The localization world is no stranger to the document interchange problems of word processors. The XLIFF format emerged to assist localizers in sharing and moving localization files between tools.

We’ll look at some of the history of XLIFF, the false starts and the issues of some of the initial versions. Then look at the current 2.0 version of XLIFF and the advantages that have come about by simplifying the format.

The Translate Toolkit is used as a file converter into PO and we’ll look at our efforts to migrate the toolkit to use XLIFF as its target format. XLIFF, being a richer format, allows us to express more of the data that we get from other files formats and that we struggle to squeeze into PO.

To achieve this we are developing a FOSS library for XLIFF 2.0. and implementing extensions to the formats we need to support. In doing so, we aim to build on XLIFF’s powerful extensibility to provide lossless storage of localisation data, and minimally lossy conversion between formats.

Once we have stabilized this library, we will be refactoring the data model within Pootle, our web-based translation tool, to make better use of these new features for conversion and data interchange.

We have identified a number of common localisation patterns that are not specified within the core XLIFF schema, so we will be working with the relevant standards body to have these use cases specified as standard extensions to the protocol.