For the compilation of the dictionary database we had
at our disposal the computer program OMBI (Omkeerbaar Bilinguaal
Bestand = Reversible Bilingual Database) that was devised for this purpose.
The main characteristics of the program are:
A short project description can be
read through the following link.
import of data in several formats |
export of data in several formats (SGML, HTML, RTF)|
easy switching between source and target language|
A publication on the program can be accessed through the following
Version 4 of the program, which became available in
February 1997 contained some special features to handle Arabic:
As for the technical details:
a root field|
Arabic alphabetic ordering|
ordering of entries within a root|
all functionalities for using vowels, sukun and
shadda in Arabic|
Below I will show some screenshots of the program:
1 The main screen
Arabic form = the entry (ßöÊÇÈ)
root = the root of the Arabic entry
Lexical Units, Examples and Idiom = Definitions of the
various Lexical Units (clicking the plus signs will show the Examples which
belong to a certain LU, see below)
Nederlands-RBN translations of ... = the
translation(s) of the selected source language unit (LU or Example or Idiom]
2 The main screen, showing Examples
Lexical Units, Examples and Idiom = This field now
shows a number of examples belonging to the LU ..., on the right we see the
translation in Dutch of the selected Example.
3 Form list
This list shows all entries belonging to a certain
root, according to the order of presentation in the dictionary.
The ordering is according to the following rules:
it was programmed in Delphi|
the underlying database is Borland Interbase|
4 Edit Form - Morphology
below shows morphological information for a
verbs in ascending degree of derivation (I to X)|
words consisting of the three radicals only (no
prefixes, infixes, suffixes)|
words starting with the first radical (containing
words starting with a prefix|
below shows morphological information for a noun:
Verb.subst. = masdar for stem I|
imperfect. kl. = vowel in imperfect for stem I|
Working with OMBI
We have been using OMBI in the various stages as
Stage 1: Translation Dutch-Arabic
Flections = plural form(s)|
Stage 2: Preparing Arabic-Dutch
import of the
RBN Dutch files |
addition of Arabic translations to the Dutch units
(Lexical Units and Examples, some Idiomatic Expressions, Descriptions).|
checking new Arabic entries by switching to Arabic
as source language. Not all Arabic translations were checked during this
stage, but new entries were. In the early stage almost all Arabic words
entered of course were new, but obviously the percentage of new entries was
Final Stage: Checking and rechecking.
Ombi has some very useful options for making
selections from the database. By queries almost all levels of units can be
marked and exported from the database.
In this way it was possible to make prints of separate
categories like morphological information, descriptions etc.
And finally all data to be printed in the dictionaries
would be exported and printed for proof reading.
expansion of the Arabic macro: addition of entries,
including Dutch translations|
expansion of the Arabic micro: addition of examples,
including Dutch translations|
The extraction of data for the final production process was done in
a different way, since Teus de Jong compiled an
Extractor to directly extract the data from the tables of the database, without
using OMBI's export options.
title: OMBI: An editor for constructing reversible lexical databases
by: Willy Martin and Anne Tamm
This article was published in the Euralex '96
Proceedings (Göteborg, editors Martin Gellerstam, Jerker Järborg, Sven-Göran
Malmgren, Kerstin Norén, Lena Rogström, Catarina Rödjer Papmehl).
The following links lead to pdf-scans of the pages of the article.
Report on OMBI and the CLVV by the Dutch Language Union
title: Constructing Reversible Lexical Databases
by: Lisanne Teunissen