Content uploaded by Martin Ebner
Author content
All content in this area was uploaded by Martin Ebner on Jan 10, 2021
Content may be subject to copyright.
Experiences Based on a Major
Information Server
Delilovic, Namik; Ebner, Martin; Maurer, Hermann; and Zaka, Bilal
Abstract: In this paper we describe lessons we
have learnt from building up a webserver with
reliable information over a period of ten years.
We show how we have tried to overcome often
encountered weaknesses and thus encourage
the community to follow our suggestions. We
also point out that much research and
development work remains to be carried out.
Index Terms: Intelligent Web server; digital
library; flexibility in Web servers; Austria-Forum
1. INTRODUCTION
A team involving the authors has been involved
in developing a substantial “quality-controlled
information server” called Austria-Forum [1], [2],
[8], [33] concerning information of interest to
Austrians since 2009. The aim of this paper is
to correct a misunderstanding: Despite a large
number of research results and
recommendations for practical use, austria-
forum.org, has been misunderstood by many as
just “another collection of partially interesting
digital information”.
This is a serious mistake that needs to be
corrected. Of course, the server is offering
interesting information on a huge variety of
topics, and has gained some respect for this.
But the collection of this material has a second
reason: Large quantities of information and
users are needed to conduct research on how
to handle huge amounts of data and users. One
approach is to study how to represent reliable
information and how to make access to it useful
and easy for users. Another one is how to let
users discuss important questions. A third one
is the problem how to make sure that the
material is not just attractive to one age group
but to young and old, from school student to
persons already having left active life for some
time. And a fourth and may be most important
in our times: to make sure it allows to
differentiate between fake news and real news,
between half-truth and objective reality. Much
substantial research in those areas seems to
have sometimes gone unnoticed. It is the aim of
this paper to report on some of the research and
development achievements that are influencing
the Web, and are needed so that the Web is not
just a new way into chaos [7].
2. NUMBERS
The Austria-Forum has some 3 million different
(mainly German speaking) users at the time of
writing (October 2019), over 1.1 million media
objects and close to six million page views per
year. It has been successful in showing how to
improve critical areas of Web presentations and
is attempting to be a kind of role-model for
similar attempts by further research and
development. Here is an overview of the
structure of the rest of the paper:
Section 3: How to avoid the presentation of
fake information by using a repository of
carefully documented reliable information.
Section 4: What is the role of digitized books?
Section 5: How can one assure meaningful
answers from search-engines.
Section 6: How can the software help to provide
serious communication between well-defined
groups and avoid outpouring of trivialities and
fake information as is typical for some social
media?
Section 7: How to fight fake news.
Section 8: How to reach all age groups.
Section 9: Conclusion.
All ideas and developments presented in the
sections that follow are by-products of research
and development of a group supported by a
number of universities and sponsors. The
development is supported by numerous
institutions from within Austria and other
countries and a board of some 200 honorary
editors and researchers.
Disclaimer: Many of the results have been
published in research papers, theses, keynote
presentations and the like before. We cannot
discuss all issues involved and are leaving
some to be studied in detail using the
references provided.
3. QUALITY CONTROL OF CONTRIBUTIONS IN
AUSTRIA-FORUM
One important way to assure high quality is to
make sure that contributions (that are in
searches presented, are written by well-known
persons, usually with a CV showing their
competence. This is also much supported by
big efforts like [33].
Users can register (even with an arbitrary pen-
name) in a way so that only the system knows
the E-Mail address, but still they can contact
authors without revealing their identity or E-
Mail. The system allows authors to answer,
without needing to know the identity of the user
who contacted them.
There is a second alternative to provide
feedback: Each Web-page in Austria-Forum
has a green “feed-back button” [13] that allows
to send, without any registration, messages to
the administrator, who can act on them,
including forwarding them to one of the editors,
if this seems appropriate.
These two facts, i.e. that authors commit
themselves to what they write rather than to
hide in an anonymous group and that two ways
of communication for questions, comments etc.
are provided, assure a much higher standard of
contributions than it is found in anonymous
systems where users have no easy way to react
and do not know in which direction an author or
author collective is trying to color information
according to own beliefs, as is sometimes
argued about Wikipedia [34}. A second
argument that reduces the value of Wikipedia is
the fact that different language versions present
in quite a few cases very opposing views.
However, in addition to above, Austria-Forum
provides another important way to assure
quality: It contains thousands of online readable
digitized books [11], [13]. Links from any Web-
page to any point in a book (and conversely,
and from a book page also to another book
page) are possible. Hence, an author writing
about some matter can make a link to a section
of a book dealing with that matter, thus adding
to his credibility the credibility of the author and
publisher of the linked book. Even if authors (or
some member of the Austria-Forum team) have
not made links to some book sections (a
process to be supported by providing
suggestions for links automatically) users can
consult books available in Austria-Forum to
check issues themselves.
4. WHAT IS THE ROLE OF DIGITIZED BOOKS IN
AUSTRIA-FORUM
As mentioned in the previous section books can
be used by linking from a contribution to a
relevant spot in a book, or users can check up
on statements made on some Web-pages using
the books provided.
Since almost all books are full-text searchable,
this is often easy. Note that books are readable
online (but usually not downloadable). Due to
the fact that links from Web-pages lead users to
“stumble” across books (of which they might not
even have been aware of) many publishers see
the incorporation of their books into the digitized
library of Austria-Forum not as a competition to
selling the books in print or as e-Books, but
rather as showing the existence of the books to
the public.
Digitized books in Austria-Forum offer much
more than usual collection of digitized books, or
individual e-Books, or books on Kindle and
such. Usually, digitized libraries offer books in
“complete isolation”, i.e. neither with links from
the Web to a spot in a book, nor supporting a
section of a book by e.g. providing additional
pictures, or linking to audio or video-clips, or
even to some interactive material, potentially
turning the book even into an E-Learning
system. Even more is true: with new tools, it is
intended to link sections of different books (and
material on the Web) dealing with similar topics
together on demand, thus allowing users to look
at a topic in various sources and from various
points of view easily, rather than having to
laboriously search in all kinds of collections for
relevant contributions.
In addition, a new generation of digitized books
we are developing is almost ready: It will be
compatible with the International Image
Interoperability Framework™ (IIIF) [30]. This is
expected by most major libraries since it offers
many additional features as described in [5].
5. HOW CAN ONE ASSURE MEANINGFUL
ANSWERS FROM SEARCH-ENGINES
There are three different important issues to
deal with:
(a) How one can easily find what one wants
(b) How answers are to be interpreted
(c) How one can assure that answers
obtained are correct
Concerning aspect (a), search-engines have
evolved quite a bit over the last years. If ten
years ago you entered “Enzian” (the German
word for gentian, a flower often associated with
the mountains in Europe) into any good search
engine you got a long list of the many types of
the flower Enzian, but also the Schnaps
(brandy) called Enzian, the rocket Enzian built
during world War II and possibly all kinds of
“Lederhosen” or hotels or restaurants using the
word to advertise that they are close to alpine
traditions.
We tried to avoid such a multitude of answers
by structuring Austria-Forum into “categories”
and “subcategories”. Thus, when you look at the
entry page of Austria-Forum and you are
interested in the flower Enzian, you probably
should not immediately use the search function
provided, but first choose the category “Nature”
and then the subcategory “Flora”, to narrow the
scope of your search. This idea of narrowing the
scope of search was picked up in various ways
by other systems. In Wikipedia typing a word
that can mean more than one object results in
showing you alternatives from which you then
choose. Google eventually ended up with two
parallel approaches: either you can enter a full
query sentence describing in some detail what
you are looking for or you will get a few hits, but
then a set of categories. When you e.g. search
for London, you get, in addition to some pictures
and links a list of suggestions:
Searches related to London
London city
London uk
London attractions
London wiki
London population
London map
greater London
London facts
It is not clear how queries will develop. Some
believe that text/language understanding will
get good enough to allow very complex queries.
We tend to believe that a multi-step process
might be better in the end. Typically, the query
“London” might result in a few obvious hits
(Google-like), then presenting alternatives, after
you choose further alternatives, and so on, until
the system has determined in a dialogue with
you what you are really interested in. Much
information on this can be found in [14] and [17].
It is worth mentioning that the categories we use
have still another reason: Material in different
categories may well come from different
sources. The entries in “Biographies” are much
based on the huge collection of biographies the
Austrian Academy of Science [31] is offering,
yet those biographies are very technical and not
suitable for the general public for which the
Austria-Forum is intended. The category
“culture” (Kultur) is really a collection of various
special topic encyclopedias compiled under the
supervision of one of the editors, the category
“geography“ is based on reliable data from
different e.g. UN servers, etc. The category
“pictures” (Bilder) forced us to merge a number
of reliable data-bases. Since title and metadata
may differ for the same picture, and the pictures
themselves may be a bit distorted, this is not a
trivial effort [3), [16]. Encouraged by chatbots
and “automatic reporting”, we have also applied
this to some extent to geography [15] by using
interactive approaches and producing graphs
on demand [32].
Aspect (b), how answers are to be interpreted,
is often not really understood. However,
consider a query like “Number of Nobel Prize
winners in the UK”. Most search engines will
give you a number, but leave you at a loss,
since you don’t know: Is this the number of
Nobel Prize winners born in the UK? Or is it the
number of Nobel Prize winners currently living
in the UK. Or is it the number of Nobel Prize
winners who got the Nobel Prize for research
they did in the UK, etc. We believe that most
search engines ignore the fact that answers
without the definition on what the answers are
based on, are fairly useless. We have tried
always to provide this information in Austria-
Forum, particularly in our interactive section on
geographic information.
Aspect (c), how does one know if an answer
obtained is correct, is clearly very important but
also quite hard to answer. After all, more than
one well-founded opinion may exist concerning
a topic. Thus, in general, the best one can do is
to offer a set of answers as explained at the end
of Section 4. However, if answers are numbers,
one can try to do better. We have checked
certain geographic facts in Austria-Forum by
using a number of data-bases: if all give the
same value: fine. If they differ, the best we can
do is to show the difference and point to the
sources the numbers come from. So, if you want
to know the area of France, we cannot tell you
exactly, but we just tell you this:
Factbook: 643801
DBpedia: 674843
Geoname: 547030
Infoplease: 547030
Britannica: 543965
Wolfram: 551500
Here is an attempted explanation for the
differences: For France, Britannica gives the
smallest area. This agrees exactly with the area
in the largest German Encyclopedia Brockhaus
(2014): It follows the French Land register data
that excludes lakes, ponds and glaciers larger
than 1 km² and the estuaries of rivers. (This is,
by the way, very much in contrast to how the
figures are arrived e.g. for Finland that not only
includes freshwater lakes like many databases
do, but also ocean channels!). The situation of
France is particularly complicated also due to
overseas departments (by law, genuine parts of
France). But should their area be counted? Or
how about still more subtle cases like New
Caledonia or French Polynesia, with special
agreements with France? Or the French
Antarctic section that is claimed by France but
never accepted in the Antarctic treaty?
Or, not so relevant for France but very much for
low lying islands, do you measure the size at
low tide or high tide? How about political
contentious areas: is the Crimean Peninsula
now part of Russia or Ukraine?
What we have briefly mentioned concerning the
area of countries applies to many other facts, be
it the population of some city, the number of
mountains in a country, etc. Some more on this
is found in [14].
The idea and results in this section show very
clearly that much of further research will be
required to solve most issues satisfactorily, but
we are proud to say that we have successfully
begun to tackle some of them.
6. THE ROLE OF THE UNDERLYING SERVER AND
COMMUNICATION FACILITIES
Section 5 has made it clear that search facilities
on large servers or sets of servers still need
much improvement. This certainly includes
Austria-Forum; yet we want to mention a
number of issues that we have pursued or are
pursuing.
The fact that our information is structured in
categories, subcategories etc. makes searching
a bit easier, particularly because we allow
categories to overlap. Thus, e.g. essays on
mobility can also be found under traffic, but only
one physical copy exists.
We allow full-text search in each (sub) category:
This is essential, since a full-text search over
the whole server is unrealistic. The full-text
search for digitized books is restricted to one
book, but a feature to dynamically define a set
of books for full-text search is in preparation.
In a contribution, linking to others is often
desirable. Yet linking a word to some other
page, just on the basis of the fact that an entry
exists for that word, yields a multitude of links
on every page, destroying the appearance and
readability. We have tried to tackle this problem
in a number of ways: A link for a word is only
created when the destination is more than a
definition of that word; often, a link to one entry
does not make sense, hence we have collected
contributions into “topics” (“Themen”), as a first
try of collecting all relevant contributions
together as described at the end of Section 4,
and we can link to such collections.
For user-friendliness we allow to initiate a
search in Austria-Forum for a word by just
double-clicking at it.
Searches (if not full-text) are carried out on the
basis of the URL of the contribution, its title, its
major headings and associated meta-data.
Meta-data, at the moment, is mainly compiled
manually and non-structured, what is a serious
problem whose solution we have been working
on: basically, for each Web-page, meta-data
based on textual analysis should be generated
and added, or at least proposed for addition.
The addition of metadata/keywords for pages of
digitized books is planned. This will allow
additional search facilities within books and it
will also make it easier to find books dealing with
a topic of interest.
As has been explained before, registered users
can contact authors of contributions. They can
also add comments visible to the public to each
contribution: Such comments are always
verified by the administrative team of Austria-
Forum to avoid misuse, but are also a simple
tool for an online discussion that we will extend
to more sophisticated discussion facilities, often
only for well-defined user-groups in [5].
There are other aspects of communication
among users and, also, between users and the
administration. For this, it is crucial that the
identity of all contributors is assured. An aspect
of the reliability as addressed in [14].
It is remarkable that as early as 2007 the
question was raised whether the Web / Internet,
which are taking over more and more of our
cognitive work, are not making us stupid, see
[10]. In [4]. A reasonably positive answer is
given, yet if one looks at [7], a contribution
written by one of the Internet pioneers, it
appears that developments have to be
observed carefully.
There is one aspect that has not received the
attention it deserves. A link in a document is like
a “goto” in a programming language, leading to
a new place without informing if and when to
return to the place where one came from.
Already in 1987 this problem was discussed
[18]: Ted Nelson suggested “transclusions”
rather than links (in programming terminology
procedure calls rather than gotos) since he and
others [19] tried to convince the community to
not use gotos, i.e. links, but few took up the idea
[6]: In Austria-Forum a tool we call “InsertPage”
does do, however, most of what is expected.
7. HOW TO FIGHT FAKE NEWS AND STATEMENTS
THAT ARE HALF TRUE.
The comprehensive report [12] has at the
beginning a statement that describes the
situation very well. It essentially says that we
are in danger of developing from a rational (fact)
based society into an emotion (feeling) based
society simply because one cannot trust facts
found on the web. For every statement one can
find counter arguments, sometimes solid ones,
sometimes on purpose subtly wrong
statements, motivated by either political or
economic interests.
Hence [12] recommends to have servers in
every language with reliable information, one of
the attempts of Austria-Forum, but also an
attempt by others. Together with the team of the
Great Norwegian Encyclopedia [33] the
Directorate-General for Parliamentary
Research Services of the EU parliament
organized a meeting of representatives of all
countries involved in building such reliable
servers in Brussels Oct. 9-11. The exchange of
information between a total of 17 such efforts
proved very valuable. It made clear that closer
collaboration will be essential to cover all
relevant topics. Notice that there are two kinds
of major topics of concern to all beyond the
usually expected information of mainly national
interest.
One set are important developments from
energy generation to mobility, from climate
change to protecting the environment, just to
mention some examples. It is really essential
that if one e.g. wants to find out “what is the
future of hydrogen energy” (in general or for
mobility), one can find a reliable contribution on
this topic. This clearly applies to a large variety
of topics. It is our feeling that this can only be
handled by having groups of experts for specific
topics assigned to some country, the outcome
translated into all languages (probably semi-
automatically with some polishing afterwards).
Another set is information on political decisions
on a national or European level. It is
unacceptable, as it happens today, that after
some (parliamentary) decision subtly modified
versions (by sloppiness or on purpose) are
distributed over social networks, influencing
many receivers without them being even aware
that they are manipulated. We believe the
proper way to do this is to set up a discussion
forum for each such topic, let everyone voice
their opinion, but extract from the minutes of the
meetings the exact decision that was taken, and
if the vote was not anonymous maybe even
showing who voted in which direction. In the
soon to be released software [5] this facility will
be available.
8: HOW TO REACH ALL AGE GROUPS
Running servers with reliable information is
important for all age groups. Yet many servers
tend, by the large amount of textual material, or
by their design, to appeal more to one age
group than to another. It has to be accepted, like
it or not, that reading long coherent pieces of
texts or full books is done significantly less by
younger generations that grew up with smart
phones and information from SMS, from social
media, from YouTube, etc. Thus, servers will
have to be designed to appeal not just to one
age group, but to all, i.e. incorporating more
pictures, clips and less text in some version and
more traditional ways of information in others.
Or to put it more generally: whether servers or
digital libraries, new developments of IT
technology and their effects have to be taken
care of.
Encouraging active participation in discussions,
even competition (e.g. on the level of schools or
classes), maybe quizzes with even some small
rewards, or interactive experiments that can be
carried out, will appeal to some, may even turn
the traditional information server into an
instrument supporting learning [9], [20].
9. CONCLUSION
We have tried to argue in this paper that a
general knowledge server, offering high quality
information, has to support a number of ways to
assure such quality; one of the most important
ones is to add digitized books with very flexible
ways of linking them to other material and to
provide functions for feedback and potentially
moderated discussion groups.
It should also incorporate different ways of
accessing material to make it attractive to more
than one age groups and replace the paradigm
of only retrieving static web pages by accessing
data in such a way that reports, as the user
desires, are generated dynamically. We have
implemented much of this in Austria-Forum (1]
as a prototype and are about to roll out an
advanced version (particularly concerning
digitized books).
We believe that such a server is not just of
interest for the public in general and a valuable
tool for teaching and learning, but can also help
much in revealing fake news as such and give
fair presentations of all the many global
problems facing mankind, hopefully in various
modes corresponding to different groups of
users.
REFERENCES
[1] Maurer, Hermann: Austria-Forum and Beyond; IPSI BgD
Journal, Transactions on Internet Research; (2014), vol. 10,
no. 2 (ISSN 1820-4503), 3 - 8
[2] Kulathuramaier, Nara; Maurer, Hermann: A Survey of
Communications and Collaborative Web Technologies;
CIT- Journal of Computing and Information Technology vol.
23, no.1 (2015), 1- 18
[3] Mehmood, Rizwan; Maurer Hermann: Merging image
databases as an example for information integration;
Central European Journal of Operations Research, vol. 23,
no. 2 (2015), 441-458)
[4] Does the Internet Make Us Stupid?; CACM, (2015), vol:
58, no. 1, 48-51
[5] Zaka, Bilal: Investigating interaction activities in digital
libraries: The Networked Interactive Digital Books Project.
(In this Issue)
[6] Krottmaier, Harald.; Maurer, Hermann: Transclusions in
the 21st Century; Journal of Universal Computer Science
(2001) vol. 7, no. 12 (1125–1136
[7] Lanier, Jaron: Ten arguments for deleting your social
media accounts right now; (2018), Penguin, Random
House, UK
[8] Maurer, H., Mueller, H. (2013) Can the web turn into a
Digital Library; Intl. Journal of Digital Libraries, Springer
Pub. Co. vol. 13, no.2, 65-75
[9] Scerbakov, A., Ebner, M., Scerbakov, N.: Using Cloud
Services in a Modern Learning Management System,
Journal for Computing and Information Technology (CIT)
(2015), vol. 23, no.1 (1), 75-86
[10] Brabazon, Tara: The University of Google (2007)
[11] Delilovic, Namik; Zaka, Bilal; Maurer, H.: Libraries of
Interactive Books as Powerful Tool for Information
Communication; AACE- Proceedings of ED-Media (2019),
1353- 1359
[12] Bentzen, Naja: Europe's online encyclopaedias,
European Parliamentary Research Service (2018) — PE
614.657
[13] Delilovic, Namik; Maurer, H.: A Note Concerning
Feedback and Queries for Web pages, Journal of Universal
Computer Science (2019), vol. 25, no. 7, 733-739
[14] Azfal, Tanvir; Glatz, Mathias; Maurer, Hermann:
Finding Reliable Information on the Web Should and Can
Still Be Improved CIT- Journal of Computing and
Information Technology (2018), vol.26., no.1, 1-6.
[15] Maurer, Hermann; Mehmood, Rizwan: Semi-automatic
Story Generation for a Geographic Server (2017), Journal
of the Croatian Cartographic Society, vol.16, no. 27, 12-25
[16] Mehmood, Rizwan: Techniques for Consolidation and
Exploration of Information on a Webserver; Ph.D. Thesis
Graz University of Technology (2016)
[17] Glatz, Mathias: Access to Knowledge on the Web;
M.Sc. Thesis, Graz University of Technology (2018)
[18] Nelson, Ted: Literary Machine (1987)
[19] Rubin, F. (1987): GOTO considered Harmful;
Communications of the ACM 30 (3), pp. 195-196
[20] Maurer, Hermann: Problems and solutions for using
computer (networks) for education; Journal of Research in
Innovative Teaching & Learning (2017), vol. 10, no. 1, 63-
78
[30] https://iiif.io/
[31] https://www.oeaw.ac.at/
[32] https://austria-forum.org/af/Geography/Lab.
[33] https://meta.snl.no/Great_Norwegian_Encyclopedia
[34] https://www.wikipedia.org
[35] austria-forum.org
Delilovic, Namik. Attended the Polytechnical High School
in Zavidovići (2003 – 2007), got his B.Sc. in Mobile
Computing from the Applied University of Hagenberg,
Austria and is currently finishing his M.Sc. In Software
Development for Business/ Informatics at Graz University of
Technology. Diu an Erasmus stint at Bosphorus University
- Istanbul (2009 – 2010), worked for a number of companies
including Samsung Mobile Austria and was Samsung Best
Mobiler - Seul (Korea) . He has good technical skills in IT
and WWW, and already a number of publications.
Ebner, Martin. Priv. Doz. Dipl. Ing. Dr. is the official head of
Austria-Forum, but is officially mainly responsible for
networked learning at Graz University of Technology. He
has numerous publications on e-Learning and multi-media
and other topics and is author of the standard book on e-
Learning in German , the “Lehrbuch für Lehren und Lernen
mit Technologien” (http://lt3.eu). He is considered a leading
expert in his field.
Maurer, Hermann. Got his Ph.D. from the University of
Vienna and spent a number of years at Canadian, USA and
New Zealand Universities He was founding Professor and
later Dean at Graz University of Technology. He has
obtained numerous national and international awards and
has publications numbering in the hundreds. . More than
you ever want to know about him you can find e.g. under
www.iicm.edu/maurer .
Zaka, Bilal. He is a veteran IT professional with over 17
years of experience in industry and academia. Bilal
managed an offshore consultancy venture with Eletel Inc.
USA. He headed a team of developers in Pakistan who
worked on a product range for mobile devices. Besides
regular offshore software development activity Bilal
provided wide consultancy services. Bilal earned HEC’s
overseas scholarship for PhD studies in 2005 and
proceeded for a Ph. D, in Austria. Bilal joined Institute for
Information Systems and Computer Media (IICM, now
ISDS) at Graz University of Technology Austria. Bilal
worked on a number of industrial projects under the
supervision of Prof. Dr. Hermann Maurer, and Prof. Dr.
Frank Kappe.
