary

to let the reader know the context in which this

experience has been constructed, thus letting him

adopt the adequate perspective.

From another point of view, this need to better

link tags to their creation context can also be found

in the work of Shirky (2005). While analyzing

Del.icio.us, the author relates: “You can see there's a

tag to_read. A professional cataloguer would look at

this tag in horror -- This is context-dependent and

temporary”. As noticed by Golder (2006), this tag

can be considered as a Task Organizing tag in the

context of one of its creator’s specific task. In the

same idea, Kipp (2007) suggests that users may

relate tags to time or emotional reactions. Heckner

(2008) and Monnin (2010) advocate for associating

a tag with its creator’s intention. A tag like to_read

may be considered far from our interest in this

paper. It does not seem to carry some user’s

knowledge, but seems exclusively dedicated to a

private use. Some works like those proposed by

Knerr (2006) and Lohmann (2011) have defined

tagging ontologies that take care about tags

visibility. We also believe that it is important to let

users define private or public tags, and to let them

manage this visibility. Indeed, we can notice that

when the private to_read tag is made public, it

corresponds to a semantics change. The tag goes

from private organizational information to an actual

transmission of its creator’s experience. This same

tag then represents an advice intimately related to

it’s creator’s knowledge, and it can only be fully

understood in the context in which this advice is

provided. This is for example the case when a

teacher tags a particular resource for helping

students in the context of a specific course. The

course, its objective, the tagger’s role, the other

actors’ identities, other resources (that may thus be

considered as less important), and even linked

activities (related to the university) altogether

participate to the semantics carried by the tag.

2.2 Ontologies and Multi Viewpoint

The semantic web research domain also specifically

focuses on tag semantics. Ontologies offers meaning

to more or less formally weave the tags used to label

and categorize entities. Human and computer

systems can use such weaving to better understand

the meaning of entities. It is also possible to infer

new relations and perform complex semantic queries

thanks to ontological reasoners.

The general approach developed in the semantic

web tries to reach some consensus in order to

propose global and shared reference ontologies in

specific activity domains. Top-down approaches

involve domain experts and/or knowledge engineers

who develop the ontologies that will have to be

accepted by all the concerned knowledge workers

(Kotis, 2006). Unfortunately, literature shows that

reaching such a consensus is a real difficult task.

The main difficulty comes from the fact that

defining an ontology always corresponds to provide

a particular viewpoint about the domain’s entities.

Yet, the viewpoint of knowledge engineers is

usually not the same as the domain experts’ one.

And even experts in the same domain do not always

share the same viewpoint (Zhitomirsky-Geffet,

2017). As a result, different ontologies dedicated to

the same domain have been released. A large part of

nowadays research tries to define means that will

help to merge or to link these existing ontologies.

To palliate this problem, researchers like Dong

(2015) propose to adopt the inverse approach by

trying to learn structured knowledge from social

tagging data. Indeed, Garcia-Silva (2014) indicates

that emergent vocabularies turn folksonomies into

interesting knowledge sources from which

ontologies can be developed. Folksonomies are then

expected to capture all the viewpoints provided by

domain actors, and thus to facilitate the creation of

ontologies that would be a-priori accepted by them.

However, this task is also difficult since it needs to

extract the semantics of the folksonomy’s tags, thus

leading to the numerous context-related problems

we have underlined in the previous part.

Another approach tries to find equilibrium

between different expertise and contributions, and

different solutions have then been proposed for

allowing diverse actors to co-construct their

ontologies. This is for example the case in HCOME

(Kotis, 2006), and DILIGENT (Pinto, 2009). For

HCOME, Kotis notes that workers need to map

others’ conceptualizations to their own and put them

in the context of their own experiences. This can

result in new meanings since concepts are seen

under the light of new experiences. Pinto advocates

for letting people retain a part of the shared ontology

and modify it locally. Indeed, these two propositions

offer collaborative tools that allow users to

personalize a shared ontology, to adapt it to their

own experience, and then to integrate some of these

adaptations into the shared global ontology.

A new trend that can be represented by the

proposition made by Zhitomirsky-Geffet (2017) also

nourishes our own thinking about tag semantics. The

author remarks that most of the above frameworks

still force the users to reach a consensus on their

final ontology. She however argues that ontology

users are also interested in a variety of viewpoints on

the knowledge domain. Our understanding of the

state of the art and the above-mentioned issues also

lead us to think that there is a need for a new type of

ontology that allows multiple viewpoints on the

domain to co-exist. Like in HCOME and

DILIGENT, the main idea is to let users develop

their own personal ontologies reflecting their own

contextual experiences. But in our approach, these

ontologies are not intended to be finally merged in a

global one. They will co-exist, and will be closely

and explicitly linked to the description of the context

they describe and from which they have emerged.

2.3 Tagging in Activity’s Context

The above studies show that tag semantics takes

great advantages in being linked to its creation

context. The numerous improvement in researches

about folksonomies and ontologies have all

proposed new means to better take this context into

account while linking created entities to their

creators, or to their point of view. From our point of

view, the context that should be considered is

broader: it is synthesized in the concept of activity.

According to Ning (2012), the activity can be

used to glue knowledge item and knowledge context

such as people, resource and environment, in a

semantic way, thus providing enhanced knowledge.

While considering activity as a central concept for

contextualizing knowledge, tags can be explicitly

linked to their creators, to their creators’ viewpoints,

to their role in the activity, and also to the whole set

of correlated entities like the resources used, the

activity’s products, the other actors and roles, and

even other related activities. Moreover, each tag in

an activity is itself part of the knowledge context and

thus becomes the background of the other tags.

Such solution implies to know in which activity a

tag is created. This need is coherent to those

identified in the context of research about multi-

viewpoints ontologies that advocates for dealing

with the way people develop their

conceptualizations in the context of their day-to-day

activities, in a seamless way to their working

practices (Kotis, 2006). Following this direction, our

approach is to let users explicitly indicate the

activity related to the experience they are sharing.

Explicit reference to users’ activities has been

successfully used in several project dedicated to

information management. These propositions

criticize the omnipresent hierarchical structure used

to store and retrieve personal resources, and propose

activity-based tagging systems to solve the identified

problems (Voida, 2009). Even if the main purpose of

these systems is different from ours, their results are

instructive. In particular, Oleksik (2009) categorized

several benefits from using the activity concept in

the tagging of resources. Kersten (2012) has shown

that the activity concept matches well end-users’

representation of their work while fitting with real

world organizations, and facilitating the

management of their resources while often switching

from one task to another. These findings advocate in

favor of an environment emphasizing the activity

concept for contextual tagging.

2.4 Main Objectives

We aim at providing a tagging system that allows

users to share their experience about web resources

involved in specific activities. According to the

above-mentioned literature, our approach is to allow

users to tag resources while keeping explicit links

between these tags and the activities in which the

resources were used, and where the tags fully make

sense. The main purpose is to help users in sharing

their viewpoints about sets of entities involved in

specific contexts, and to let others discover, browse,

understand and be inspired from them. Such

approach can somehow be qualified as folksonomic.

It allows users to browse the whole set of shared

viewpoints and to create clouds in which tags are

pondered by their occurrences. It is however also

possible to know each specific contextual activity

related to a user tagging action. And finally, one can

filter and search the viewpoints according to criteria

built from the system structure.

We want to take benefits from ontological tools

to have access to linked data, semantic queries and

inference. It however has to be noticed that even if

they inspired us, our aims differ from those of

Zhitomirsky-Geffet (2017) in building ontologies.

The approach of the author is to guide the users to

construct multiple viewpoint ontologies and then

integrate them together, where multiple viewpoints

are part of a central unified ontology. We do not aim

at competing with engineering methods and systems

for building centralized ontologies. We aim at using

ontological mechanisms for supporting users in

sharing their own personal experience. According to

Zhitomirsky-Geffet (2008), we think that the

viewpoints are not limited to different visualizations

of the same information relationships: the

relationships themselves may differ as well. In our

approach, each viewpoint corresponds to a

(potentially) different set of semantic entities and

relations between them. We thus define each

viewpoint as a single contextual ontology, a term

that can be correlated to the personal knowledge

ontology proposed by Hsieh (2008) for personal

information management. A contextual ontology is

intrinsically linked to a specific context and

represents the experience developed by an actor in a

specific activity. However, according to the real-

world activities they represent, contextual ontologies

can be interrelated. Thus, the acceptance and pooling

of some contextual ontologies’ concepts developed

by different actors in some shared, correlated, same

type, or same domain activities can be envisioned.

Such acceptance by conscious integration of some

others’ concepts into one’s contextual ontologies

may help to dynamically represent the consensual

viewpoint of a group of actors, and picture some

shared and evolving contextual group ontologies.

3 THE EVOXEL FRAMEWORK

3.1 Using PROV to Model Context

We have chosen the Provenance data model (PROV)

and especially the PROV Ontology (PROV-O) to

represent the activities, the resources they use and

produce, and their actors. PROV-O is an owl

ontology providing a simple data model proposed by

the W3C for which provenance is defined as “a

record that describes the people, institutions, entities,

and activities involved in producing, influencing, or

delivering a piece of data or a thing” (W3C, 2013).

One of the main ideas from which PROV has

emerged is that provenance of information is crucial

in deciding whether information is to be trusted and

how to give credit to its originators when reusing

it. Such intention clearly matches our needs focusing

on the provenance of the tags and the shared

experience they are related to.

The PROV data model is constituted of three

core concepts: Agents participate in Activities,

which can use and produce Entities (such as

documents, web sites, etc). We thus use prov:Agent

to describe our users sharing their experience,

prov:Activity to describe the contexts in which this

experience emerges, and prov:Entity to describe the

resources used as well as the activities’ products. We

also mainly use PROV’s basic properties (prov:used,

prov:wasAssociatedWith, etc.) to describe the

relationships between these entities. PROV classes

and properties are further detailed in (W3C, 2013).

3.2 Putting Tags in the Context

Using PROV thus helps us in describing activities,

i.e. the contexts related our users’ shared experience.

As stated in the previous section, each particular

context is described in a corresponding contextual

ontology.

Figure 1: Fragment of Greg’s contextual ontology

describing his point of view on the activity.

Figure 1 describes some fragments of Greg’s

contextual ontology corresponding to its Writing

KMIS article activity. KMIS article is a prov:Entity

representing a product of this activity. Zhitomirsky

et al. 2017 and Saab, 2010 represent two of the

resources used as references in the current article.

The entity Zhitomirsky et al. 2017 is tagged journal

article, and also scientific publication. In fact, Greg

only applied the journal article tag, and scientific

publication is inferred by the system. This is due to

the fact that these tags correspond to owl classes,

and that the former is defined has a subclass of the

latter by Greg’s in its tags hierarchy for this

contextual ontology. Yet, the framework’s structure

allows using all classical semantic reasoners’

capabilities in the context of each specific activity,

thus providing contextual semantic reasoning.

Greg also tagged this resource as related work in

the current article. It is obviously a contextual tag

which sense is closely related to this specific

context. Indeed, in another article, the same resource

can be tagged differently (e.g. “case study”) because

it has been quoted for other reasons (e.g. basing new

research on the provided study about the effect of

diet on health). This tag is then applied to this entity

only in this contextual ontology.

Figure 2: Contextual Ontology Structure.

Figure 2 represents the generic structure of

contextual ontologies. It shows that each user’s

contextual ontology is also itself considered as a

prov:Entity. Each user’s experience in a particular

activity is crystallized in a contextual ontology,

which is itself considered as a product of the user’s

activity it describes.

3.3 Connecting Personal Ontologies

The framework generates one contextual ontology

by actor and by activity, each ontology being split in

two parts: a private and a public one. Some activities

can however be shared by multiple actors. For

example, Writing KMIS article actually involves two

co-authors. Such activity shared by two actors is

described in two contextual ontologies, each one

containing PROV data and tags associated to the

elements referenced by each user in this activity.

Some entities can be involved in multiple

contextual ontologies. For example, the two authors

realizing the Writing KMIS article activity both

share the same references and thus, these resources

are part of their respective contextual ontologies.

They also may share some tags applied to these

resources, but not necessary all of them if they

consider these resources from different viewpoints.

Moreover, the same tags may be part of different

(inheritance) hierarchies representing different

meaning for their creators.

From a global viewpoint, we use the

dcterms:references and dcterms:isReferencedBy

relationships to keep track of the bilateral links

between the many elements (identified by their IRI)

like users, activities, resources, products and tags,

and the contextual ontologies referencing them.

3.4 Framework Capabilities

The framework allows browsing all the contextual

ontologies in which a particular entity is involved.

Starting from an entity (e.g. a resource) and

selecting a contextual ontology using it lets discover

the other entities that participate to the realization of

the targeted user’s activity, and their specific

relationships. This is achieved while providing the

associated contextual tags, thus letting one better

understand the related experience. The framework

also gives the opportunity to list all the tags tied to

an entity by “merging” the several users’ viewpoints

in a folksonomic fashion. Selecting a particular tag

can lead to all the activities in which it has been

used, and thus let discover its meaning in the

different users’ activities. It is also possible to search

the environment by performing queries based on the

many entities, entities types (tags) and their specific

relationships in contextual ontologies.

To put these functionalities in action, we are

developing a web application and Google Chrome

plugins connected to a first implementation of the

framework through a JEE server using the OWL

API and the Openllet semantic reasoner. We tested

some scenarios enhancing some existing web tools.

For example, tools like Google Scholar can help in

looking for specific articles, searching with

keywords or author name. Given an article, it shows

the list of other papers referencing it. A same paper

could be cited for several reasons, depending on the

context. Discovering (or remembering) why a paper

references another one can help in faster

understanding a researcher’s viewpoint, and then

evaluate the (contextual) level of interest of other’s

research work. Thanks to our framework, a plugin

injects these new data in a references list, and

proposes the contextual and potentially inferred tags

associated with each reference in this context of use.

We investigated the domain of scrapbooking,

where many people use Blogger to exhibit their

creations and design methods. Each blog article

gives details about a creation like the materials used

and how they have been associated together. Thanks

to EVOXEL, the blogger is able to contextually tag

materials. Afterwards, exploring the blog lets users

discover that a certain piece of paper, which has

been tagged as background in a creation, has also

interestingly been used as an embellishment while

being associated with other specific materials (or

materials types) in another one. Such discovery,

letting one discover unexpected use and/or

unexpected (associated) resources, becomes a source

of new inspiration.

Instructional designers also used the framework

to describe their pedagogical activities in an

approach somehow similar to Merlot. They were

afterwards able to use web forms applying semantic

query features and for example search in others’

contextual ontologies for all the lectures using

SCRUM in an active pedagogy process.

More generally, the framework also allows the

creation of tags from query results. For example, a

query for retrieving all the scientific articles that cite

a particular researcher’s papers as related work can

be transformed in a new (private) to read tag (i.e. an

equivalent class). The inference will then apply the

tagging on the corresponding elements, even if they

are integrated well after the tag’s definition.

4 CONCLUSIONS

The web nowadays offers many tools that allow

people to share their own experience about the

resources they use. In this context, tagging systems

can help in organizing and navigating through the

shared knowledge. Moreover, tags carry semantics

that can help in understanding it. Building on a

literature review, we have shown that tags could take

great benefits from being even more closely related

to their creation context. We thus have proposed the

EVOXEL framework that emphasizes the concept of

activity for contextual tagging. The framework relies

on semantic web technologies and introduces the

concept of contextual ontology to link tags to their

creation contexts, and let them benefit from

ontological structures, mechanisms and tools.

We have shown that the framework yet offers the

basics for supporting our needs, but EVOXEL is still

under development. We are currently further

developing the web applications and the Google

Chrome plugins offering its functionalities at our

end-users’ abstraction level. The framework is also

itself currently enhanced by integrating a

collaborative dimension for better supporting users

who share activities and/or integrate and reuse

entities borrowed from other contextual ontologies.

EVOXEL’s model also offers new opportunities and

we are already working with other researchers on

new functionalities based on contextual ontologies

similarity measurement. We expect those new

features to even better support end-users’ experience

sharing, discovery, and inspiration.

REFERENCES

Cernea D.A., Del Moral E., Gayo J.E.L. (2008). SOAF:

Semantic indexing system based on collaborative

tagging, Interdisciplinary Journal of E-Learning and

Learning Objects 4, 137-149.

Dong H., Wang W., Liang H.N. (2015). Learning

Structured Knowledge from Social Tagging Data: A

Critical Review of Methods and Techniques, Proc. of

8th SocialCom, Chengdu, China.

Garcia-Silva A., Garcia-Castro L. J., A.García, Corcho O.

(2014). Social tags and Linked Data for ontology

development: A Case Study in the Financial Domain,

Proc. of WIMS14, Thessaloniki, Greece.

Gayo, J.E.L., De Pablos, P.O., & Lovelle, J.M.C. (2010).

WESONet: Applying semantic web technologies and

collaborative tagging to multimedia web information

systems. Computers in Human Behavior, 26(2), 205-

209.

Golder S., & Huberman B.A. (2006). The structure of

collaborative tagging systems, Journal of Information

Science, 32(2), pp. 198–208.

Heckner, M., Mühlbacher, S., & Wolff, C. (2008).

Tagging tagging. Analysing user keywords in

scientific bibliography management systems. Journal

of digital information, 9(2).

Hsieh J-L., Lin I. W., Sun C. T. (2008). A Web-based

Tagging Tool for Organizing Personal Documents on

PCs. Proc. of CHI 2008, April 5-10, Florence, Italy.

Kersten, M., Murphy, G. C. (2012). Task context for

knowledge workers. In Proceedings of AAAI 2012

Activity Context Representation workshop.

Kim H.L., Breslin J.G., Yang S.-K., Kim H.G. (2008).

Social semantic cloud of tag: Semantic model for

social tagging. In Proc. of KES AMSTA’08, pp. 83–92.

Kipp, M. I. (2007). @toread and cool: Tagging for

time, task and emotion. Proceedings 8th Information

Architecture Summit. Las Vegas, Nevada, USA.

Knerr, T. (2006). Tagging ontology-towards a common

ontology for folksonomies.

Kotis K., Vouros G. A. (2006). Human-centered ontology

engineering: The HCOME methodology, Knowl. Inf.

Syst. 10(1), pp. 109–131.

Lohmann, S., Díaz, P., & Aedo, I. (2011). MUTO: the

modular unified tagging ontology. In Proceedings of

the 7th Int. Conf. on Semantic Systems, pp. 95-104.

Monnin A., Limpens F., Gandon F., Laniado D. (2010).

Speech acts meet tagging: NiceTag ontology. In Proc.

of I-SEMANTICS’10, pp. 31:1–31:10. ACM.

Ning, K., & O'Sullivan, D. (2012). Context modeling and

measuring for context aware Knowledge Manage-

ment. International Journal of Machine Learning and

Computing, 2(3), 287.

Oleksik G., Wilson M. L., Tashman C., Rodrigues E. M.,

Kazai G., Smyth G., Milic-Frayling N., Jones R.,

(2009). Lightweight tagging expands information and

activity management practices. Proc. of the SIGCHI

CHI’09, April 04-09, Boston, MA, USA.

Passant A., Laublet P. (2008). Meaning of a tag: A

collaborative approach to bridge the gap between

tagging and linked data. In Proc. of LDOW’08.

Pinto H.S, Tempich C., Staab S. (2009). Ontology

engineering and evolution in a distributed world using

DILIGENT, Handbook on Ontologies, pp 153-176.

Saab D. J. (2010). The Ontology of Tags. iConference

2010.

Singer, L., Figueira Filho, F., Cleary, B., Treude, C.,

Storey, M. A., & Schneider, K. (2013). Mutual

assessment in the social programmer ecosystem: An

empirical investigation of developer profile

aggregators. In Proc. of CSCW 2013, pp. 103-116.

Shirky C. (2005). Ontology is overrated: Categories, links,

and tags. http://www.shirky.com/writings/ontology_

overrated. html.

Van Setten M, Brussee R, Van Vliet H, Gazendam L, Van

Houten Y, Veenstra M (2006). On the importance of

who tagged what. In Proc of the Workshop on the

Social Navigation and Community based Adaptation

Technologies, Dublin, Ireland, June 20, pp 552-561.

Voida, S. & Mynatt, E.D. (2009). “It feels better than

filing”: Everyday work experiences in an activity-

based computing system. Proc. of SIGCHI CHI '09.

W3C (2013). Prov-dm: The prov data model. W3C

Recommendation, World Wide Web Consortium.

Zhitomirsky-Geffet M., Bar-Ilan J., Miller Y., Shoham S.

(2008). A Generic Framework for Collaborative

Multi-perspective Ontology Acquisition, Proceedings

of WWW 2008, Beijing, China.

Zhitomirsky-Geffet M., Erez E.S., Judit B-I. (2017).

Toward multiviewpoint ontology construction by

collaboration of non-experts and crowdsourcing: The

case of the effect of diet on health. J. Assoc. Inf. Sci.

Technol. 68(3), 681-694.