|
Academia 22
Home |
Press |
Academia |
22 |
art_18
E-Dictionary How will they change our life?
Academia Nr: 22 (März - Juni / marzo - giugno 2000)Johann Gamper, Judith Knapp
Dictionaries in paper form are out of style. New computer technologies make easy-to-use dictionaries possible and they also promise to save time. This article describes what kind of technical problems have to be solved by implementing an electronic dictionary to make it as efficient as possible for the users.
You may already be acquainted with the following situation: You read an article in a foreign language and discover several expressions which you don't understand. There is a huge dictionary lying on your table. Every time you come across an unknown word you spend a lot of time looking it up. And then you still have to check its meaning in the given context. Wouldn't an electronic dictionary be more efficient? You could obtain information simply by clicking on a few buttons or in technical terms, by navigating through a hyperspace.
At the European Academy we are currently developing an electronic learner's dictionary called ELDIT (Elektronisches Lernerwörterbuch Deutsch ITalienisch) for those studying the German or Italian languages. The dictionary will contain about 3,000 word entries for both the Italian and the German languages. The dictionary should serve a dual purpose: It should enable the students to encode what they do not understand and, at the same time, enable them to produce their own texts. Thus, the user of this dictionary will not only find the words and its translations but also a great deal of additional information: More complex expressions of a word, compound words, derivations, idiomatic expressions, and graphics which show semantic relations between the single words as well as pictures and sound files which illustrate the given information.
The dictionary will be implemented in the form of an adaptive hyperbook, a special technology which makes it possible to personalise the dictionary for the individual user. This technique facilitates fast access to the information needed.
As the developer of such a project one has to cope with the following problems: First of all, there is a vast amount of data. How can one get an overview? How can the data be entered into the computer? When the data is finally in the computer, how should the pages be generated?
What is an Adaptive Hyperbook?
A hyperbook is nothing more than a book on computer. But because of the electronic nature of this book there are many additional possibilities. Developers can include not only text and pictures like in an ordinary book, but multimedia content as well (e.g. sound or short films). To go from one page to another you have to click on a few buttons. Starting from every page you can gain access to further "next pages". Consequently, hyperbooks do not have a linear structure. There is no first or last page. You start somewhere and click around until you find the information you are looking for.
Just that "clicking around" can waste a lot of time as well. The adaptive hyperbook can help solve exactly this problem. The presentation of the content no longer depends on fixed rules. The book contains a user model which stores information such as the user's current needs and learning goals, previous knowledge, or learning abilities. Depending on this information, the book presents an individually designed page which should match the needs of the individual user. For more information see [Brusilovsky 1996].
Bibliography
Peter Brusilovsky (1996), "Methods and Techniques of Adaptive Hypermedia", User Modeling and User Adapted Interaction Journal, vol. 6, pp. 87-129.
A clear structure to feed the computer with 3000 words
The first step is to structure the data which the book will contain. Figure 1 shows a simplified version of the data model and shows the information packages stored in the dictionary as well as their mutual relationship.
Figure 1: It's easier than it looks. This are the data packages which the dictionary will provide when the user requests a certain word.
The dictionary will contain words, lots of words. Four categories of words are distinguished: verbs, nouns, adjectives, and others. Depending on the type of word, the dictionary will show different pieces of additional information. Let us have a look at the noun class. For every noun we give the lemma, phonological information (e.g. a sound-file with the correct pronunciation of the word), morphological information (which article to use, how to construct the plural, etc.), possible idiomatic expressions, and a few derivations. A noun might have more than one meaning; for every meaning the dictionary will show the appropriate translation, a picture, compound words, and collocations which belong to this meaning. Not only the meanings but all complex information pieces are provided with their respective translation.
As an example, let us have a look at the German word "Haus". The lemma of the word is "Haus." The sound file "haus.wav" will give you the correct pronunciation. You can also find idiomatic expressions and derivations. The word "Haus" has more than one meaning, such as "Heim" or "öffentliches Gebäude." The translation for the meaning "Heim" is "la casa." For this meaning we have compound words like "die Haustür" or "das Hochhaus" and collocations like "zu Hause sein" or "ein Haus bauen."
Figure 2: The German word "Haus" and some of the possibilities about how to use this word. Behind the easy-to-use e-dictionary
What are the basic modules of our system? How can a user interact with the system? What happens, when a user makes a request to our dictionary? These and similar questions are discussed in the following section.
Figure 3: What the users of e-dictionaries do not see. A complex process is needed to tell the computer which words we would like to see and how we would like to see them.
Figure 3 shows the architecture of the ELDIT system and its interaction with users and administrators. The user can use any WWW-Browser to access the dictionary via the Internet; the best known browsers are probably Internet Explorer and Netscape. Such programs are also called a "client".
There is another program which administrates the distribution of the pages of our hyperbook, which is called a "server", and which runs on the computer upon which the dictionary is located. This server is extended with an additional program which is called "servlet".
When a user wants to see a particular dictionary entry, the client program sends a request to the server. The ELDIT-servlet seeks all information packages in the database which belong to the requested word, prepares a WWW-page and sends this page back to the client.
Now we will have a look at the right side of the graphics. Language experts put the data (the single information packages) into the computer. From among various possibilities, we decided to investigate how a quite new technology – namely the XML language - could be used for our purposes. Additionally we are using the Xeena program. The language expert can insert the data via a graphical user interface, and Xeena generates the corresponding XML-files. These files are transferred into database elements and stored in a database.
Improving the Current System
There are many possibilities for improving our dictionary, which is still under construction: Our dictionary should help the learners with producing their own texts, actually with learning the language. Thus, we will add exercises and tests to find out how well the user has understood the information; we will give hints and solutions to the tests and provide additional help for defining new learning goals; etc.
Until now we have not concentrated on the adaptivity of our hyperbook. Possible parameters for an adaptation are: novice or advanced language learner, interest in a special field, different learning goals such as grammar, vocabulary, exercises, etc. All these pieces of information have to be stored in a so-called user model. According to the information in the user model, individual pages are generated for individual users. There are different possibilities for obtaining such information: the program can directly ask questions about the user's preferences and learning goals, or, it can memorise the user's actions and draw conclusions about the present knowledge of this very user. To conclude, adaptive hyperbooks seem to have a promising feature as a support for both teachers and students; they support individual and collaborative studying and they can help avoid problems associated with both language acquisition and teaching.
Dr. Johann Gamper, researcher in the section "Language and Law" at the European Academy Bozen/Bolzano
johann.gamper@eurac.edu
Dr. Judith Knapp, researcher in the section "Language and Law" at the European Academy Bozen/Bolzano
judith.knapp@eurac.edu
Figure 1: It's easer than it looks. This are the data packages which the dictionary will provide when the user requests a certain word
Figure 2:The German word "Haus" and some of the possibilities about how to use this word
Figure 3: What the users of e-dictionaries do not see. A complex prcess is needed to tell the computer which words we would like to see and how we would like to see them
|
|
