Survey of requirements of digital research services in libraries.

goldbashedAI and Robotics

Nov 15, 2013 (3 years and 8 months ago)


Survey of requirements of digital research services in libraries

Dominique Stutzmann

First of all, I wish to thank

the French Institute and Anne
Elisabeth Buxtdorf for inviting me

In this session about “the library of the future”, t
very short
(not more than 15
minutes) will try
speak aloud
the librarians’ questions and fears, but
to give us a chance
of hearing
the users’ voices. If we ask the following questions “Will the library be a lean
information machine?” or “cultural space?”, o
r even “will the library as a space vanish?” and
if we believe that all the answers are given by the users’ behavior on the internet

account the specific relationship that the libraries can have and maintain with their
users and what the
libraries can do in order to keep a role in the new information
environment, then it will lead us to false conclusions. That’s true that everyone can see on
almost every website the major changes which came with the once so
called “web2.0”:
, crowdsourcing, folksonomies on the one hand, that is user acting as
publishers and being proactive in the field of information literacy, and
on the other hand we
a mistrust against “institutional experts”, that has been clearly demonstrated by severa
studies. By saying that, it may seem that all conditions are given for the libraries to disappear:
the users are acting as librarians and the experts are not valued any longer, and we don’t see
why the librarians as experts in information literacy should

. Nevertheless
there is a
good reason for being confident: the users
do not w
ant the libraries to disappear.

In the following minutes, I will present you two users’ studies organized by the French
National Library, the first on behalf of The Europ
ean Library, the second one done for getting
a better knowledge of our users.

We may
sum up the results even though th
ere may be some
show the ambivalent relationship between libraries and their

Part 1

The first

survey deals with the semantic search. As you may know, several search engines on
the web try to give a more accurate access to the different resources and to the different kind
of resources (Hakia, Sensebot, Cognition, Powerset, Kosmix…; Sinequa, Exalead
, Kartoo). In
the domain of the printed book, a well
known search engine already uses data mining tools in
order to improve the searching services, especially by giving back the meaning of the words
to the text, what we can call “semantic”. As a result, ci
ty names may appear on a map
represent the
geographical meaning


will some day become searchable through a
graphic interface. Some libraries already give access to geography
related resources. Two
examples: one map giving access to novel accordin
g to the city where the story takes place;
one second map giving access to travel and tourist guides

in the holdings of a little public
library in Southern France
. This map makes it possible to perform the search easily whereas
the subject index of a libra
ry could be too precise for the user.

The next step will probably be to provide a graphic interface to the navigation by subject.
Many attempts have already been made, but none is totally satisfying (Kartoo

: map as a
disamibuisation tool
, ThinkPedia

ation tool
, VxInsight
and the interactive
map based on the social indexing on LastFm

that show composers and musician through the
cooccurrences in the tags

of search engine,
the TELplus project of The European Library has one part

dealing with semantic search engine and a prototype has been developed under the direction
of the French National Library.

This prototype offers not only a multilingual search
interface, but also multilingual search services

you can type in “c
heval”, “travel” and
“berg” and find results with “horse”, “travel” and
“mountain” or “Pferd”, “Reise” a
nd “Berg”.
There is also a way of navigating in the relevant subject headings in the facets on the left side.

In this semantic prototype, two di
fferent keywords extractions are present: the first
one, above, is purely statistic and linguistic; the second one, below, is the result of matching
the full text with the structured subject headings authority data from the library domain. Both
are very re
presentative of what each technology can do: the first mixes up singular and plural,
show as relevant the word “whom”, because it is frequent in the text and more in 1889 as it is
nowadays. If you give a closer look, you will see that the second one is muc
h more accurate
and has got real subject headings which give us a better and more relevant account of the

The libraries are not only a major content
most of
books and resources
being digitized
are kept in libraries)
, this
prototype demonstrates that the libraries
play a role in the information environment

as they are one of the few cross
domain providers
that maintain not only data, but also structured metadata.

We had a user survey evaluating this prototype with

2 groups of 15 people during a session of
an hour. And the results were quite confusing. For both questions “Which functionality is the
most useful?” and “Which is the less useful?”, the “keywords” were the most frequent
answer. And the most frequent sugg
estion made by the user was also to have clickable
keywords for starting a new search.

If we consider these seemingly contradictory results, we can draw the following conclusion:
the users were not satisfied with what the prototype is offering, but their
main interest is
keyword or concept search and they expect that the librarians can offer better search
functionalities in this domain than an average search engine.

[slide] Part 2

The second survey I want to present briefly was made for the
French Nationa
l Library

We wanted to
address the question of
made of
large digital libraries


Aims where to get a better knowledge of our users and of what they want
. What
does the public

What sort of online communi
ties are using the digital
resources? Are there “invisible” or “self
ignorant” communities we could bring together with
networking and sharing services?

[The question we didn’t ask is: what is the benefit for the library and for the users?]

The BnF had the

survey done by a consulting firm who offered to combine


methods in three successive phases. The results of each phase were used to create new
questions for the following phase. First, we put a questionnaire online

to get a quantitative
w and to get to know the public and its online habits. In three weeks, more than 1,200
forms were filled out. In addition to this online poll,

the second
phase included forty semi
interviews with users in France and abroad. Though forty may not
seem like very
many, it is a satisfactory number for this sort of survey.
In the 3

phase, there were 3


First of all, the survey gave us more accurate information about the users than we had
with online tools such as Xiti or Google
Analytics. The gender and age analysis shows that
the typical digital library user is male and urban.


More precisely our users can be broken into two major groups: male retirees or
working professionals over forty

on the one hand
; and students of

both sexes

on the other
. Women in all categories are younger than men, and there are more female students than
male students.

All in al
l, the potential for discovery
in a large digital library has the consequence that 75% to
90% of all users don’t kn
ow what they might

find in the digital library, even if we can split
the users into two separate and equal groups: those who know what they are searching for,
and those who are visiting and exploring.


People with sharing and online publishing hab
its represent a quarter of all users, which was
quite a lot at the
of 200

in France
, before Twitter and Facebook became well
known services

But we don’t really find the expected willingness to share in the digital


All users e
very positive feelings about the digital library. Even when
they were comparing with other digital repositories, they think a library’s digital resources are
superior to those of a non
. That’s why they also all agree t
hat the
library should not be “spoiled” by foolish comments: if they use forums and discussion lists

or collaborative websites, or
use the rating functionalities on, the users declare
they don’t want to see this happening within the digital libr
ary, because they would rather see
the librarians rating the works or enhancing the reference works.

It may seem paradoxical, but the same people who say that they don’t want to see what other
people do, are willing to share what they do themselves.

do not feel this is contradictory
It is
a consequence of the positive perception of the digital library

users are willing to help
In a certain manner, users express the desire not to collaborate together, but to
collaborate with the library,
thus benefiting all of its users.

The distinction between people wanting or willing to share with the digital library and
wanting or willing to share with one another is not a matter of age or gender or frequency of
use. There is a
separation be
tween research professionals and amateurs. The research
professionals, faculty members and students, have their own sharing tools within the
university system: not only do

they mostly
not wish for their comments to become available
to anyone, which means t
hat there should be restricted access to some comments in the
library, they also cannot set a limit on their activities

comment only on the online
resources of one digital library. For us librarians, this is a significant restriction to our hope of
proving the access, service and quality of our data with social tagging and comments
sharing, as the most qualified people are not likely to share within our digital rooms.


Also again the debate:
social tagging is not only about being fun and user
friendly, it is
a way to give better access to poorly cataloged resources. And in doing that, we are putting
the service and its

use back in the spotlight, from where it shouldn’t have been removed.
is not necessarily given as such and
could b
e delivered
in a more nuanced way
semantic search engines.

retrieve information from the
text and
this information with authority data (subject headings, classification, and so on) on the one
hand and with social rating and

indexing on the other hand. By this mean,
users don’t have the
feeling the librarians are leaving

them to fend for themselves.

At the same time, libraries have to develop a “no barrier” approach and make their content
searchable and usable
, exactly the s
ame way

as the others’ content would be used from
various websites and services. They
become more visible on the web and their content
be found by users who do not look for them.

[conclusion slide]
I conclusion, I’d say that
asking if

libraries as a space will vanish”
can only be a self
realizing prophecy. The missions of libraries don’t change and the library
of the future should be pretty much the same as it is now. The industries need
structured data
ies, in order to imple
ment services that will benefit to the end users. The
domain knowledge of librarians
and the information literacy, the possibility of accessing and
searching licensed resources or to search the relevant databases or open archives repositories.
are new ways of fulfilling the needs of the users and that is where libraries have to be
the most creative so that their role in long term preservation and access and their role as
resources centers remain clear to the communities they serve.