Well maintained Translation Memories are the basis for fast and accurate automatic translations. However, good maintenance of the files is hard work and should be done by native speakers. Regretfully, there is no common and open source for Translation Memories for MS Windows programs or for Delphi programs. The interest of the Delphi open source community for localization seems limited (as is the interest for creating help-files). There are different formats often used for translation memories. Currently, Linlocalize offers (with some limitations) support for:
and adds his own specific .glo file format. The .glo text file format is most simple: source and translation strings are stored in a single line Unicode UTF-8 text format and are delimited by tabs, while tabs and line delimiters within strings are coded by respectively /t and /n, and / as //. A glo-file is therefore sufficient readable by any Unicode text editor, most notably Notepad++.
It makes sense to keep the glossaries in one place. You can define a central directory in the Tools | Options dialog. This directory name is used when creating or opening glossaries, but there is no obligation to use it. It is recommended to place all projects and the glossaries directory in a central place.
When you change the directory name, you can start afresh and move only the necessary glossaries to that new directory.
First the project glossaries are searched, then the global glossaries. When there is no project attached to the translation, only the global glossaries are used. The files are searched in the order you have placed them. As soon as a translation is found the search process is exited.
It makes sense to
In general, it does not make much sense to use both as it slows down the auto-translation process.
A glo-file can be added to the glossaries used for automatic translation in two ways:
Each source and translation in a project and in a glossary is coded with a language code and an optional country code. Examples are EN, EN-US and EN-UK for respectively the English language, the English language as spoken in the U.S.A., and the English language as spoken in the U. K. Internally, only the language code is used and not the country code. However, it is sensible to use country codes for the languages in the glossaries, so that you can identify for which country they are intended. When you want to create a translation for a specific country, use in the project the language-country code (such as for instance DE-CH for German as spoken in Switzerland) and place glossaries (DE-CH, which have specific translations for German-Switzerland) before DE or DE-DE glossaries (which have German-Germany strings).
The languages are written to the Linlocalize.ini file. You can add or delete language names. The language and optional country identifications (for instance [DE] or [DE-CH]) has to be in capitals.
See also: Creating and maintaining a translation