Koha Swedish Translation Work

Lars Noodén

Koha uses primarily the Perl scripting language and the Structured Query Language (SQL) for its database. Behind the scenes, the text for Koha's two interfaces are managed by Perl using gettext. The database is managed using MySQL.

There are translations for the OPAC, the Intranet and the MARC21 labels:

• OPAC
• Intranet
• MARC21

Two User Interfaces

Koha has two interfaces. There is a separate gettext file for each one:

• OPAC - seen by the public and the librarians - find a book, etc.
• Intranet - seen by the librarians - cataloging, circulation, subscription, etc.

gettext

gettext are a set of utilities which depend on keeping the text messages in a program separate from the program itself. In that way, the program can have text in different languages without having to re-write the program. One advantage to that is that the translation work can be done by non-programmers.

Kbabel or emacs in gettext mode, or other tools can be used to edit the language files.

MARC21

There are five types of MARC21 records, about 3400 fields and subfields in all. It can be confusing because in all five, the fields are identified by counting numbers and the subfields by letters.

• Authority (~995) - authorized forms of names and subjects to be used as access points in MARC records and the interrelationship between those forms.
• Bibliographic (~987) - commonly includes titles, names, subjects, notes, publication data, and information about the physical description of an artifact.
• Classification (~429) - classification codes and the captions associated with them that are formulated according to a specified authoritative classification scheme (~e.g. DDC or SAB or YKL)
• Community (~577) - Descriptions of non-bibliographic resources such as people, organizations, programmes or services, events, etc.
• Holdings (~411) - Information about where the artifact is physically located and is it one piece, multi-part, or serialized.

Method

The goal is to minimize the time, motion and number of repetitions needed to collate the Swedish and English labels with each other and with the field and subfield identifiers.

Many methods work. My current method is to work with the MARC21 labels in a spreadsheet, with Swedish in one column and English in another on one part of the screen. On another part there is a web browser with a pair of tabs, one for the English translation and one for the Swedish.

The MARC21 fields, subfields and labels were extracted from the Library of Congress example using a combination of scripts and regular expression pattern matching and then imported into the spreadsheet. These are then checked against the corresponding Swedish labels.

When that part is finished, the columns can be exported from the spreadsheet into a format which can be imported into the SQL database.

References

MARC21. U. S. Library of Congress. 2000 Edition. Update No. 5. Oct 2004.

LIBRIS Formathandbok (MARC21). The Royal Library of Sweden. Revised 25 Mar 2004.

Perl / CPAN

MySQL

Kbabel. KDE Internationalization project. 17 Oct 2004.

GNU gettext utilities. 6 May 2002.

Last updated: 11 Mar 2005
Lars Noodén