This file contains some information that is relevant for the course Computational Lexicography II (in Swedish: Datamaskinell Lexikografi II VT96) that is taught at the University of Uppsala in Spring 1996.

Accessing machine readable dictionaries

In this text I will describe some facilities which you can use for accessing a machine readable dictionary. I will assume that the dictionary has a very simple format: they are stored in a collection of tables with each field combined with the field name on a seperate line and a blank line between the rows of the table. I will use the # sign to denote that the rest of the line in the dictionary is a comment. An example:

   # dictionary entries
   ENG.WORD: dog
   ENG.TYPE: regular
   SWE.WORD: hund
   SWE.TYPE: en1

   ENG.WORD: lion
   ENG.TYPE: regular
   SWE.WORD: lejon
   SWE.TYPE: ett1

   # type definitions
   ENG.TYPE: regular
   ENG.PLUR: ~s
   ENG.DEF:  the ~
   ENG.DEFP: the ~s

   SWE.TYPE: en1
   SWE.PLUR: ~ar
   SWE.DEF:  ~en
   SWE.DEFP: ~arna

   SWE.TYPE: ett1
   SWE.PLUR: ~
   SWE.DEF:  ~et
   SWE.DEFP: ~en

You can use this script searchScript for accessing your dictionary. It will search in the dictionary if you invoke it like:

searchScript yourSearchString

in a terminal window (aixterm, dtterm or xterm).

The script will select all lexical entries matching the word you are searching for. It will create a temporary web page with these entries and start netscape to display this page (this takes some time). The script is not perfect and you should adapt it to generate the output that you want. This might be difficult so feel free to ask me anything you do not understand or do not manage you do yourself.

After you have made the script work like you want it to, we will put it on the web so that you can access your data via a web form. An example of the script in work can be found at Kamal's Kurdish-English dictionary


Last update: May 30, 1996. erikt@stp.ling.uu.se