Finalreporteurosis3 1

EUROSIS
Task 3.1
« Webmapping of science and society actors in Europe »
Final Report
http://webatlas.fr/exhibition/eurosis/
1
EXECUTIVE SUMMARY.................................................................................................................................................3
INTRODUCTION..............................................................................................................................................................5
MAPPING THE WEB OF SIS IN 12 EUROPEAN COUNTRIES...............................................................................6

WHAT IS OUR APPROACH?............................................................................................................................................6
Web mapping as a Sociology..........................................................................................................................6
New tools and methodology for a new data source: the web.....................................................................................7
Web mapping as Network Sciences............................................................................................................................8
CONSTRAINTS.................................................................................................................................................................9
Differences between NCPs.........................................................................................................................................9
It is only a “Web picture”.........................................................................................................................................10
The set of countries...................................................................................................................................................10
ABOUT THIS EXPERIMENT............................................................................................................................................11
NCPs practicing the Web exploration.......................................................................................................................11
How works a distributed network of experts?..........................................................................................................12
Building a common description of the resources.....................................................................................................13
RESULTS..........................................................................................................................................................................16
GENERAL PICTURE........................................................................................................................................................16
Summary...............................................................................................................................................................16
General Map.........................................................................................................................................................17
Connectivity structure of the types of actors.........................................................................................17
Interactions between types of actors........................................................................................................20
NATIONAL MAPS........................................................................................................................................................... 21
Specificities...............................................................................................................................................................21
Armenia....................................................................................................................................................................23
Belgium.....................................................................................................................................................................26
Bulgaria....................................................................................................................................................................30
Czech Republic.........................................................................................................................................................33
Estonia......................................................................................................................................................................37
Finland.....................................................................................................................................................................41
France.......................................................................................................................................................................44
Hungary....................................................................................................................................................................47
Italy...........................................................................................................................................................................50
Montenegro...............................................................................................................................................................53
Poland.......................................................................................................................................................................56
Portugal....................................................................................................................................................................59
MAIN AXIS OF INTERPRETATION.................................................................................................................................61
Hierarchy emerging from countries' relations..........................................................................................................61
 Where do websites aggregate?......................................................................................................................61
 Small-World in the countries......................................................................................................................................61
 The signature of complexity.......................................................................................................................................62
 Observing the hierarchy.............................................................................................................................................63
 Interpretation of the hierarchical relations..................................................................................................................64
FUTURE STEPS..............................................................................................................................................................69
PROPOSITIONS FOR A WEB INFORMATION SYSTEM DEDICATED TO SIS (OR SIS²).........................................69
BIBLIOGRAPHY............................................................................................................................................................70
GLOSSARY......................................................................................................................................................................71
ANNEX 1...........................................................................................................................................................................73
EXECUTIVE SUMMARY............................................................................................................................................3
INTRODUCTION..............................................................................................................................................................4
2
MAPPING THE WEB OF SIS IN 12 EUROPEAN COUNTRIES...............................................................................5
WHAT IS OUR APPROACH?............................................................................................................................................5
Web mapping as a Sociology...........................................................................................................................5
New tools and methodology for a new data source: the web.....................................................................................6
Web mapping as Network Sciences............................................................................................................................7
CONSTRAINTS.................................................................................................................................................................8
Differences between NCPs.........................................................................................................................................8
It is only a “Web picture”............................................................................................................................................9
The set of countries.....................................................................................................................................................9
ABOUT THIS EXPERIMENT............................................................................................................................................10
NCPs practicing the Web exploration.......................................................................................................................10
How works a distributed network of experts?..........................................................................................................11
Building a common description of the resources......................................................................................................12
RESULTS..........................................................................................................................................................................15
GENERAL PICTURE........................................................................................................................................................15
Summary...............................................................................................................................................................15
General Map.........................................................................................................................................................16
Connectivity structure of the types of actors..........................................................................................16
Interactions between types of actors.........................................................................................................19
NATIONAL MAPS........................................................................................................................................................... 20
Specificities...............................................................................................................................................................20
Armenia....................................................................................................................................................................22
Belgium.....................................................................................................................................................................25
Bulgaria.....................................................................................................................................................................29
Czech Republic.........................................................................................................................................................32
Estonia......................................................................................................................................................................36
Finland......................................................................................................................................................................40
France........................................................................................................................................................................43
Hungary....................................................................................................................................................................46
Italy...........................................................................................................................................................................49
Montenegro...............................................................................................................................................................52
Poland.......................................................................................................................................................................55
Portugal.....................................................................................................................................................................58
MAIN AXIS OF INTERPRETATION.................................................................................................................................60
Hierarchy emerging from countries' relations..........................................................................................................60
 Where do websites aggregate?......................................................................................................................60
 Small-World in the countries......................................................................................................................................60
 The signature of complexity.......................................................................................................................................61
 Observing the hierarchy.............................................................................................................................................62
 Interpretation of the hierarchical relations..................................................................................................................63
FUTURE STEPS..............................................................................................................................................................68
PROPOSITIONS FOR A WEB INFORMATION SYSTEM DEDICATED TO SIS (OR SIS²).........................................68
BIBLIOGRAPHY............................................................................................................................................................69
ANNEX 1......................................................................................................................................................................70
EXECUTIVE SUMMARY
Background and objectives
In 2007, the conjunction of two favourable elements made the EUROSIS Webmapping project
possible:
3
- a promising national study had just been carried in France on a related topic: it consisted in a
webmapping of Science in Society stakeholders in France (SiSmap);
- the publication of a call for proposal dedicated to the “Science in Society” National Contact
Points (SiS NCPs) network. While the NCPs built up an answer to this call, it appeared that
there was a need for them to get a better knowledge of science and society communities.
The main objective of the project was to make it possible for SiS NCPs to acquire a better
knowledge of science and society potential projects’ partners. Based on a specific webmapping
of contact areas between actors of the science and society debate, the project aimed at
providing NCPs with an adequate tool, in order them to visualize the patterns of the science
and society relationship, and to understand which stakeholders, communities, grassroots
protagonists, are the most represented and the most active in the science and society
dialogue.
Methodology and results
In order to reach these objectives, the project partners selected a subcontractor, namely the
Web Atlas association. Together, they developed a back and forth process:
 NCPs would provide various data on national SiS stakeholders

 Web Atlas processed the collected data and correlated them so that findings take the shape
of graphic interactions corresponding to various SiS domains
 NCPs analysed and adjusted the results according to an analysis grid
 Web Atlas delivered a SiS’actors study.
While defining the methodology, a series of possible biases appeared:

 The first obvious one, which is our choice, is that this study is based on the Web only. As a
consequence, it does not show the “real” SiS components of each Member State, but the
dynamic interactions between grassroot protagonists on the Web
 For each country, the task of collecting data has been made by the NCPs. However, NCPs
have different backgrounds and knowledge of Science and Society communities. As a
consequence, their better knowledge of specific SiS networks may have created
heterogeneity and impacted the final mapping
 All participating partners were volunteers. As a result, the project did not benefit from a
very coherent set of countries. This had a noticeable impact on the results.
The partners agreed on having the map address, at the national as well as European level:
 Major themes (scientific and/or citizen) of the science and society relationships
 Actors and communities engaged in these interactions
 All categories of interactions, going from dialogue and cooperation to controversies and
open conflicts.
The results the project are of two kinds:

 A series of deliverables: a report on the Science in Society actors in Europe, a list of Science
and society stakeholders (NGOs, government, research centers...) with contacts details and
158 “printed” maps
 Scientific results. A comparative study of the national studies led by each SiS NCP in its own
country shows strong national specificities, although general trends can be identified: a
three-level hierarchical structure, universities and the world of academic research acting as
a bridgehead or as a backbone in the overall architecture. This suggests a pregnant top-
down system; our main recommendation would then be to facilitate more fluid two-way
exchanges between the various layers of the system, in order to encourage bottom-up
participation.
4
5
Introduction
Introduction
After a call for interest, organized by the French Science in Society National Contact Point (SiS
NCP) as task leader of the Webmapping project, 12 NCPs chose to take part in the task
“Webmapping Science in society actors” These NCPs are representatives of Armenia, Belgium,
Bulgaria, Czech republic, Estonia, Finland, France, Hungary, Italy, Montenegro, Poland,
Portugal (for the list of participants and their individual profile, see annex 1). These experts in
the Science in society programme (7th framework programme) were actively involved all along
the different stages of the task:
 They provided data (websites, tags...)
 They took part in two two-day Workshops in Paris (26 and 27 May, 2008 and 1 and 2
December, 2008)
 They validated the final deliverables.
The first aim of the EuroSiS Webmapping project is to explore and understand the large, open
and dynamic topics dedicated to science in society on the Web through the collective
construction of a mapping tool. The results consist in a collection of SiS-maps including themes
and actors at different levels of analysis (common European themes, international and national
neighbourhoods SIS themes, particular themes through national and European dimensions…).
It consists also in the elaboration of a common description schemes and a collective
vocabulary for resources description among 12 NCPs involved.
As regards to methods and tools, the EUROSIS mapping project is based on :
a) Use of visualization for communication/promotion of EuroSiS project
b) Production of graphical synthesis of large sets of data
c) Extraction of visual patterns from web data analysis (centre, periphery, clusters, levels of
hierarchy, borders…)
d) With different levels of data (zoom) and analysis of SIS European themes.
e) And preparing an information archiving process on specific topics and, beyond, a possible
Web Information System (or portal) dedicated to European SIS domains.
6
Mapping the web of SiS in 12 European countries
What is our approach?
Web mapping as a Sociology

One of our fundamentals is the “Actor-Network Theory” or ANT. Despite this name it is not a
theory, but rather a methodology. Its main purpose is: ”just observe”, and tell what you see.
And its difficulty takes place in this “just”, because as the French sociologist Bruno Latour
says, it is not easy to “just” observe. We do not judge what is good or bad, who is right or
wrong. We do not try to show something. We just observe what is happening between the
actors of Science in Society.
The ANT makes us aware of one of the biggest issues in this kind of sociology: it is very hard
to define what – or who – is an actor. At one moment one institution is an actor, but at
another moment it is a network of people acting on their own. The actors involved in Science
and Society split or gather depending on various reasons, including the issue they are
confronted with. A laboratory may have a specific action in a specific field (ex: promoting
nanotechnologies) and some of its researchers may act in another way (ex: alerting people
about risks).
This is the first complication: how to describe a large amount of entities, when describing only
one entity is so difficult?
Let us say we have got our entities – many entities. How are they linked together? How to
define the interactions – many interactions – between these entities? Let's say one actor is
linked to another. Do they work together? Or does one actor give money to the other? Do they
share common objectives? Or just know each other? Is this link positive (they cooperate)
negative (they compete) or both (“coopetition”)?
This is our second complication: how to describe a large amount of interactions, when
describing only one interaction is so difficult?
Science already provides us with very relevant studies that deal with those complicated issues.
But reality is not only complicated. It is also complex – and it is not the same thing. Even
simple things can draw complex networks. We all know that bees are much simpler than the
human being, but we do not succeed in understanding how the bee society acts, as a whole
entity, as an actor-network. And it is not because bees are complicated: if we work enough we
will understand any complicated thing. It is because their society is complex, and this means
that it is out of our understanding – unless we have some help from adapted tools like
computers. No matter how much time you spend to understand one only bee. It is in the
complex network of their interactions that hides the answer.
How to explore and describe the geography of an open, dynamic and large scale hypertext
system such as the Web? Can we identify and map such a distributed organization? A web
mapping project, even with social network analysis, lead us to a very new kind of information
architecture as A.-L. Barabasi described it in “Linked” or D. Watts in “Six Degrees”. With
EURO-SIS data, we are far from a library…
7
A lot of studies consider one or few types of actors, or one or few types of interactions. We
know some things by studying an issue (ex: gender studies). We know some things by
studying some type of actors (ex: media and agenda setting) or interactions (ex: the business
model of universities). But it is not enough to get the big picture, because the way these
studies link together is too complex to be easily understood. As a result the context is missing,
so that it is even more difficult to get a perspective. The different actors and issues are not
separated like suggest these studies, there are a lot of transversal links, from one actor
involved in a field to another actor of another type in another field... And these problematic
interactions are not always complicated to explain. What is the link between a politician
involved in health issues in southern countries, and a scientific journalist involved in gender
studies? They are married: a simple explanation but that does not fit in a simple grid of
analyze. A lot of interactions between actors are due to specific cases, and we used to consider
these unexpected links as something irrelevant. But this is an important part of the big picture
we fulfill. This is a part of the complexity we have to deal with, and we must adopt a new
approach to observe it.
New tools and methodology for a new data source: the web
The idea of getting the big picture of complexity is not a new one. But something changed that
makes it possible: Internet. We use the web as a field to collect a large amount of
heterogeneous data. In the physical world it would be difficult (and very expensive) to collect
data about many actors and their interactions. The fact is they belong to various spheres, and
each of these spheres has its own rules. You cannot ask for information to a university the
same way you ask it to a ministry, or a NGO, or a newspaper... And if you usually know quite
easily which are the interactions inside one sphere (governmental, science, media...) it is far
more difficult to get information about “transversal” interactions, from one sphere to another.
The web makes everyone projected to the same surface. A website is a website and all of them
can be compared no matter which sphere they belong to. A hypertext link is a hypertext link.
You probably will not know which “real” interaction has made this link possible, but it exists.
This is a problem for detailed analysis, but here is the gain: all links fit to the big picture.
Remember that the big picture is the context that we need. Once we get it, we can “dive” in
data to get details.
Bruno Latour calls this “quali-quantitative analysis”. It is a classical methodology in network
sciences. Our data are as large as they are detailed. We collect many websites but each of
them contains a lot of useful information, including: who is this actor, where does it come
from, who are its members and/or partners, what does this actor do... You can get some
detailed analysis on some actors, but not all of them (because it is too much work). Where is it
useful to get a detailed analysis? When you interpret a map, you may know where to “dig” or
“dive” to the detailed level. The map is the context that gives details their meaning, and the
details help you to get a better interpretation of the general level. That is why we always move
our point of view from general to detailed, from detailed to general. It is neither a qualitative
nor a quantitative approach. It is both of them, because the web is a new source of knowledge
that makes it possible.
The quali-quantitative approach is the best methodology to understand the complexities we
are confronted with. That is why when our first task is to collect, the second one is to expose
data. Our goal is not to provide a statistical analysis, even if we actually use a lot of statistics.
Our goal is to show the big picture, whatever its complexity is. Each map we build, each
statistic we compute, is linked with the others so that you can switch your point of view from
the general to the detailed and so on. We will provide some conclusions, but these data tell a
lot more than we can possibly interpret in this report. The value of this kind of work is to
expose many things that you can understand if you get involved in the field of Science in
Society.
That is why our tools are not “big black boxes” that compute some list of numbers. They are
made to show what is happening, and we altogether discuss the best way to show the data we
have collected.
The main work that each NCP provided is collecting and discussing the relevance of the
8
maps and the other statistical visualizations.
Web mapping as Network Sciences

A few years ago in the United States, it was decided to create a new discipline: the Network
Sciences. The scientific work done by many researchers like A.-L. Barabasi, from whom this
work is inspired, did not fit into another existing discipline. These researches started as
mathematicians (like Duncan Watts) discovered a new kind of structure called “scale-free
network”. This kind of structure can be considered as the signature of the complexity: it is a
non-hierarchical network that we find in various fields like biology, ethology, sociology,
economics, engineering... and the Web. These networks are special because they have a high
degree of clustering (ex. The friends of my friends are probably my friends) but also a small
diameter (the famous “six degrees of separation”).
We can tune each network between complete order and complete disorder (central
feature on the left). As expected on the Web, EuroSiS map belongs to a special class of
networks with two properties : composed with clusters and short distances between nodes
(on the right, general EuroSiS map at 2nd workshop).
A lot of strange properties stick to these structures (strongly connected components, power
law distribution of degrees, the preferential attachment modeled by Barabasi...). We do not
know so much about these strange objects but we can build tools to analyze them. The
mathematical concept is the “graph” and we use these graphs to build our maps: sites
connected by links. The “InfoViz” discipline helps us to build tools to manage graphs,
altogether with computer sciences. And with these tools we design maps following the basics
of semiotics (J. Bertin). As previously stated, the whole methodology can be understood in the
perspective of the sociology and the Actor-Network Theory. All these disciplines are necessary
to handle complex structures. That is why we consider this approach in the perspective of the
interdisciplinary field of Network Sciences.
9
The basis of network sciences is the graph theory, which is a branch of the Mathematics that
tries to understand different aspects of the graphs: nodes connected by edges. At the
beginning this field is close to geometry and topology, and is quite abstract (it applies to any
“relational system”). European mathematicians Erdös and Renyi tried to explain why, when
you make a random graph grow, a “strongly connected component” appears no matter which
are the parameters. But these works did not help to understand complex networks until Watts
and Strogatz understood that the “random graph” was not the good model. At this moment
they worked about how animals can synchronize, and more precisely some crickets that emit a
pulsating light. Their problem was: how do these insects observe each other, so that they
perfectly synchronize through large distances? They tried to draw the network of the insects
observing each other to understand how it was possible. The solution was found in a new type
of networks they imagined, that is neither a regular nor a random graph, but something in-
between. This graph has some shortcuts that break the symmetry and allow nodes at the
opposite to be connected. These crickets were looking at their neighbors, excepted some of
them looking far away to distant other crickets (shortcuts). And they noticed that this new
model explained some animal phenomenons, but also some social ones like the famous “six
degrees of separation” that connect almost everyone on earth (identified by Milgram); that is
why this type of networks was called “small-world”.
After this discovery different works established that these “small-world” networks could be
found in various places, like the power grid of the USA, neural networks, ecosystems and of
course, Internet. The Web had to be studied, and various “network sciences” works were
dedicated on it. Kleinberg's works on graphs were decisive there. He created his famous
algorithm HITS that allows to build search engines, and he worked on the existence of
“aggregates”, strongly connected clusters of resources that share the same topic. Barabasi is
also a “father” of network sciences, as he tries to determine the diameter of the Web and
conceptualized the “preferential attachment” as the source of the small-world networks (which
he calls “scale-free networks”). His book “Linked” is a good synthesis of all these fields, and it
has been read by a lot of people including many non-specialists. Tim Berners Lee, who created
the HTML and the Web, leads a network-oriented program called “Web Science”. On his side
Barabasi leads a program called “Network Sciences” at the Sante Fe Institute, that interests in
the Web but also other applicative fields like cure for cancer or studies on biodiversity.
Nowadays Web studies are only one branch of Network Sciences amongst other. Its concepts
spread to various fields, now including Science in Society.
Constraints
Differences between NCPs

One of our goals is to compare the countries. But for each country the task of collecting data
has been made by the NCPs. That is why the differences observed between countries could be
due to differences between the NCPs. Indeed, some of the NCPs were specialist of the Science
in Society fields whereas others were specialist of the FP7 procedures (for the NCPs profile, see
annex 1). As a consequence, they all had a different approach of the way of collecting data.
This is an internal problem of our methodology, we could not avoid it. In a matter of facts, we
observe such a bias. We observe it because we are aware of this issue since the beginning. We
10
deal with it.
Our way to eliminate the main part of this bias is to actually compare the profiles of the
countries, and to confront it to the NCPs. The NCPs are our experts and we count on them as
objective analysts. In our workshops we discussed about these differences, and asked them for
an interpretation of the differences between the data of each country. The self-evaluation
came spontaneously and we had no difficulty to point to these biases.
As we will see, we are not able to provide a general analysis grid for all countries. This is one
of our results: there are big differences between the way that each country “thinks” Science in
Society. So the bias due to the differences between NCPs still remains in our analysis .
But we will point to the biased results as far as we know them. We will deal with it to
highlight the main real differences between countries' profiles. On a detailed level it is
difficult to separate biases from true results. But most of our observations are so clear and
general that it is not the case.
It is only a “Web picture”

We studied the Web. Not the “reality”. But our results are besides objective, and as real as
they are scientific. Now let's clarify this point.
The Web is real. Naturally it is not the “physical world”, but we do not consider it as “virtual”.
It is not a “virtual reality”, it is rather a “real virtuality”. We observe Websites of actors that
really exist and these actors really published the information that we collect. These Websites
are not “the actors” but we consider that they represent them in a certain way. It is useful to
study the Web and that is why we do it, but we do not take an actor for its website.
We also consider the hypertext links as the only relation between actors that we observe. We
do not know how to describe or give a value to such a link. It is neither good or bad, strong or
weak... We just take it as the trace of an interaction, whatever it is. Because we do not need
to analyze it at a detailed level, it is sufficient. But we can track for more details by “diving”
into data and analyzing the context of the link (analyzing the page: what does it tell, what is
the meaning of the link...).
The Web does not exist in a separated reality. What exist in the Web has good reasons to
exist. The Web is produced by the “physical reality”. The Web is a distorted prism of the
“reality”. At a general, non-detailed level, it is a useful representation of what “really”
happens, despite the distortions (that we always try to manage). And it is not only that: it is
the only way that we have to get the whole picture.
The set of countries

We asked for volunteers to participate to this task. As a result we did not get a very coherent
set of countries. This has a noticeable impact on our maps.
The first problem is linked to the size of the countries. Large countries propose more relevant
actors (and websites) than small countries. How to balance the amount of websites
demanded? We chose to ask for the same amount of actors to each NCP. This means that
small countries have more small actors: big countries selected only the “top” actors, those who
often have the biggest websites.
This produces a map where each country has the same amount of actors, but their
connectivity strongly varies. That is why the mean connectivity of each country has to be
balanced with the size of the country: some smaller actors do not appear in the set of the
biggest countries.
The second problem is about the countries that did not participate. This impacts the
connectivity of smaller countries. We think that most of the big countries have a lot of
interactions with various other countries. But smaller countries may have a large part of their
interactions with few big countries close to them. As you will see Estonia has a lot of links with
Finland. But some of our small countries like Montenegro look poorly connected because we do
not have their main partners (Greece, Germany, Austria...). Due to the absence of several
big countries, some small countries may look less connected than they are.
11
About this experiment
NCPs practicing the Web exploration

NCPs are not supposed to be experts in watching, even if they are experts in the SiS field.
Exploring the Web is a difficult task. It needs a lot of energy and beginners often feel lost in
the huge amount of pages. Our goal is to help them to find their way in this ocean of
information.
Layers of the web:

- Higher layer: the most visible
(Google, Microsoft, Adobe...)
- Medium layer: websites

aggregate here (on-line
communities...)
- Deep web: databases...
Our theoretical model tells that the Web is organized in layers (see image above). The higher
layers contain the biggest websites, associated to the most notorious actors. The more a
website is in a higher layer, the more it is connected and the more it is general. That is what
we usually call the highways of information. But low layers also contain a lot of relevant
information. You may find there the websites of many actors that you do not know. These are
more specialized websites, corresponding to various topical sub-categories. The higher layers
with the most notorious actors are quite easy to explore, and they not contain original
information: NCPs already know what is in there. The interesting but difficult part of the work
is about collecting the websites in the lower layers. That is why we taught NCPs how to use the
WebAtlas tools. We wanted them to confront with the lower layers and discover new resources.
The main tool is called Navicrawler. It is published by WebAtlas and available on their website.
It is a free and open source software that works with the famous free browser Firefox. This
tool is design to semi-automatically collect websites (and links between them) and to bring
useful information about what is collected to the user.
WebAtlas usually teaches how to use the Navicrawler to researchers like sociologists. Several
sessions of practice are generally enough to let these users explore the web on their own. The
Navicrawler is complex but the basic functions are quite easy to understand. The difficulty does
not come from the tool itself, but rather from the complexity of the Web. Even helped by the
Navicrawler it is difficult to explore it. We had only one session to guide the NCPs (the first
workshop). WebAtlas explained the main functions of the Navicrawler and presented
theoretical models of the Web and the complexity. But NCPs had to practice on their side once
back in their country. We had contacts by phone and e-mails, but we could not be there to
help them practicing.
The NCPs have different backgrounds. Some of them were very interested by the methodology
and the tools, because they saw that it could be useful for their own work, out of this project.
Some NCPs were just following the methodology as a part of this task, and some of them had
no problems while others found it difficult. Before the first workshop we asked NCPs to give us
some URLs, and we made a map with it. The map was not very good but they had a sample of
what would be done with their Navicrawler practice. This was a good point, and the NCPs got
involved in the methodology during the first workshop. But after that those who had difficulties
could not get the help they needed from us. They collected the URLs we needed by their own
12
mean (search engines, databases, by their relational network...). It is not a big deal but we
think that they did not benefit from an opportunity to discover new resources. As a result we
asked some countries to get more resources after the second workshop, to balance the
amount of websites-actors between countries.
Depending on the profile of the NCP, it may or not be relevant to improve their Web watching
skills. For NCPs who are not involved in knowledge management, it is more difficult to
stick to this task and benefit from it. This issue due to the various profiles of NCP could be
resolved by asking NCP to get help from colleagues that get involved in knowledge
management.
When WebAtlas teaches the practice of Navicrawler to researchers, they usually have
difficulties. Compared to them, the NCP had a better understanding of this methodology. But
researchers often work with other researchers, and they help each other so that they succeed
in practicing. We think that it could be useful to make NCPs help each other. It is useful not
only for the practice of Web exploration, but also for sharing knowledge about what is
collected.
It is difficult and expensive to make all the NCPs travel to a single place to attend a workshop.
If we had to make this work for every European country, it would even more difficult. That is
why we think that it would be useful to propose local workshops, with only local
countries, in addition to the common workshops. Each local workshop should be available
for a group of 4 to 8 countries, like for example northern countries and Baltic republics. It
would take place in a place where it is easy for these NCPs to travel. It would be useful for
these countries to share their results because they probably collaborate a lot. And it would be
the good way for us to help everyone practicing.
How works a distributed network of experts?

A distributed network of experts is a key-feature for monitoring SiS policies and societal issues
in European Community. Building up an efficient network of observers is quite difficult and
even more difficult in the first period of the system. We then had to take account of many
differences between members of the new “community”.
In theory, a distributed system of observers (each one involved in identification and
description of Web resources) “dialogues” with a central component (i.e. an online database
accessible through a personal workspace). In this central component, a process of indexation
and archiving contributes to a general library of documents. Centralization of data and
computational processes allows graph calculations on a large set of documents, especially in
order to build cartographies (European and national ones).
The EUROSIS system can be considered as a sum of local expertises: qualitative tasks to
select and describe Web resources) collaborating in order to build a general corpus on which
quantitative tasks are possible. In this way, the set of EUROSIS cartographies can be
considered as the result of a quali-quantitative process.
13
A distributed network is designed to fulfill collective goals (ex. European cartography and its
common trends) as well as individual goals (national reporting, national cartography of SiS,
local index of actors…).
The practice of such a distributed work implies the association of theory and technological
experimentation. The actors of this network need to master the basis of the theory. It was not
easy to transfer this knowledge because it is not always related to the profile of the NCPs.
Nevertheless some theory is unavoidable as soon as we try to share knowledge about a new
and complex object like Web mapping. The network needs to understand what it is working on
in order to coordinate. The theory and the practice of tools benefit one another. The strength
of such network resides in its capacity to coordinate the practices of actors, in their theoretical
aspect as well as in their technological aspect.
The NCPs have to manually work on the corpus in order to guarantee its quality. We do not
rely on automatic processes on this point. But automatic processes are needed to manage big
masses of data. Manual as well as automatic processes are necessary. NCPs have to switch
from one to another many times. The theory is required to be efficient on the manual process,
while experimentations allow to successfully monitor automatic processes.
Building a common description of the resources

To get interesting data, we need to describe our resources. Each actor, represented by its
website, has to be described. Some descriptions depend on facts, and it is easy to deal with
these. The country of an actor is well determined, even if there are some exceptions. So as the
social type: NGO, institution, university... But describing the actors by topics is far more
difficult.
Each actor has an activity in one or several topics or themes. These can be ethics, gender
studies, nanotechnologies... Our initial point of view was to let the main topics emerge. That is
why we asked each NCP to describe its resources with free tags. Our idea was to build big
topical clusters by grouping these tags. How to gather tags? Because each NCP could use
many tags to describe each actor, we used the “co-occurring” criteria: two tags are link when
they appear often together to describe actors. With this criterion we wanted to compute a
graph and to observe emerging clusters.
The NCPs used 1200 different tags to describe their actors. But most of these tags meant the
same thing. We manually gathered the synonymous tags. We then had a set of 200 tags. Take
a look at the graph of co-occurrence:
14
This graph is typically a scale-free network. No cluster emerges. There is a minority of hugely
connected tags and a majority of poorly connected tags. Why are there no clusters?
We know that some clusters exist. We know that the “information society” is not the same
thing as “gender issues”. We know that “nanotechnologies” is not the same topic as “nuclear
energy”. Why every tag seems connected to every other tag? Because for each pair of tags,
you will probably find an actor that is involved in both. You will find some actors that work on
gender issues and information society. You will find someone involved in nanotechnologies and
nuclear energy. This is typically the signature of complexity. Despite the fact that we tend to
separate everything, there are a lot of transversal links that remind us that knowledge does
not fit in a simple set of boxes.
So we failed to observe the topical clusters, because what we observed was the massive
presence of trans-topical links. We could have used an algorithm to force the clustering, but it
is not the good way to proceed if you want to understand how things work. So we tried to
make the clusters ourselves. But if you look at these 200 tags you will see that they are clearly
different one from another, even when they are close. Would you melt e-learning and e-
government, or research funding and expertise? We thought that these questions were
important enough to share them with the NCPs.
We discussed about this during the second workshop. And we underestimated the complexity
of this once again... We started this work by an open discussion about the good way to gather
the tags. But no common description emerged. When some NCPs proposed to gather some
tags in a big topic, there always were other NCPs to find it irrelevant. Because this open
discussion failed, we then tried something more constructive. We splitted the NCPs into three
groups, each group working on a good list of topics to describe the corpus. And we then
compared the results to find the common points. Two groups provided similar results whereas
the last one came out with a very different approach.
 Groups 1 and 3: Agriculture, Chemistry, Communicating science, Economy and politics,
Energy, Environment and sustainable development, Ethics, Food, Gender issues,
Geology, GMOs genetics and biotechnologies, Health, ICT and telecommunication,
Nanotechnologies, Physics, Science education, Security, Social sciences and culture,
Transport.
 Group 2: Education methods, Ethics in science and technology, Gender, Governance of
science, Science communication, Universities and public research bodies.
15
Note that both sets of tags are closely related to the 7 th Framwork programme (FP7). But as
the first set describes the topics of science quite the same way as the FP7, the second set
describes the topics of the Science in Society program of the FP7. Both sets have topics in
common, like “gender”. But some tags take place in completely different perspectives. In the
second set you will find “governance of science”. This tag is not in the first set, but it can be
associated to any of these tags (governance of chemistry, governance of geology...). The
perspective is different. The tags of the second set are about science, while the tags of the first
set are parts of science.
After discussion, all NCPs agreed on the first set with some modifications due to the second
set. We do not know if these tags are “the good ones”. There are probably other relevant sets,
and we know that this choice is satisfying. So, finally we sticked to the FP7 perspective.
And we could have done that since the beginning. Should we have?
We need to reduce some complexity of data to observe significant patterns in the maps. We
want to look at the complexity of the interactions between actors, that is why the complexity
of thematics is not our priority. We will build maps where each topic has its own color: if we
have 200 topics then the map cannot be read. We need a very synthetic description of the
topical aspect of actors.
But the question remains. Our conclusion is that another relevant approach of Science in
Society could be the analysis of the topical complexity. We know that knowledge does
not fit inside Dewey is classification. With the Web we can observe it and we could have some
surprises.
By the way, we now know that it is a loss of time to work on the topics in a mapping of
interactions. We need to reduce complexity somewhere, which is why we have to choose
between analyzing interactions and analyzing the topical landscape. NCPs have different points
of view on the good way to separate Science in Society in a set of topics. Maybe different
countries just do not see the science as the same thing. That is why no common set of
topics can simply emerge from this kind of distributed work. There will always be
contradictory descriptions of the resources. Our chance is that the European Commission is
making this work of coordination.
16
Results
General Picture
Summary
The European SiS Network only exists as series of national SiS networks. But despite their
differences, these networks share a common structure.
The core of the SiS network is the “Academic Science”: Universities and Secondary Schools,
Research Centers, Sciences Centers and Museums... And it is important to notice that Media
also participate to this core structure. Actors tend to interact a lot within the core like a
community (the national scientific community).
But there is an extension to this core that we call a “bridge”, that gathers actors concerned
by science: NGO-CSO, Web-specific actors (as Portals) and Policy Makers and Governmental
Organizations. These actors do not necessarily constitute a community, but extend the core
in different directions that allow other poorly connected actors (networks of organizations,
personal websites, companies...) to connect to the core.
The bridge is the most important part of a SiS network because it allows science to spread
out of its own sphere. The “academic” core is present in almost every national SiS network
but it is remarkable that the strongest networks always have this extension of the core. We
suggest that, if the first condition to build a strong network is to get a core academic
subnetwork, the second condition is to extend it through civil society (NGO-CSO), independent
actors (websites, portals) and Policy makers.
17
General Map
The general map shows that every actor binds more to actors of the same country. That
is why the clusters in the general map essentially represent countries (that is why we colored
the actors by country).
General Map: actors aggregate by country (colors by country)
Naturally this result was expected. Actors interact more inside the same country, and this is
strengthened by the fact that on the web, different languages tend to separate (and European
countries have many different languages).
To understand how actors interact we have to look inside every country. We will discuss
about the differences between national SiS networks in the next section. The good way is to
look at interactions between different types of actors rather than different nationalities. The
general map does not help us for this task, but national maps and statistics do. That is why we
will rely a lot on the statistics about types of actors:
A) NGO-CSO
B) Portal or other independent Websites
C) Science centers and museums
D) Research centers
E) Advisory bodies
F) Universities and secondary schools
G) Network of organizations
H) Policy makers and governmental organizations
I) Media
J) Events/projects
K) Companies
L) Subnational and local actors
M) Personal websites
N) International
Connectivity structure of the types of actors

The way an actor links to others strongly depends on its type. For example we observe that
Universities bind a lot and constitute the backbone of several national SiS networks. They also
tend to bind a lot with other types of actors such as Research Centers or Science Centers. On
the contrary, companies do not link each other, and are poorly cited by other actors. What is
more there are many Universities and Secondary Schools while there are few companies in our
European SiS network. It is clear that Universities and Companies play a very different role,
18
have very different types of interactions. When comparing these different types of actors, a
global pattern emerges. This pattern may be interpreted as a hierarchy, and because our
network is a “small-world” (or “complex”) one, we choose to represent this pattern as
“connectivity layers”.
The three connectivity layers: pattern for the types of actors
We found three connectivity layers. It is important to understand that an actor is in a layer

if it fits its profile, and that this profile is determined by how actors link inside the layer, and
how they link with other layers. That means that the way to determine which actor is in which
layer has to be made recursively. Nevertheless the profiles are strong and there are big
differences between the three profiles.
– The higher layer is the easiest to determine. Its main characteristic is the following:
only an actor in the higher layer is more or almost as much cited by other actors in the
higher layer than it cites them. In other terms, there are more links from other
layers to the higher layer than from the higher layer to the other layers, and
actors in the higher layer link a lot together.
Furthermore each type in the higher layer has many actors. This layer gathers the
“academic science” part of SiS and Media:
– Universities and secondary schools
– Media
– Research Centers
– Science Centers and Museums
– The intermediate layer contains other types of actors that are well connected, but
they are less cited by the higher layer than they cite it. What is more, the
intermediate types of actors have almost symmetric links with the lower layer.
This layer gathers three types of actors:
– NGO-CSO
– Portal or other independent websites
– Policy makers and governmental organizations
– The lower layer contains poorly represented and poorly connected actors. In
particular these types of actors have more links with other layers than with their
own. These types of actors are:
– Network of organizations
– Personal websites
– Subnational and local actors
– Advisory bodies
– Companies
– Events/Projects
19
How types of actors link depending on the three layers
Two things are remarkable in this structure. The higher you are, the more you link to
actors in the same layer. And the higher you are, the more you are cited (by lower
layers).
Interactions between types of actors

The higher layer (Universities, Research Centers, Science Centers and Media) is the backbone
of the European SiS network. As a matter of facts the European SiS Network is illusory, it is
rather a conglomerate of national SiS networks. In other words the higher layer is the
backbone of a typical SiS network.
The common point of Universities, Research Centers and Science Centers is that Science is the
main part of their mission, of their activity. We will call these categories “the academic
science” (even if it is not exactly the case). Anyway these three categories share a theme and
a common structure. They often draw a strong and common subnetwork in the national
networks (see the analysis of the national maps for more information). They take part of the
same structures but some countries are more Science-Centers-oriented (such as France) while
other are more Universities-oriented (for instance, Finland). And we have to assume that
Media actors are a part of this “academic science” structure (this category fits our criteria).
On the contrary the lower layer gathers actors that do not play an important role in the
SiS network. But we have to take into account that each of these categories are poorly
represented; it would have been difficult, for a category that gathers about 5% of the actors of
the corpus, to be linked a lot by other types of actors. Nevertheless it may be statistically
possible. It just means that none of the categories in the lower layer can be considered as a
remarkably connected type of actor. In other words the links concentrate where many
actors gather. The global strength of the SiS network does not come from a small amount of
hyper connected actors, as in an explicitly hierarchical network, the strength of the SiS
network comes from its strong subnetworks (the higher layer). This does not mean that
there is no hierarchy, but it is what we may call a “2 nd degree hierarchy”, a hierarchical
structure that does not exclude distributed patterns, that is strengthened by these distributed
structures. In this perspective the lower layer is just the “miscellaneous” part of the network.
Now that we have seen the role of the higher layer and the role of the lower layer, it is easier
to understand the role of the intermediate layer. Even if they are well connected, these
actors play a different role than the higher layer, for two reasons. First, the higher layer is
exclusive: its characteristic is that you have to be as well cited by the higher layer than you
cite it to be in it. In other words, the intermediate layer has more links to the higher layer than
the higher layer has to it. On this point the intermediate layer is like the lower layer. But the
intermediate layer has nevertheless many more links with the higher layer than the lower layer
has, even if they are asymmetric. Second reason, the actors of the intermediate have almost
as many links to than from the lower layer. For these two reasons we interpret the
intermediate layer as a bridge between the lower layer and the higher layer.
The “academic science” is the backbone in the SiS network and the intermediate
layer is a bridge to it. These intermediate actors represent those who, if they are not directly
dedicated to science, plug to the academic community because they are concerned. NGO-CSO
are often dedicated to environmental issues and then plug to the academic science on
environment (see national analysis on this). Portal or other independent websites, meaning
web-specific actors, are more topically diffuse but play by nature the role of a bridge (because
it is the role of portals in an information system). And policy makers also play this role. The
difference with the higher layer is that these actors are less science-dedicated, and more
society- or politics-oriented. We may interpret the intermediate actors as “those who are
concerned by science”.
– CORE : Academic Science (Universities and Secondary Schools, Research Centers,

Science Centers and Museums, Media)
 Strong national subnetworks
– BRIDGE : Science Concerned (NGO-CSO, Portals and other independent websites,
Policy makers and Governmental Organizations)
 Well connected, extends the core as a bridge to other actors
– OTHER : Miscellaneous (Network of organization, Personal websites, Subnational and
local actors, Advisory bodies, Companies, Events/Projects)
 Poorly connected, but links to the bridge and the core
National maps
Specificities
The national SiS networks draw various structures. There are big and small, strong and weak
networks. But two tendencies emerge.
– The bigger and stronger networks tend to be more connected to other countries
– The strong national networks are always supported by an academic subnetwork, but the
presence of NGO-CSOs and web-specific actors (such as portals or independent
websites) makes the difference for the strongest.
Armenia
A quite poorly connected network, with a lot of Research Center and few Web-specific
actors
Few links with other countries
Belgium
A very strong network, with well distributed types of actors, and where Wallon and
Flemish actors tend to separate
Some dedicated subnetworks, in particular about Sustainable Development
Many links with other countries
Bulgaria
A strong network with many science institutions counterbalanced by many NGO-CSOs
Few links from other countries, but many links to other countries
Czech Republic
Inconsistent network with many science institutions counterbalanced by many NGO-
CSOs
Estonia
Strong network with many science institutions counterbalanced by many NGO-CSOs,
and many media
Some dedicated subnetworks about Environment
Finland
Very strong network where most of actors are science institutions, in particular
Universities.
France
A quite weak network with many science institutions, in particular Science Centers
Museums tend to aggregate but most of them are totally disconnected, and there is a
subnetwork of NGO-CSOs about Environment and Democracy
Hungary
A strong network, with well distributed types of actors including many web-specific
actors that play an important role
Italy
A weak network with no institutional subnetwork
Montenegro
A very weak network with many companies and media, policy makers and advisory
bodies and some dedicated subnetworks, in particular to scientific institutions
Very few links with other countries
Poland
A quite weak network with many science institutions that draw a strong subnetwork
Portugal
A quite strong but very small network with many Science Centers that draw a strong
subnetwork
Armenia
118 actors have been collected, which is few compared to other countries. We found 218
hypertext links between these actors, which means a low links density of 0,016.
Armenian actors are quite poorly connected together, compared to other countries.
Note that like for every country, there are many more internal links than links with actors in
other countries (it is a general and quite obvious result).
The most represented actors are Research Centers, and they massively occupy the Armenian
SiS territory. There are also many companies and NGO-CSOs. There is no subnational or local
actor, no portal or other independent websites, and no personal website, and very few media
or networks of organizations. This may indicate that the Web in Armenia does not play the
role of a public space for discussing Science.
The main Armenian actor in the national network is the National Academy of Sciences. It
attracts 12% of the internal links. It is classified as a Research Center, and a quick look on the
national map shows that it is the center of the research centers.
Zoom on the Research Centers cluster (in light blue) in the Armenian map
The National Academy of Sciences is also the main hub. 11% of the links to other
Armenian actors come from it. This is due to the fact that it links to most of the Research
Centers.
Research Centers do not constitute a strongly connected network, but the hierarchy is very
strong. As you see in the zoom above, the structure draws a star and its center is the National
Academy of Sciences. This means that the different institutes do not link together, probably
because the structure reflects an institutional network where the only necessary link is the
National Academy of Sciences itself.
Armenia has very few inbound links. It seems that the Armenian network is not a major
resource for the European SiS network.
Armenia has quite few outbound links to other countries. The main target is the
International or European level (67%). Armenia has strong interactions with no other country
in our set (but maybe with other European countries that did not participate to this task).
The most represented themes are “Governance, expertise, ethics” and “Information and
communicating technologies”.
Belgium
204 actors have been collected, which is very much compared to other countries. We found
1489 hypertext links between these actors, which means a very high links density of 0,036.
Note that density is expected to lower as the count of nodes increases; the contrary happens
here and it is remarkable.
The Belgian SiS network is very strong, compared to other countries.
The different categories of actors are well balanced. The most represented are NGO-CSOs,
Portals and Research Centers. There are few media, and no companies. The “society”
part is well represented, and with the strong presence of “web-specific” actors (portals,
personal websites), it means that the Belgian SiS network is active and well developed.
The main Belgian actors in the national network are the Université Libre de Bruxelles and
the Belgian Science Policy (cited respectively by 56 and 51 other Belgian actors).
Research.be is the main hub (links 91 Belgian actors) and Belgian Science Policy is the
second one (56).
Research.be is the main portal
SiS in Belgium draws a single and well connected network, but some components appear. First
we have to notice that the Flemish actors and the Walloon actors globally occupy
different areas, besides they are well linked and often melted. This is not a surprise as the
two communities do not share language (and we know that on the Web different languages
separate).
Another component can be identified as actors involved in sustainable development (mainly
NGO-CSOs and portals). They tend to separate from the rest of the actors involved in Science,
but there are bridges that are actors involved in environment.
NGO-CSOs and Portals about Sustainable development
As we have seen, there are many NGO-CSOs in Belgium. Most of them distinguish from the
main “science” component. We identify two small clusters of NGO-CSOs. The first is Dutch
language and dedicated to philosophy in science. The second one is dedicated to women and
gender issues.
NGO-CSOs cluster about philosophy (related to science)
NGO-CSOs cluster about women
Belgium has very much inbound links (184). The main source is France (23%) but other
countries are important too (Bulgaria, Finland, Hungary, Poland and International). Belgium is
well linked by many countries and it denotes an important role in the European SiS
network.
Belgium also has very much outbound links to other countries. The main target is also
France (41%) and in second the International or European level (33%). Other countries are
poorly cited by Belgian actors (excepted Finland, 10%). This reinforces the hierarchical
importance of the Belgian SiS network, which is well cited despite that it does not cite many
countries.
The most represented themes are “Communicating Science” and “Governance, expertise,
ethics”.
Bulgaria
135 actors have been collected, which is a lot compared to other countries. We found 581
hypertext links between these actors, which means a high links density of 0,032.
The Bulgarian SiS network is strong, compared to other countries.
The most represented are NGO-CSOs, Universities and Research Centers. Every other
category is represented. The “science” part is strong but is well counterbalanced by NGO-
CSOs.
The main Bulgarian actor in the national network is the Bulgarian Academy of Sciences
(cited by 38 other Bulgarian actors). This actor is also the main hub (cites 43 Bulgarian
actors).
Most of the other important websites are universities: they draw a strong subnetwork.
The Bulgarian Universities subnetwork is strong (actors in blue)
But the universities subnetwork is not strongly connected to the institutes (research centers).
The research centers draw a very hierarchical structure around the Bulgarian Academy of
Sciences. There are few connections between institutes. This is probably due to an institutional
approach of building websites, where the main bridge has to be the BAS.
The hierarchical network of Research Centers draws a “star” around BAS
Bulgaria has few inbound links (51). The main sources are Poland (31%) and the
International level (25%). Several countries do not cite Bulgaria (Czech R., Estonia, France,
Italy). The Bulgarian SiS network is not a main resource for the European SiS network.
But Bulgaria has many outbound links (124) to other countries. The main target is
International (38%) but other countries are also well cited (Belgium 15%, Hungary 13%,
Poland 14%).
The most represented themes are “Governance, expertise, ethics” and “Communicating
Science” and “Science education”.
Czech Republic
70 actors have been collected, which is very few compared to other countries. We found 51
hypertext links between these actors, which means a very low links density of 0,011.
The Czech SiS network is nearly inconsistent, compared to other countries. There is too
few links between actors to bind them in a single component.
The Czech Republic SiS Network is inconsistent:

Actors are poorly connected, or totally disconnected (list on the left)
The most represented actors are NGO-CSOs (30%). The “science” component is also well
represented (Universities 23%, Research centers 17%, Science center and museums 13%).
There are few companies and media and no advisory bodies, no events/projects, no
personal websites and no portals. This means that the web as a public space is not
developed at all in the Czech SiS Network.
The main Czech actor in the national network is the NGO-CSO Gender Studies (cited by 8
other Czech actors). The main hub is Czech Helsinki Committee (cites 6 Czech actors), and
the Forum 50% is the second authority as well as the second hub (cited by 5 actors, cites 5
actors, in Czech Republic).
NGO-CSOs draw the main subnetwork, that gathers most of the internal links in Czech
Republic. The three actors mentioned above take place in this subnetwork. It is the only part
of the Czech SiS (non-)network where resources significantly gather.
The Czech NGO-CSO subnetwork
Czech Republic has very few inbound links (16). The main source is International (56%) and
other sources are not significant (most countries do not cite Czech actors at all).
Czech Republic also has few outbound links to other countries (72). The main target is also
International (38%) but other countries are cited too (Belgium 8%, Finland 10%, France 10%,
Hungary 8%, Poland 8%, Portugal 14%).
The Czech SiS Network is not important in the European SiS landscape (it does not
mean that “individual” actors do not play any role). It is remarkable that 9 actors (13%) have
connections with actors in other countries while they have no connection in Czech Republic.
This is probably a consequence of its very weak internal connectivity. Note : this situation
probably makes that search engines do not give a good rank to the Czech SiS websites.
The most represented themes are “Communicating Science”, “Governance, expertise, ethics”
and “Science Education”.
Estonia
981 hypertext links between these actors, which means a high links density of 0,024.
The Estonian SiS network is strong, compared to other countries.
The most represented are NGO-CSOs (18%), Media (17%), Portals (16%) and the
“science” component (universities 13%, science centers 11%, research centers 16%). There
are no advisory bodies, no companies, and no personal websites. The presence of media
is remarkable. The Estonian SiS network is very specific, because it is almost exclusively
composed of Science, Media, Portals and NGO-CSOs (total 91% of the actors).
The main Estonian actors in the national network are Eesti Päevaleht and Postimees (cited
respectively by 37 and 36 other Estonian actors). These two actors are medias.
Nature Pages is the main hub (links 58 Estonian actors) and is a portal.
In general the authorities are Media and NGO-CSOs while hubs are (obviously) Portals, and
also Media and NGO-CSOs. Media are very strong in the Estonian Network, but also NGO-
CSOs.
The main authorities, that are media, are very central (as expected). They gather and occupy
the center of the bottom-right area of the map.
Main authorities (media): Eesti Päevaleht and Postimees

They take place in the center of the biggest component.
On the contrary, the main hubs take place around a zone (that we call “bridge”) that connects
the main component (bottom-right) and a smaller component (top-left). We will explain below.
Main hub, a portal, below the bridge.
Second hub, a media above the bridge
The main component contains most of the Estonian actors, including museums, schools,
universities, media... We may identify two centers. The first is constituted of the main
authorities, and is closer to the center of the map (top-left of the component). The second
one is the Estonian Ministry of Education and Research.
The main component. First center, top-left: Eesti Päevaleht and Postimees
Second center, bottom-right: Estonian Ministry of Education and Research
The smaller component is dedicated to Natural Parks. Most of its actors are Research
Centers. And the “bridge” zone that connects the two components contains mainly actors
dedicated to the Nature issues. It contains many NGO-CSOs.
Smaller component (top-left of the map): Natural Parks (Research Centers).
Bridge (between the two components): NGO-CSOs about Nature.
Scheme of the structure of the Estonian SiS network
As observed in other studies 1, the Natural Parks separate from the rest of the SiS network
1
WebCSTI by WebAtlas
http://webatlas.fr/download/docs/WebCSTI10.pdf
(besides bridges always exist). We suggest that it is due to the strength of the “Nature” field
on the Web. The authorities are not cited a lot by Natural Parks. This makes this subnetwork
distinguish on the map. But Hubs take these in account and it explains their situation. They
connect to many actors including in the main component, the bridge and the Natural Park
subnetwork.
The Estonian SiS network is very hierarchical. Many actors are disconnected (19%) but hubs
and authorities have many more links than other actors.
Estonia has few inbound links (51). The main sources are Italy (25%) and Finland (24%).
Other countries do not cite a lot Estonia. Despite its strong constitution, the Estonian SiS
network is not a main resource for the European SiS network.
Estonia has very few outbound links to other countries. The main target is Finland (48%)
and then the International or European level (24%). Other countries are poorly cited by
Estonian actors. Estonia is clearly citing Finland in priority. We suggest that Finland plays the
role of a bridge between Estonia other European actors. Furthermore, Estonian SiS
network seems self-sufficient (national actors do not rely a lot on foreign actors to expand
their local network).
The most represented theme are “Science Education”, “Communicating Science” and
“Environment”.
Finland
hypertext links between these actors, which means a very high links density of 0,088. The
density is the highest, by far, in our whole corpus (the second one is Belgium, 0,036).
The Finland SiS network is very strong (the strongest), compared to other countries.
The different categories of actors are strongly unbalanced. The most represented are
Universities (51%). There are few Science centers, media, portals and personal
websites, and no companies, no events/projects, networks or subnational actors. The
“web-specific” actors (portals, personal websites) are remarkably underrepresented while
Universities occupy a huge part of the network. The Finnish SiS network is mainly
academic, that makes it very strong but it does not work well as a public space compared
to other countries.
The main Finnish actor in the national network is the University of Tampere (cited by 37
other Finnish actors), but we have to notice that many other universities are almost as well
cited.
The Academy of Finland, the Finnish IT center for science and the Finnish science and
technology information service are the main hubs (they link respectively 43, 42 and 41
Finnish actors). Contrary to the authorities these are not universities, but respectively a Policy
maker, a NGO-CSO and a Portal.
Finland has very much inbound links (141). The main source is Poland (30%) but other
countries are important too (Estonia 18%, Belgium 13%, Hungary 11%). Finland is well linked
by many countries and it denotes an important role in the European SiS network.
Finland also has very much outbound links to other countries. The main target is also
Poland (26%) and in second the International or European level (23%), and after Belgium
(17%) and Hungary (16%). Armenia, Czech Republic, Italy and Montenegro are not (or very
poorly) cited by Finnish actors. The Finnish connections with other countries are almost
symmetric, even if Finland is a little bit more selective when citing other foreign actors.
As expected, the actors that point the most to foreign actors, as well as the actors that are the
most cited by foreign actors, are mainly Universities. The Universities subnetwork is so strong
and massive that it determines the global profile of the Finland SiS network. In particular the
Universities subnetwork is well distributed and not strongly hierarchical. Most of the
universities are very well connected, so that no particular hub or authority emerges. We
suggest that this behavior of distributing connectivity might explain the symmetric connectivity
profile of Finland amongst other countries (each country points to Finland with almost the
same amount of links as it is cited by).
The Finland SiS network is typical of a regular, strongly connected, weakly hierarchical
network. This might be due to a good mastering of the connectivity policies by most of the
actors. But it also denotes that no public initiative emerges in this part of the Web, so that
there is no independent subnetwork.
The universities network in Finland is massive and well distributed.
The most represented themes is “Science Education” (according to the importance of

universities).
France
676 hypertext links between these actors, which means a low links density of 0,016.
The France SiS network is quite weak, compared to other countries, and it is partly due to
the amount of disconnected actors.
The most represented actors are Science centers. There are few companies, and no portal
or independent website. The part of Network of organizations is important compared to
other countries.
The main French actors in the national network is the Cité des Sciences et de l'Industrie
(cited by 47 other French actors). Agrobiosciences is the main hub (links 37 French
actors).
The relative weakness of the French SiS network is partly due to its important count of
disconnected actors (26%). The remaining connected actors draw a quite strong
network. Amongst the disconnected actors we observe an important proportion of museums.
These might be expected to constitute a subnetwork, but it is the exact opposite. Many of the
museums ignore (and are ignored by) the other SiS actors. By the way there are some
connected museums, and they draw a small subnetwork. This also means that they do
not connect to actors of another type.
Most of the French museums are connected to no other actor
The remaining connected museums draw a small subnetwork
Two components distinguish in the French SiS network. The first (and main) component
contains many “CSTI” labeled actors (Scientific, Technical and Industrial Culture). This
component occupies the bottom-right part of the national map. The second component
occupies the top-left part of the map. This component contains a subnetwork of NGO-CSOs.
These two components give two different orientations to the France SiS network, but these are
not clusters and many actors are out of these components. The main hub, Agrobiosciences, is
close to the NGO-CSOs component, but the main authority, Cité des Sciences et de l'Industrie,
is in between at the center of the map.
The “Scientific, Technical and Industrial Culture” component
The NGO-CSOs (green) component

France has very much inbound links (146). The main source is, by far, Belgium (53%).
Almost every country has at least 1 link to France, but otherwise it is not well linked by other
countries. France SiS Network is a noticeable resource in the European SiS network.
France has many outbound links to other countries (74), but it remains less than inbound
links (it is the only country with Hungary). The main target is also Belgium (58%) but most
of the countries are not cited by France (Armenia, Bulgaria, Estonia, Hungary,
Montenegro, Poland, Portugal). This gives importance to the France SiS network, that is quite
well cited despite that it does not cite many countries.
The most represented theme is “Communicating Science”.

Hungary
hypertext links between these actors, which means a high links density of 0,029.
The Hungarian SiS network is strong, compared to other countries.
The different categories of actors are well balanced. The most represented are NGO-CSOs
(18%) and Universities. There are few advisory bodies and events/projects, and no
networks of organizations. The “web-specific” actors are well represented (portals, personal
websites) while “science” is not overrepresented. The Hungarian SiS network is active and
well developed.
The main Hungarian actor in the national network is the Hungarian Academy of Sciences
(cited by 33 other Hungarian actors).
The Institute for Political Sciences is the main hub (links 39 Hungarian actors).
NKTH is the second authority as well as the second hub (cited by 24 actors and pointing to
28).
The left part of the map contains more “science” websites, while the right side contains more
Portals and NGO-CSO. The left part is science-oriented while the right side is society-
oriented. But despite this distinction, the two parts are interconnected in a single national
cluster.
Structure of the Hungary SiS Network: two interconnected parts
Hungary has many inbound links (88). The main sources are Finland (25%), Poland (20%)
and Bulgaria (18%). Bulgaria is linked by many countries and it denotes a noticeable
importance in the European SiS network.
Hungary also has many outbound links to other countries. The main targets are
International (28%) and Belgium (26%), and also Finland (17%).
Science”.
Italy
118 actors have been collected, which is few compared to other countries. We found 136
hypertext links between these actors, which means a very low links density of 0,010.
The Italy SiS network is very weak, compared to other countries.
The most represented actors are Portals (19%) and Science centers (19%). There are many
Subnational actors compared to other countries, and each other category is represented.
The “society” part is well represented, and with the strong presence of “web-specific” actors
(portals, personal websites), it means that the Italy SiS network is active.
Because the network is weak, there are no strong authorities. The most cited website has 7
inbound links, and the following 6 or 5. On the contrary there are strong hubs. EXPLORA-La
tv della scienza is the main hub (links 33 Italian actors) and Torino scienza is the second
one (29).
The Italy SiS network, organized around two hubs
SiS in Italy draws a single a poorly connected but quite well structured network, if we take
into account many disconnected actors (39%). No specific cluster emerges, and in
particular contrary to other countries, there is no strong subnetwork dedicated to scientific
institutions. The network is organized around the two hubs, but there are still links between
other actors.
The absence of an institutional network is remarkable. In other countries there is often a
network of Universities that makes the national network stronger, but there are few
Universities in the Italian corpus (is it due to the NCP's selection ?). Nevertheless it is also
remarkable that various actors link each other (portals, subnational or local actors, media
and events/projects).
Italy has few inbound links (37). The main sources are Belgium (22%) and France (19%).
Italy also has few outbound links to other countries (71). The main targets are International
(30%) and France (27%).
The most represented themes are “Communicating Science” and “Environment”.
Montenegro
118 hypertext links between these actors, which means a very low links density of 0,007.
The Montenegro SiS network is very weak, compared to other countries.
The most represented actors are Universities. There are few Science centers, few Portals,
and no Network of organizations, no Personal website. The science is underrepresented
while advisory bodies, companies and media are very well represented.
The main Montenegrin actors in the national network are the University of Montenegro and
the Ministry of Education and Science (cited respectively by 11 and 10 other Montenegrin
actors).
Ministry of Education and Science is also the main hub (links 12 Montenegrin actors) but
other actors have almost the same amount of outbound links.
Three components emerge from the Montenegro SiS network.
The first component is dedicated to telecommunications and engineering and contains
many Companies and Media.
The Telecom and Engineering component in Montenegro SiS Network
The second component contains many Faculties dedicated to various fields, and linked
together by the University of Montenegro (that is the best authority).
The Faculties component
The third component is quite similar to the second one, but different. It contains High
Schools and is organized around the Ministry of Education and Science.
The High Schools component
The Montenegro SiS network is very weak and contains many disconnected actors (47%).
Nevertheless it is well structured, clearly organized in an institutional way (subnetworks by
types of actors) where some institutions are the center of the network (Ministry of
Education and Science, and University of Montenegro).
Montenegro has very few inbound links (5 links only), the minimum in our corpus of
countries. It also has very few outbound links to other countries (22, also the minimum).
The main target is International (45%). This poor external connectivity corresponds to the
weakness of the network.
Science”.
Poland
hypertext links between these actors, which means a low links density of 0,013.
The Poland SiS network is quite weak, compared to other countries.
The most represented are Universities (42%). Each other category of actors is represented,
and the “science” component represents more than the half of the corpus (59%).
The main Polish actors in the national network are the Ośrodek Przetwarzania Informacji
OPI Warszawa, the Polish Academy of Sciences and Nauka Polska (cited respectively by
22, 21 and 20 other Polish actors).
Polskie Towarzystwo Fizyczne is the main hub (links 46 Polish actors).
Many actors are disconnected in the Poland SiS Network (29%). They are responsible for
the weakness of the network. On the contrary the remaining, connected, actors draw a well
connected network.
Poland: many disconnected actors (left) and a quite strong network (right).
This makes the Polish SiS network globally weak
The strong subnetwork is actually the network of the Universities. The strength of the
Poland SiS network is due to the amount of academic links between Universities. This does
not mean that no other type of actors is important. In particular the main hub is a NGO-CSO
(Polskie Towarzystwo Fizyczne) and it links many universities. But we suggest that the lack of
variety in this network may exclude some actors and explain why there are so many
disconnected actors.
The main hub is a NGO-CSO

Poland has many inbound links (86). The main source is Finland (42%), and Bulgaria is also
an important source (20%) but other countries do not link much Poland. Poland is well
recognized by some countries but its sources of links are not well diversified. This may be due
to the academic orientation of the national network.
On the contrary Poland also has very much outbound links to various other countries. The
main targets are also Finland (24%) and the International or European level (24%). Other
countries are well cited by Polish actors like Belgium (14%), Bulgaria (9%) Hungary (10%).
The academic network is active but the asymmetry between the inbound links and the
outbound links (more various actors) denotes and intermediary position of Poland in the
European SiS Network. It is a quite important resource for some countries but not widely
recognized.
The most represented theme is “Communicating Science”.

Portugal
hypertext links between these actors, which means a high links density of 0,021. Note this
high density is not necessary significant because there are few actors.
The different categories of actors are strongly unbalanced. The most represented actors are
the “science” component: Universities (31%), Science centers (25%) and Research
centers (22%) for a total of 78%. There are no advisory bodies, no Events/projects, no
Media, no NGO-CSO, no personal website, no Portal or independent website. The
network collected by the NCP is essentially the academic network in Portugal. There is a quite
strong academic subnetwork that contains well connected Science Centers and
Universities, but the Portugal SiS Network remains globally weak (despite its good density).
The main Portuguese actor in the national network is the Ciencia Viva (cited by 11 other
Portuguese actors). Ciencia Viva is also the main hub (links 20 Portuguese actors).
Portugal has few inbound links (38). The main sources are Czech Republic (26%) and Poland
(26%).
Portugal has very few outbound links to other countries (56). The main target is
International (29%) and in second France (21%).
The most represented themes are “Governance, expertise, ethics”, “Communicating Science”
and “Science Education” (note that this theme is overrepresented, and it is probably due to the
important role of Science Centers in the National SiS Network).
Main axis of interpretation
Hierarchy emerging from countries' relations
Where do websites aggregate?

The structure of the links is strongly determined by the countries. As you can see in the
general map, actors bind more when they are in the same country. But we have to look
further... How do actors link in the same country? How do actors link from one country to
another, how do countries link?
As we have seen, actors bind more inside the same country. But different things happen in
different countries. Some countries have a very strong structure, like Estonia. The Estonian
NCP has collected a lot of actors, and they strongly link each other: almost 10 links per
website. Some other countries have a weaker structure, like Czech Republic: almost 2 links per
website. But something does not vary: inside every country the websites tend to link to those
of the same type.
We need to make again an important remark here. It is easy to understand that we talk only
about websites that we actually collected. But do not forget that the links we talk about are
only the links between the websites that we collected. This means that a website can have a
lot of links to other sites, and appear weakly connected in our map. This is the case when
these links point to websites out of our corpus. So when we say that Czech has only 2 links per
websites, we talk about 2 links from a collected Czech website to another collected Czech
website. This low number is not necessary abnormal. On the contrary it is significant that there
are 10 links per website inside the Estonian corpus.
The topic of the websites (representing actors) is important to understand how they bind inside
a country. We often observe strong networks of institutional / governmental websites, and
strong networks of universities, but there are some exceptions. We observe sub-networks in
each country. But the size of these subnetworks and their links vary a lot from one country to
another. We will see later if we can determine different profiles. At the moment we will focus
on the question of strong networks and subnetworks.
We have got some strong networks in countries and some subnetworks inside some of them.
Websites aggregate at two different scales: the national level (one aggregate by country)
and the local level (some sub-networks in a country). There is no strong network with
websites of several countries. We might also think that the European level is one big
aggregate; it is probably true but we cannot tell with these data.
Note that we will not tell anything here about the specific case of each country. We will stick to
the general question of data interpretation, even if it requires taking a close look on some
countries' data.
 Small-World in the countries

How different countries link each other? We first observe that countries link more when they
are close. The typical case is Estonia binding to Finland. This could be a consequence of
historical relations as well as the result of the similarities between the NCPs' profiles. These
questions fall out of the scope of this study. We may presume that most of the countries that
strongly link in the map have strong ties in the “real world”. But the relations between “big”
and “small” countries are more subtle, and this reflects some hierarchy that exists in actual
interactions at the national level. Because we did not analyze every country in Europe, it will
be difficult to describe precisely this hierarchy. Nevertheless we need to explain why it exists
and what it does.
When we observed the connectivity between countries, we were surprised to observe a
“complexity effect”. We will explain what it means, but for the moment let's say that we
realized that countries had a specific type of relations. We were surprised because the general
map does not look like a typical complex network. The networks that we usually collect on the
Web look like an incredible mess, where it is difficult to see where websites link together.
Compared to that, the general map is well organized: each country clearly occupies its
territory. What is more, countries obviously link to their neighbors. But this “obviously” was
hiding something: the links between close countries are asymmetric. This appears when you
compare the amount of inbound and outbound links from a given country to another. And what
is more we observed that this asymmetry is not random at all. This is typical of small-world, or
complex, networks.
 The signature of complexity

The structure of the inter-country links has the signature of complexity. This can be
determined with two observations. On one hand, almost every country links to every
other country. This point reflects the presence of many transversal links. On the other hand,
the quantity of links between two countries strongly depends on the “size” of each
country. The way countries link depending on their “size” is very specific, and we will take a
close look on this.
These “transversal links” are well known in the science of networks. They are identified as
“bridges” or “shortcuts”, because they allow to travel quickly through the entire graph, from
one side to another. In the image above, you can see such transversal links crossing the
graph. Without these, the graph would not be a “small world” one.
When we say “big” or “small” countries, or when we write about their “size”, we do not
necessary mean the size of the geographic area they occupy. We first mean that a “big”
country is one that has a lot of strongly connected actors and a lot of connections with other
countries. On the Web, being “big” means that you have a lot of incoming links, a lot of
access: many visitors and a high ranking in search engines. We may assume that the more a
country is economically strong and populated, the more its research gets fundings, the more
powerful its network is. It is a very rough approximation but it works well with our 12
countries; that is why we say “big” and “small” - but do not forget the quotes. We consider
Finland as a “big” country and Armenia as a “small” country. This may draw a scale that tells
us which country is “bigger” than which one: we certainly will not write the list, for obvious
reasons! But we need to start our analysis with relations between clearly “big” and clearly
“small” countries. We need this approach to observe a certain type of interaction that is
typically complex.
Finland has many more actors than Armenia, but it also has more links per website than
Armenia. That is why Finland has many more links than Armenia. Armenia has not many links
with other countries, and it has a lot more outbound links than inbound links. Finland is “big”
and Armenia is “small”.
Finland has quite the same amount of inbound as outbound links. Who is Finland pointing to?
Mainly the international or European (IoE) actors that we have in our corpus. Few links are
from IoE to Finland, many links are from Finland to IoE. If we assume that the IoE level is
“bigger” than Finland, then we can say that Finland has many links to what is “bigger” while it
has few links from it. And as a result, if we except links with the IoE level, Finland has more
incoming links than outgoing. Then take a close look at this: in Finland many inbound links are
from Estonia, but there are few links to Estonia. If we assume that Estonia is “smaller” than
Finland, then we can say that Finland has many links from something “smaller” and few links
to it. Most links are from the “small” to the “big”.
If we look at Armenia, we will see that there is no country that has many links to it. But there
are many links from Armenia to the International or European level (IoE). Armenia too is
pointing to something “bigger”.
If we observe all countries, we will always observe this phenomenon. Most of links are from
the “small” to the “big”. We call this “attraction to the top”, and it is typical of the
scale-free networks. The links tend to point to more connected nodes. Most of the links start
in lower layers and lead to high layers. From “small” nodes to “big” ones. As a result we
observe something else. The “small” nodes tend to have few inbound links, because there is
nobody “smaller” that could point to them. As well “big” nodes tend to have many inbound
links because there are many “smaller” nodes pointing to them. And this leads us to another
consequence: there are few “big” nodes and many “small” nodes, drawing pyramidal structure.
This is the signature of complexity:
 Some countries have few inbound links and more outbound links. We call these the
“small” ones.
 Some countries have many inbound links from various sources and a less or equal
amount of outbound nodes. We call these the “big” ones.
 There are more “small” than “big” ones. The “size” is totally relative, so that you are “big”
or “small” only compared to others. And there is an important gradient between “big” and
“small” ones.
 Observing the hierarchy

Let's illustrate that with the links between Finland and Poland. From the Finnish point of view,
the situation is well balanced, and the Poland seems to be an equal partner: Poland has a good
amount of links to Finland, and Finland has a good amount of links to Poland. The amount of
links incoming as well as outgoing is well balanced.
Inbound and outbound Poland

But let's look at the Polish point of view. Poland is “smaller” than Finland, and we can see that
because its connectivity profile is not well balanced. Contrary to Finland, Poland has a lot more
outbound links as inbound. The more valuable links are those that come to you, not those that
go out of you. If others put links to you on their website, then they recognize you, and you
may become an authority. Despite its outbound links, Poland has not so many inbound links.
And this is changing the balance with Finland. Because even if there is the same amount of in
and out links with Finland, the part of Finland links is not equal. The part is more important in
inbound links than in outbound links. Finland is actually the country (that we have) that points
the most to Poland. But it is not the case for outbound links. Where do Polish links go? First,
they go to the IoE level (attraction to the top) and then to Finland and Italy. Finland is not the
main destination of the links. But it is the main source.
Inbound and outbound Poland
As we have seen, Poland and Finland seem balanced from the Finnish Point of View. But it is
not the case in the Polish point of view. This is because Finland is well cited, and Poland is not
as well cited. Actually Finland and Poland are not in the same layer of connectivity; Finland is
in a higher layer of connectivity.
Now that we have explained how to observe this relation, we can look at it for every country.
What do we observe? Finland is an important source of links for Hungary and Estonia, and as
we have seen, Poland. But Poland is at its turn an important source of links for Armenia and
Czech Republic. And we need to add that no country is an important source of links for Finland,
and that Armenia as well as Czech Republic are an important source of links for no other
country. We have here a nice example of hierarchy, which we can represent as a pyramid with
Finland at its top.
Hierarchy Finland
This is the main hierarchy of our set of countries. It does not include all of the countries
because they do not all have hierarchical relations. France and Belgium for example form a
balanced couple.
Couple FRance-BElgium
Some other countries do not really have hierarchical relations. It is the case of Montenegro,
because it is an important source of links for no country and it has not enough inbound links to
tell who is an important source of links. It is also the case of Bulgaria, but despite it has many
inbound links, no country plays a particularly important role there. As we have told before we
have only a part of this hierarchy, and we cannot fully draw it. For example Bulgaria may be
an important source of links for a country that we do not have, or some country that we do not
have may be an important source of links for Finland. In particular it would be very interesting
to get the countries that could play an important role for other countries, like United Kingdom,
Germany, Greece...
Let is make an important remark about this hierarchy. One may think that it is an
administrative, or institutional, hierarchy. It is probably not (only) the case. An “official”
hierarchy is a tree-like structure. This means that you are linked with entity that is directly
above you, and the entities that are directly below you. What is specific in these “classical”
hierarchies is that you are not allowed to have relations with the “top”. You are not supposed
to be in relation with the superior of the superior of your superior, you are supposed to use
intermediates (your direct superior and that is it); and the same to the bottom of the
hierarchy. We will call this a first degree hierarchy. The hierarchy we observe is different, it
is a second degree hierarchy and that is why it is not probably (only) due to “institutional”
relations. It is also the reason why it is difficult to observe. Compared to a 1 st degree
hierarchy, it is like if each entity is connected to all superior entities, and not only the entity
strictly above. In particular every country is connected to the European on International level,
which is the “top”, and in this 2 nd degree hierarchy the top is the core. This makes so many
links that it is difficult to detect the hierarchical relations, and you have to notice that countries
are not connected in a tree-like structure. The links do not draw a 1 st degree hierarchy.
However, the distribution of the links makes another hierarchy emerges, the 2 nd degree one.
So there is no 1st degree hierarchy but a 2nd degree one, and we had to take a close look on
links to observe it. That is why we call it “2nd degree”.
 Interpretation of the hierarchical relations

We cannot draw the full hierarchy and if we had the full set of European countries, our
observations would certainly be different. But there is something that we would nevertheless
observe: the hierarchical relation determines different layers. We do not think that we would
discover a country “under” Armenia as well as “above” Finland. And clearly we have nothing
like that in our set of countries, even if there are balanced relations like France and Belgium
(the hierarchy is not strict).
At this point we have got to reformulate some things. It is not necessary to think on the basis
of “big” and “small” countries. This was good for understanding and observing the
phenomenon, but the relation is self sufficient. The relation determines which country is
“bigger” or “above” which country, not the inverse. We are going to leave this and determine
the specificities of this hierarchical relation from straight observations. The main criteria that
we used for now is “who is an important source of links for who”. This works well but it is only
a part of the phenomenon. Here is the big picture.
We have got to explain something that one may find paradoxical. A characteristic of the
relation is that there are more bottom-up links than top-down. And another characteristic is
that the upper country is an important source of links for the lower one. We have already seen
that, with the relation between Finland and Poland, but it is useful to look at it again.
So there are more bottom-up links but there are top-down links that are important for the
lower country. This is the point: it is not because top-down links are important for the lower
country that there are many of them, or that they are important for the upper country. It is
the contrary! The top-down links are not important for the upper country, because it has more
links and because the outbound links are less important. This is precisely the problem of the
lower countries: they get less inbound links while these are more important. The lower a
country is, the more it lacks of inbound links.
We observe a specific relation between countries. This relation gathers this set of
characteristics:
• The relation is asymmetric and we define its direction as vertical: the
upper country and the lower country
• The lower country has fewer inbound links than the upper one
• The lower country has less various inbound links than the upper one
• In proportion, there are more bottom-up links than top-down
• The upper country is an important source of links for the lower one
And these relations chain each other so that we make these observations:
• The relations draw a partial and non-strict hierarchy, strongly
determined by connectivity: if AB and BC we never observe CA, knowing
that we observe some DE
• We call the vertical levels connectivity layers and there are less
countries in higher layers - the European/International layer is the highest
• This relation is typical of scale-free (or small-world) networks
NB: do not confuse the hierarchical relation with the links: there are links between
unrelated countries
Our hierarchy is originated in the direction of the links. This is the key: inbound links are
rare. And there is a competition to get them, even if we do not see it at an eye's glimpse.
Fundamentally the hierarchy is the consequence of this competition; we have got to assume
that the hierarchy we observe reveals a competition between countries. And it is quite easy to
understand it in the field of science, because we know that researchers struggle to be well
cited. But we will discuss this analogy later. For now we will describe precisely what happens.
There can be a competition only if there is a limited resource to collect. This resource is the
inbound link. And this is the limit: for every link there is only one node that is on the good
side, the target side. As a caricature, we could say that each link has a winner (the target) and
a loser (the source). The goal is to get many inbound links, and because it is a limited
resource, we may think that those who get many inbound links prevent others to get enough.
But as always when it is complex, it is more subtle... Because even if there is a limited amount
of inbound links compared to the amount of links (the half is not it?), anybody can freely add
more links. We can consider that when you publish links to others, you are making an
investment. By adding more links (where you are on the wrong side, the source side) you are
increasing the gap between your inbound links and your outbound links, and thus defining
yourself as a “lower” actor. But you may hope that by doing this, you will get links back as
target, where you are on the good side. What are the conditions to get these links?
The works of A.-L. Barabasi show that the hierarchy we observe may emerge from a behavior
called “preferential attachment”. This happens when new nodes, for example freshly created
websites, link to already strongly connected nodes, such as very famous and big websites. By
doing this, they reinforce the gap between strongly connected nodes and poorly connected
ones. In our case it is impossible to track back the genesis of the websites of our actors. But
we know well some things that look like the preferential attachment. We know that it is a
common behavior for small and badly known actors to put links to main and notorious actors.
But it is a beginner mistake, because these links are not useful (since the target is already
notorious) and because their source actor will not benefit from it, and certainly not by getting
links back. And we also know that institutions usually do not put links to other non-institutional
actors, especially when they are small, while they always require links to their websites
(usually with their logo, as soon as they give money). This behavior is driven by the idea that
a symbolic reward is needed in exchange to something, and the web is very sensitive to the
symbolical transactions.
But there is a problem... Because the model of Barabasi predicts that “the winner takes all”:
the destiny of such a system is to become a hyper-hierarchical network where there is one
node in the center (the winner) that collects all the links from all other nodes. This is clearly
not going to happen, neither to our countries nor to the Web in general. There is no clear
explanation but WebAtlas' previous works suggest that it is due to the strength of
communities. The countries play this role as we can see in the general map: the country level
is so strong that it prevents main national actors from being lowered by foreign and bigger
actors on the same field. Some cultural attraction is at play here, but it is mainly because
there is a need for leading actors in every language. It is well-known that the Web is quite well
separated depending on different languages... With the specific situation of the English
language, unfortunately this study does not include United Kingdom.
The science is a community, and we will not discuss here on the details. The fact is that
researchers usually collaborate. It is clear that when two countries collaborate a lot, like
Finland and Estonia, they have many interactions that justify hypertext links. But this is not
enough to prevent Finland from being in a higher connectivity layer than Estonia, and the main
reason is that Finland has more various sources of links than Estonia. The hierarchy does not
only emerge from the quantitative relations between countries, it mainly emerges because
some countries have many partners while others have few (even if missing countries introduce
a bias on this measure). And we insist on the notion of partnership, because it justifies
reciprocal links. Like we have seen, unidirectional links usually are bottom-up and reinforce the
gap between connectivity layers.
This is an important point on 2nd degree hierarchies. Even if actors try to have a balanced
relation, and even if each pair of actors tries to have a balanced partnership, the 2 nd degree
hierarchy may emerge. It may emerge because even if the relations are balanced, their
distribution tends to be unbalanced (to a certain point). Two actors may have equal
interactions together, if one has more partners than the other, the relation is actually
unbalanced. So if balancing interactions is a good way to regulate 1 st degree hierarchies, it
does not operate on 2nd degree hierarchies. What is more, the “preferential attachment” makes
that the 2nd degree hierarchy tends to transform in a 1 st degree one. As soon as a country has
many partners, it tends to unbalance all of its relations (because it becomes more attractive).
The 2nd degree hierarchy is more diffuse but stronger than the 1 st degree one, so that it is
more difficult to regulate. In our case the 2 nd degree hierarchy was “hiding” behind a network
that seemed well distributed, non-hierarchical, at first glance. We do not know if it is possible
to counterbalance such a hierarchy. But we know that it is self-regulating to a certain point,
elsewhere we would observe a “winner takes all” effect. We suggest that for countries that get
few interactions, it is easier to get some from other poorly connected countries. This prevents
some countries to lose all of their relations. Existing partnerships at the bottom of the
hierarchy prevent the whole pyramid to become a hyper-hierarchical structure where the top
(the “winner”) gets all of the connectivity (“takes all”). In the Web in general it is very clear
that Google, if we call it the potential “winner”, does not prevent other websites to freely
connect, even if a huge mass of users rely on search engines to access resources. Partnership,
particularly in the lower levels, is a good strategy to get inbound links.
Of course the main goal of a partnership is not to get inbound links on your website!
Nevertheless you would expect for it. Why do you expect to be linked? Because it brings more
access to you. And it is not only because a hypertext link is a path for users to browse to your
website. This is just the way access exists on the Web. What you want is access, whatever it
looks like. It is about appearing in the papers of other researchers, if you are one. It is about
being cited as the laboratory of a famous researcher, if you are a laboratory. It is about being
cited as a support of a main event or project, if you are an institution. As soon as an actor
publishes something, it is to be read. And to be read, this actor needs to be accessed. This is
wider than just the Web. But in the Web symbolic transactions appear very well. We track
interactions by analyzing the traces they leave: these traces are symbolic transactions, and
the most obvious shape of a symbolic transaction in the Web is a hypertext link.
Depending on the situation, a symbolic transaction is incredibly cheap or incredibly expensive.
Usually two persons do not make any effort to shake hands: it is almost nothing. But think of
the effort to make the Israeli leader and Palestinian leader shake hands: this symbolical
transaction is huge. We do not mean that the effort is made just for hands shaking. But
without this effort, no symbolical transaction: this is the value of certain symbolic things. Such
things happen with hypertext links. There is a story about one of the most famous Web
designer. He wanted to make a big German car company put a link on its Website to another
big German car company. The effort to make was so incredibly high that he never succeeded.
At another level, these phenomenons happen on the Web of SiS in Europe.
An actor will not put a hypertext link to another without a good reason. Hypertext links exist
only as traces leaved by interactions that worth the effort to put these links. Generally, just
knowing the existence of an actor does not worth the effort to put a link to it. It seems clear to
us that many of the actors of SiS know more other actors than they point to. We propose
several reasons to put a link to another actor:
– It is asked by contract (links to institutions: bottom-up). In France, as soon as you
are supported by the CNRS (National Center for Scientific Research) you are obliged to put the
CNRS logo and usually the link (and the CNRS is the biggest authority in the Web of Science in
France).
– In the “links” page, the target is an important actor (often, meaning “at least as
important as me”: mainly bottom-up). Typically, NGO's will put links to the national
institutions that correspond to their field, even if they are “non-governmental”, because
governmental institutions are main actors.
– Propagating relevant information (event, publication: notorious actors are more
read, so this is mainly bottom-up). Example: a researcher writes a post in his blog about a
publication he finds noticeable, and puts a link to it; and this publication is probably written by
more notorious actors (that is why our researcher watches their publications).
– Showing a partnership (by definition, this is half bottom-up and half top-down).
When a consortium is created for a new project, the actors usually show it in their websites
and then put links to each others.
Most of these types of connectivity produce mainly bottom-up links. They apply well to the
national networks of actors. But have to notice something for the network of countries.
Obviously there is few institutional linking from one country to another, except to European
institutions. We observe a massive connectivity to the International or European level, totally
bottom-up. We cannot track it because it is too large, but we think that the links to the IoE
level are mainly institutional links, because the website of the European Commission is by far
the most cited website of our study. The study of the European Web by RTGI shows that this
domain has a strongly connected core shaped around the website of the European
Commission, thus confirming our analysis.
In the same perspective, we think that links from a country to another are mostly published to
show a partnership or to propagate relevant information. These motivations produce more top-
down links than the others.
We do not have enough elements to determine whether the situation is this one. But with it we
can propose an important hypothesis. If the country to country links in the SiS field are due to
strong interactions such as partnership, and this might be due to the importance of
universities, then a lack of links means a lack of strong interactions. The point is that we
observe a lack of links in lower countries (in the hierarchy) like Armenia and Montenegro. Of
course the hierarchy is not complete. But at least it would mean that there is a noticeable lack
of strong interactions between the upper and the lower countries that we have in our set.
Maybe it is not a mystery that some small countries do not have many partnerships with their
European neighbors. But we may bring a context to that, and propose some hypothesis about
the way things happen. For now we propose the hypothesis that in the SiS field, the main
interactions between the countries that we have are mainly due to scientific activity, and
particularly through the activity of universities. Then the weak connections observed to certain
countries would be the sign of the lack of partnerships with these countries. This point should
then be put in perspective with the fact that the more attractive countries tend to gather more
and more resources, and that partnership is an important mean to counterbalance this
asymmetry. Thus promoting strong interactions with countries that lack the most could be a
strategic action to prevent the SiS field in Europe from being dangerously concentrated.
This hypothesis is a proposal to read this work of mapping, and it will be one of our axes to
analyze different aspects of the data we gathered. Actually this hypothesis simply reflects that
besides the countries have hierarchical relations, we find some countries very isolated. That is
the way we propose to interpret the hierarchy of connectivity between countries is the
following:
 The network of national actors in each country is stronger than any transnational
network. On one hand it protects the situation of local leading actors, by preventing
stronger foreign actors from draining their resources. On the other hand a hierarchy
appears between countries where few emerge as leaders.
 The access provided by inbound links is a limited resource, and it is shared on the basis of
certain interactions. We think that countries compete for this resource (access) while
cooperating to counterbalance the tendency of some countries to drain too much of this
resource. It might reflect a network of interactions that is wider than the Web.
 We think that some countries might be excluded from this game, because nobody
cooperate with them. We have not enough countries in this study to be more precise, but
the network of interactions seems noticeably unbalanced to us, especially concerning
weakly interacting countries.
 We suggest that the best way to counterbalance this might be an action on universities,
because they play the main role between countries in the SiS field. We suggest too
that institutions might interact essentially with their nation and the international or
European level, but not directly with other countries.
This is not a conclusion but an hypothesis that should be tested in other (further) works and
that will help us to put some results on perspective.
Future steps
Propositions for a Web Information System dedicated to SIS (or SIS²)

EuroSiS task 3.2 (partner search) conclusions could be used to design an information system
for a permanent and global process of description. This Web Information System (W.I.S) could
involve the 27 countries of European Community operating through a large distributed network
of observers. Such a system dedicated to SIS-mapping could be also useful for “science
policy”, decision making and political survey. This knowledge-based system of Web resources
extracted in European area could be a tool for the coordination with national activities and the
European Commission.
Both at national and European level, this information system produces indications on thematic
changes and emerging trends. Network mapping tools can display some temporal patterns of
evolution in big sets of data as Web documents (time tracking). How information spreads in
EURO-SIS network? How evolves distances and position between countries, organizations,
actors through time and Web data? Such systems for tracking time series in data exist
nowadays and include visualizations and statistical indicators.
Activated over 27 countries, observers could be involved in regular activities of knowledge
sharing and Web data mapping. Such a distributed network of observers could help in
identification of common resources, common trends, common interests and common
challenges at the most appropriate level, be it national or European. To achieve this goal, the
Baagz software has to be re-designed to become a fully personal workspace, including
functionalities for online corpus monitoring, resources tagging and up-load/downloading space
for NCP’s networking. The system must also be used for national reports.
Finally, as the SINAPSE system (or in association with), this system could publicly display URL
and tagged resources through an on-line portal. This king SIS-search engine could encourage
public engagement ('downstream' and 'upstream') in scientific research, in particular, through
involvement of civil society in debating and shaping the research agenda. An experiment must
be carried out in this field to build a democratic knowledge based society and to address
societal issues in European research.
Obviously, in a distributed network for SIS-mapping and tracking, personal workspaces must
be extended to all kinds of documents: Web data but also dynamic information (such as
“news”, RSS fluxes…) and “static” ones as scientific publications, national databases, even
multi-formats documents (pdf, ppt, txt, odt…). This technical extension could permit to
encompass all fields of SIS policies and civil society activities. This process of aggregation of
Web Information, specific documents and dynamics news should be one the key-feature of an
Information System for strategic watch and decision making in the SIS field.
The tagging of Web resources made by observers represents a high quality work and should be
improved by continuous learning and common sharing. This regular task could lead to a
general index of European actors in SIS field, from political ones to industrial, scientific and
civil leaders involved in science and technology management. A continuous distributed work
associated to centralized processes for data analysis and mapping could be considered as the
key-feature of an Sustainable Information System in which technological development and
evolution is governed by the needs of a community as a whole and not only by computational
functionalities. In other words, Euro-SIS mapping and experience could lead to a future
Sustainable Information System for Science In Society, a SIS².
Bibliography
Bibliography
Theory, conceptual framework

A.-L. BARABASI Linked - the new science of networks, new ed. 2005.
S. JONHSON Emergence: the connected lives of ants, brains, cities, and software, 2002.
Graph Theory
D. WATTS Six degrees - the science of a connected age, 2004.
S. STROGATZ - sync: the emerging science of spontaneous order, 2004.
M. NEWMAN - the structure and dynamics of networks, 2003.
Web-Mining
S. CHAKRABARTI Mining the web, 2002.
J. KLEINBERG - Algorithm design, 2006.
InfoViz
B. SHNEIDERMAN - readings in information visualization: using vision to think, 1999.
Actor-Network Theory
B. LATOUR, Reassembling the Social: An Introduction to Actor-network-theory, Oxford
University Press, 2005, ISBN 0199256047, 9780199256044
Theory, conceptual framework

A.-L. BARABASI Linked - the new science of networks, new ed. 2005.
S. JONHSON - Emergence: the connected lives of ants, brains, cities, and software, 2002.
Graph Theory
D. WATTS Six degrees - the science of a connected age, 2004.
S. STROGATZ - Sync: the emerging science of spontaneous order, 2004.
M. NEWMAN - The structure and dynamics of networks, 2003.
Web-Mining
S. CHAKRABARTI Mining the web, 2002.
J. KLEINBERG - Algorithm design, 2006.
Web Mapping
N. ANDRIENKO, G. ANDRIENKO, Exploratory analysis of spatial and temporal data: a
systematic approach, Birkhäuser, 2006, ISBN 3540259945, 9783540259947
InfoViz
B. SHNEIDERMAN - Readings in information visualization: using vision to think, 1999.
Glossary
Actor: In our study, an organization involved in Science In Society. Because we use the web
to extract information, an actor is always represented by a website.
Connectivity: Amount of hypertext links. If a website has a high connectivity, it has many
hypertext links. The internal connectivity of a set of websites represents the amount of links
between these websites, while the external connectivity represents the amount of links
between websites inside and outside the set.
Data mining: The process of extracting hidden patterns from data. See Web Mining.
Degree: the degree of a node is the count of its links. We distinguish indegree (how many
links point the node) and outdegree (how many links “go out” of the node).
Edge: Connects two nodes in a graph. In our study it represents hypertext links.
Graph: Nodes connected by edges. A mathematical structure that we use to project data
extracted from the web.
InfoViz: Also called “Information graphics” or “infographics”. Visual representations of

information, data or knowledge. We think that this discipline is better described as
“spatialization of information”. In this study we stick to the “Exploratory Data Analysis”
philosophy: the way of making statistics is hypothesis generation oriented, rather than
hypothesis validation oriented.
Link: Hypertext link that connects two web pages. By extension we consider most of the time
links between websites (websites are linked if there exist links between their web pages). See
“Edge”.
Network Science: Network science is a new and emerging scientific discipline that examines
the interconnections among diverse physical, informational, biological, cognitive, and social
networks. This field of science seeks to discover common principles, algorithms and tools that
govern network behavior. The US. National Research Council defines Network Science as “the
study of network representations of physical, biological, and social phenomena leading to
predictive models of these phenomena.”
Node: The fundamental element of a graph. Nodes are connected by edges. In most of our
graphs, nodes represent websites that represent actors (and edges represent hypertext links).
Scale-free network: Network whose degree distribution follows a power law, at least
asymptotically. The web is considered as a scale-free network, and it's the case of the
subnetworks studied here. See “Small world network”.
Site: Website. A collection of related web pages, images, videos or other digital assets that
are addressed with a common domain name or IP address in an Internet Protocol-based
network. There is no strict definition of a website. But Internet users intuitively know when
they go out of a website and come to another. We relied here on the source page identified by
NCPs. This means that a website isn't always defined by its domain.
Small-world networking: In mathematics, physics and sociology a small-world network is a

type of mathematical graph in which most nodes are not neighbors of one another, but most
nodes can be reached from every other by a small number of hops or steps. A small world
network, where nodes represent people and edges connect people that know each other,
captures the small world phenomenon of strangers being linked by a mutual acquaintance.
Such as the networks studied here. See “Scale-free Networks”.
Web: Abbreviation of World Wide Web. A system of interlinked hypertext documents accessed
via the Internet (with a Browser). In this study we don't explore the Internet but only the web,
meaning the part of Internet that is accessible to a Brower (not the emails, the voice over IP
etc.)
Web Mapping: The process of designing, implementing, generating and delivering maps on
the World Wide Web. While web mapping primarily deals with technological issues, web
cartography additionally studies theoretic aspects: the use of web maps, the evaluation and
optimization of techniques and workflows, the usability of web maps, social aspects, and more
Web Mining: The application of data mining techniques to discover patterns from the Web.
According to analysis targets, web mining can be divided into three different types, which are
Web usage mining, Web content mining and Web structure mining. In this project we used
essentially Web structure mining.
ANNEX 1
National Contact points’ CV
Armenia Mr. TIGRAN ARZUMANYAN: is the Head of International S&T

Tigran Programmes Unit of the National Academy of Sciences of Armenia (NAS
Arzumanyan RA), and Vice-President of the Center of Ideas and Technologies NGO, and
officially nominated FP7 INCO NCP in Armenia.
Since 2003 he has been co-ordinating the activities of EU FP6 National
Information Point established at the NAS RA to promote and facilitate
participation of Armenian researchers in European research programmes,
particularly, FP6/7. In this endeavour he participated in many
trainings/information days organized in various EU/EECA countries
dedicated for FP6/FP7 NCPs.
He was Fellow of NATO Science Policy Fellowship Programme in 2001 and
2004, to carry out research and comparative studies on S&T and
Innovation policy at the Science Department of Calouste Gulbenkian
Foundation, Lisbon, Portugal, and Innovation Research Centre, IKU,
Budapest Corvinus University, Hungary, respectively. During last years he
has participated and contributed to several S&T and innovation policy
related international events organized in Romania, Bulgaria, Portugal,
South Africa, Russia, and USA.
Belgium Laurent Ghys has a PhD degree in Chemistry awarded by the Free
Laurent Ghys University of Brussels (ULB) in 2002. He is working at the Scientific and
Technical Information Service (STIS) as a scientific collaborator since
2002. He is NCP for the Belgian Federal Authority for the specific Research
Infrastructure programme of the 7th Framework Programme and
occasionally supports the work of the other NCP.
Bulgaria Dr. Ivaylo Dimitrov has a Ph.D. in Philosophy and Master degree in
Ivo Dimitrov Public Relations. He is Assistant
Professor of Epistemology at the Institute for Philosophical Research of the
Bulgarian Academy of Sciences.
Founder and chief editor of two websites for science popularization
(Democrit.com and
Green.Democrit.com). Due to his research interests and science
communication activities Dr. Dimitrov is
elected as SiS NCP in FP7.
Czech republic Mr. Michal Pacvoň works for the National Information Centre for
Michal Pacvon European Research, which is a part of the
Technology Centre AS CR. He is graduated on Faculty of philosophy on
Charles University in Prague. He
was involved in research on Centre for Phenomenological Studies. He
published some articles in the literary
and film theory revues. He is also NCP for Socio-economic sciences and
humanities and for INCO.
Estonia Terje Tuisk works for Archimedes Foundation since 1997. She has been
Terje Tuisk involved in NCP networks since
2000 - NCP for IHP program in FP 5, Science and Society NCP in FP6 and
Science in Society NCP in FP7.
She is a head of Science Popularization Unit in Foundation Archimedes -
among other duties the unit works
with Young People and Science topics as well as organizing the Science
Communication Award contest in
Estonia.
Finland Dr. Reetta Kettunen Ms., Secretary General for Committee for Public
Reetta Kettunen Information.
The Committee for Public Information is an expert body attached to the
Ministry of Education. It follows progress in research, arts and technology
and the development of knowledge in Finland and abroad. Each year, the
Committee awards grants, makes a shortlist of candidates for the State
Award for Public Information, and makes a list of domestic quality non-
fiction entitling libraries to acquisition support. In addition, it gives its
opinion to the Ministry concerning matters within its mandate and makes
proposals and takes initiative for different ways to promote the
dissemination of knowledge.
The Committee also has many responsibilities in the Science in Society
work on national level in Finland
(http://www.minedu.fi/export/sites/default/OPM/Julkaisut/2004/liitteet/op
m_213_tr28.pdf?lang=fi).
The committee coordinates the information website on R&D in Finland
www.research.fi.
The office of Committee is attached to the Federation of Finnish Learned
Societies. (http://www.tjnk.fi/en/ ).
Dr. Kettunen has an academic background with plant physiology and
molecular biology as her area of scientific expertise.
She has a strong experience in science communication and science policy.
Her previous working experience includes:
- Editor-in-chief in the field of popular science books, Art House Ltd
- Scientific secretary, Viikki Research Group Organization in Molecular
Biosciences, University of Helsinki
- Programme manager, University of Helsinki (for research programmes
funded by the Academy of
Finland).
In ERA-context, Dr. Kettunen has been working for two ERA-NETs (for
ERA-SAGE as a task leader, for ERAPG as a ncp and as a science
representative). She is also familiar with technology platforms. She is the
FP7 SiS NCP (II) in Finland.
France Mrs. Sophie Tocreau has a master degree in scientific museology. After a
Martine Roussel first work experience at the
Sophie Tocreau “Palais de la découverte” (geology department), she joined the Ministry
Delegate Research in 2000 where
she was given the responsibility of coordinating the national science
festival (2000-2005). Involved in the “Science festivals” European
network, Sophie Tocreau she has followed up the Science in society
European program, along with the official program committee
representative. In this framework, she has started mobilizing « Science in
society » potential actors. Now, she also carries out this task in the
framework of the FP7 Science in society National contact point.
Mrs. Martine Roussel has joined the ministry Delegate of Higher
Education and Research since March 2007 where she assumes the task of
NCP for the Science in society capacities program along with Sophie
Tocreau. She worked previously for the French National Centre for
Scientific Research (CNRS) as a communication officer. She was involved
in FP5 IST projects while working for the Teleport Sachsen Anhalt (DE)
Brussels office. She studied politics in Rennes (Fr) and at the university of
Exeter (UK). She graduated from the college of Europe in 2001.
Hungary Mrs. Ágnes Hegyváriné Nagy, NCP for the FP7 programmes “Science in
Agnes Society”, “Ideas” and the topics related to “Women in Science” has been
Hegyváriné working in the International Department of the National Office for Research and
Nagy Technology since 1999. She worked as an NCP for the “INCO” programme, and
as a contact point for legal and financial issues and an NCP coordinator during
the 5th Framework Programme. She was responsible for several Hungarian and
EU-funded training programmes for NCPs and Hungarian liaison offices. In
2002 she received an acknowledgement for her work from the Under-secretary
of the Ministry of Education.
She graduated at the University of Debrecen in 1996 as a teacher of
Mathematics and English, later she obtained a postgraduate degree in
European Studies at the University of Economics (Corvinus) in Budapest.
Italy Mrs. Mara Gualandi is a National Contact Point in Italy for the EU
Mara Galandi Programme "FP7-Science-in-Society".
Katia Insogna Mrs. Gualandi has been working for seven years, from 2000, in APRE with
several tasks:
-Responsible editor for the monthly electronic magazine "APRE news"
- Responsible for the relations with CORDIS and IPR Helpdesk
- Responsible for training courses organized by APRE with the presentation
of various subjects as
"Communicate the research" or "How to submit a successful proposal in
the Framework Programme".
Montenegro Mrs. Tamara Tovjanin is a senior advisor for scientific and technical
Tamara Tovjanin cooperation in charge, among other duties, of cooperation with European
Union. Her job is to prepare proposals for programs of international S&T
cooperation and participate in their approval; coordinate and stimulate
their implementation in our country through regular contacts with projects'
holders; carry out coordination with the appropriate departments of the
national and local bodies in the scope of her duties; propose measures for
improvement of S&T cooperation and follows their implementation, provide
expert opinions on draft programs, agreements, contracts and other
documents in the field of S&T, etc. She was a project manager on the
Montenegrin side of the ERA WESTBALKAN project and is in charge of the
project management of the projects ERAWESTBALKAN+ and EU-Balkan-
FABNET. She was FP6 NCP and is FP7 NCP for all of the above-mentioned
priorities.
Educational background: BA in Oriental Philology and MA in European
Studies.
Poland Mrs. Małgorzata Krótki holds a Master’s degree in Biotechnology. She also
Malgorzata studied Managing EU Projects at the Warsaw Agricultural University,
Krotki Faculty of Economics and Agriculture.
After a first work experience at the Plant Breeding and Acclimatization
Institute, Department of Foreign Cooperation (EU Centre of Excellence
project „Crop Improvement Centre for Sustainable Agriculture”), she joined
Instytut Podstawowych Problemow Techniki Polskiej Akademii Nauk (IPPT
PAN) in 2005 where she was given the responsibility of supporting role for
the FOOD NCP (2005-2007).
Since January 2007 she has undertaken NCP positions for Socio-economic
Sciences and the Humanities as well as Science in Society. She is a
member of SSH and SiS Programme Committees. She is involved in
several FP6 projects.
Portugal Mr. Mario Vilar has a Degree in European Studies from Universidade
Mario Vilar Moderna de Lisboa. In July 2004, he
joined GRICES as Technical Staff for the Helpdesk of the Idealist34. He has
undertaken several NCP positions, namely Energy and EURATOM NCP
(since January 2005), IST and Infrastructure NCP (since
January 2006), Infrastructure, SiS and SSH NCP (since January 2007).

Finalreporteurosis3 1

Uploaded by

Copyright:

Available Formats

Finalreporteurosis3 1

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Finalreporteurosis3 1

Uploaded by

Copyright:

Available Formats

EUROSIS

« Webmapping of science and society actors in Europe »

MAPPING THE WEB OF SIS IN 12 EUROPEAN COUNTRIES...............................................................................6

Background and objectives

Methodology and results

 NCPs would provide various data on national SiS stakeholders

While defining the methodology, a series of possible biases appeared:

The results the project are of two kinds:

What is our approach?

Web mapping as a Sociology

Web mapping as Network Sciences

Differences between NCPs

It is only a “Web picture”

The set of countries

NCPs practicing the Web exploration

Layers of the web:

- Medium layer: websites

- Deep web: databases...

How works a distributed network of experts?

Building a common description of the resources

General Map: actors aggregate by country (colors by country)

Connectivity structure of the types of actors

The three connectivity layers: pattern for the types of actors

We found three connectivity layers. It is important to understand that an actor is in a layer

Interactions between types of actors

– CORE : Academic Science (Universities and Secondary Schools, Research Centers,

Research.be is the main portal

NGO-CSOs and Portals about Sustainable development

NGO-CSOs cluster about philosophy (related to science)

NGO-CSOs cluster about women

The Bulgarian Universities subnetwork is strong (actors in blue)

The hierarchical network of Research Centers draws a “star” around BAS

The Czech Republic SiS Network is inconsistent:

The Czech NGO-CSO subnetwork

Main authorities (media): Eesti Päevaleht and Postimees

Second hub, a media above the bridge

Bridge (between the two components): NGO-CSOs about Nature.

Scheme of the structure of the Estonian SiS network

The most represented themes is “Science Education” (according to the importance of

Most of the French museums are connected to no other actor

The remaining connected museums draw a small subnetwork

The NGO-CSOs (green) component

The most represented theme is “Communicating Science”.

Structure of the Hungary SiS Network: two interconnected parts

The Italy SiS network, organized around two hubs

The Telecom and Engineering component in Montenegro SiS Network

The Faculties component

The main hub is a NGO-CSO

The most represented theme is “Communicating Science”.

Hierarchy emerging from countries' relations

Where do websites aggregate?

 Small-World in the countries

 The signature of complexity

 Observing the hierarchy

Inbound and outbound Poland

 Interpretation of the hierarchical relations

Propositions for a Web Information System dedicated to SIS (or SIS²)

Theory, conceptual framework

Theory, conceptual framework