Two Visions of the Web: from Globality to Localities
Fabien PFAENDER Mathieu JACOMY Guilhem FOUETILLOU
COSTECH Laboratory
University of Technology of Compiegne
France
fabien.pfaender@utc.fr mathieu.jacomy@gmail.com guilhem.fouetillou@utc.fr
Abstract
This paper presents two different perspectives of the
web : a global one that corresponds to the classical ap-
proach of search engines and a the local one that we
propose as an alternative approach. The search en-
gines perform their indexation operation on the whole
web in an automatic way and display their results ac-
cording to it by proposing a perfectible visualization.
We will review the usability of these visualizations
while examining the way search engines build their
hierarchies. That leads us to reconsider the notion of
context and the way models of the web influence our
vision of it to finally propose a new model strongly
related to its perception through alternative visualiza-
tions.
Keywords— Webming, web structure, information
visualization, usability
1. Introduction
We propose to discuss the approach of the web by
search engines considering the opportunity to build
maps of the web into observing its shapes. Both ap-
proaches are based on a common understanding of
the web as an open, heterogeneous and large-scaled
network that can be worthily projected as a graph.
We will first quickly present the way algorithms of
search engines utilize known shapes of the web, and
the way they represent it with its limits. Then, we
will stand up for the idea that the notion of locality
is essential even to establish a hierarchy of resources.
Finally, we will get onto the notion of context on the
web considering the major role of localities and car-
tographic visualizations with an experimental map.
2.Apprehending the shapes of the web :
the case of search engines
This part confronts the search engines, which ran-
king produce lists, and the possible shapes of the web
that maps can visualize. Two complementary aspects
of the web as a graph are discussed : algorithms and
visualizations.
2.1. Principles of web graphs calculation
by search engines
Initially the observable shapes of the web allowed
to conceive the now classical algorithms on which is
based a large part of the search engine technology. Ho-
wever, the evolution of this technology was provided
by the observation and the analysis of the behavior
of internet users. Knowing that the projection as a
graph is a foundation of classical approach of the web,
we point divergences between exploiting the shapes of
the web and producing empirical hierarchies.
2.1.1.The first ranking algorithms : measure
the support of web navigation
First ranking algorithms, aiming to return the most
pertinent resources, are developed into exploiting the
properties of the web as a graph. The graphs can be
easily handled as matrix, and search engines make
the most of this way of coding data[1]. Kleinberg’s
HITS algorithm exploits the bipartite structures of
web graphs to build a measure of authority. The mea-
sure of authority is a measure of pertinence only if bi-
partite structures are global properties of the web. For
Kleinberg ”the creator of page p, by including a link
to page q, has in some measure conferred authority
on q”. The success of Kleinberg’s algorithm and his
derived give him reason. However, the social aspect
of the hypertext link is not completely analyzed to-
day, and that’s why it is difficult to evaluate an algo-
rithm thanks to sociological arguments. On the other
hand, it is possible to observe the web into bringing
out invariant properties, like bipartite properties, and
building appropriate algorithms while analyzing these
invariant properties in the field of social sciences.
In 1998 too, Sergey Brin and Larry Page pu-
blish the paper that describes their search engine
Google[2]. Google uses the PageRank algorithm to
build a hierarchy of resources. Once again, the intui-
tive justification proposed by authors is based on the
forms of the web. We insist on the fact that besides
argumentation, the algorithm is not conceptualized
by the internet users, but by the shapes of the web
supporting the navigation. Today, the social aspect of
the web is exploited without theory, and like in 1998,
social tools appear on the web thanks to the only in-
tuition. The algorithms of search engines are based
on shapes of the web that are possibly social, and not
on the analysis of internet practice.
2.1.2.The search engines released themselves
from the shapes of the web
The massive use of search engines by people at large
changed the situation. For a search engine, the effi-
ciency of results returning is the main issue : algo-
rithms don’t need to refer to shapes of the web. Of
course, the model of random surfer is widely used. But
search engines, to be the most efficient possible, use
different optimizations that make algorithms difficult