Extracting / Localising Strings Embedded in Source Code

I am a professional translator who has been engaged to coordinate the localisation of a large software system written in Informix 4GL and 4js. The immediate problem facing this project is that the strings of text for translation are embedded directly in the ".per" code that is the source of the screen forms. They are present as field labels, as button labels and as comments. The field strings are often of fixed, predetermined length. These circumstances obstruct my ability simply to put the source code through a machine assisted human translation program such as are frequently used by localisers. Each string has to be painstakingly extracted, and the translated forms require a great deal of post- editing to get the fields properly lined up again. I wondered if others on the list had faced similar problems with the localisation of software developed in Informix 4gl and 4js, and if so, what solutions were brought to bear? My firm is thinking of developping its own tools for the extraction and replacement of the strings, not only in the forms themselves but also for all the messages pertaining to 4GL itself. We would be grateful for accounts of experience with this sort of thing, or indeed, if anyone knows of pre-existing add-on tools that serve this purpose. Many thanks to anyone who can advise. Elisa Francesca Roselli Localiser Volume Software Tours, France Sent via Deja.com http://www.deja.com/ Share what you know. Learn what you don't.