Releases

Globalese 1.4 released

Globalese 1.4 includes many useful new features and various fixes.

Integration

  • Translation memories and term bases on memoQ servers can now be easily pulled into Globalese for training purposes.
  • Translation memories and term bases in the MemSource Cloud can now be easily pulled into Globalese for training purposes.
  • XTM XLF files can be translated.

Language support

  • Thai language support
  • Faster processing of Japanese text

Additional enhancements

  • Term bases can now be uploaded in the MultiTerm MTF XML export format
  • Profiles are now called engines in Globalese.
  • The new Engine Wizard allows for easier engine creation for beginner users.
  • Retrain/Repair option for models enables retraining even when the underlying corpora hasn't changed (useful after upgrades).
  • Visualised tuning history for engines.
  • Possibility to choose tuning files from the repository (users are not required to upload them if they want to use something that is already in the repository).
  • Better error handling and display.
  • New checks implemented for the language pair of uploaded tuning/translation/evaluation files, to prevent the uploading of files with wrong language combinations.

Issues have been fixed concerning:

  • Training
  • Translating and evaluating SDLXLIFF files
  • Sending back memoQ files to a memoQ server

Globalese 1.3.3 released

Globalese 1.3.3 is out with the following fixes and enhancements:
  • Improved tokenization for English, Italian and Swedish.
  • Segments that were not translated by Globalese (due to the quality threshold) appear in a different category in the evaluation report.
  • Improved normalization rules for special characters.
  • The translation of memoQ .mqxlz files is now supported, meaning easier export from memoQ.
  • When reimporting translated files to a memoQ server, document version numbers are now continuously assigned.

Globalese 1.3.2 released

Globalese 1.3.2 delivers the following fixes:
  • Improved handling of multilingual TMX and TBX files.
  • memoQ file parser bug fixed.
  • Handling of invalid language codes in TMX or TBX files (invalid language codes won't appear as a new language code any more).
  • Improved tokenization rules.

Globalese 1.3.1 released

Globalese 1.3.1 brings some improvements in file management and under the hood as well:
  • Post-edit evaluation reports can be downloaded in CSV format.
  • Improvements in the tokenizer.
  • Possibility to translate selected files only in a project.
  • Some minor UI enhancements.

Globalese 1.3 released

Globalese 1.3 is here with some new features and fixes.
  • MemSource Cloud support: connect to your MemSource Cloud projects to pull and push files.
  • If you work with offline MemSource .mxliff files, that is also supported now (both as a resource and a translatable file).
  • It is now possible to evaluate machine translation against multiple (post-edited) reference files.
  • Various UI and speed improvements.

Globalese 1.2.2 released

Globalese 1.2.2 is out, with the focus on improving profiles through the Unknown words report. Fixes and new features include:
  • It is now possible to download the report in a TXT file, and process it in your CAT tool
  • Consequently, it is now also possible to upload bilingual CAT files (.mqxliff, .sdlxliff, .ttx., .txml and .xliff) to the Repository.
  • Fix for Japanese issues with the Unknown words report.
  • TTX issue with extra tag insertion fixed.
  • Some further improvements have been made to speed up project management.

Globalese 1.2.1 released

Globalese 1.2.1 comes with some new features and a number of fixed issues:
  • The new Unknown words report helps you improve the quality of your engine by identifying unknown words in a project. You can download the list of words and phrases, translate it, and upload it to the Repository to include it in your profile.
  • Globalese now supports connecting to memoQ server projects with more than one target language.
  • Various improvements to speed up project management.
  • Improved SDLXLIFF and TTX filters.

Globalese 1.2 released

Globalese 1.2 includes a couple of important new features and various fixes.

memoQ server integration

Globalese now offers integration with memoQ server 6.2 through the memoQ WS API.
  • Files prepared in a memoQ server project can now be directly retrieved to Globalese, machine translated and then sent back to the memoQ server project again without the need to export, import, download or upload anything.
  • Once post-editing is finished in memoQ, the post-edited files can be retrieved once again into the Globalese project for evaluation and to improve the engine quality through retraining.

Additional new features

  • It is now possible to evaluate an entire project by one click.
  • PER scores are now also available when evaluating a file or a project
  • Now keeping count of machine-translated segments
  • Changed colouring indicator of machine-translated segments in memoQ for easier separation.
  • Evaluation now works with segments that have been joined or split during post-editing.
  • Evaluation of Wordfast .txml files.

Fixes

  • Various issues concerning the translation of .ttx and .mqxliff files have been fixed.
  • For large projects, the user will only see the number of files in the translation or evaluation queue, not a list of all the files.
  • Improved tag replacement.
  • Better group statistics.
  • Shorter translation time when translating from Chinese.

Globalese 1.1.1 released

Enhancements in Globalese 1.1.1 include:
  • Through the new Recycle feature, it is possible to send all human-translated files in a translation project to the profile used in the project, and use them to tune the profile; or to the repository, and use them to rebuild the models.
  • The total number of segments of a translation file is displayed next to the number of translatable segments.
  • The total number of segments has been added to the evaluation results window.

Globalese 1.1 released

Globalese 1.1 includes various fixes and some new features:
  • Through the newly introduced Threshold attribute, it becomes possible to filter machine-translated segments by expected post-editing analysis rate. Users can set a rate between 0% and 100% that will correspond to results of a comparison against a final human-translated version of segments. Globalese will only insert translations into those segments where this threshold is met; otherwise it will leave the segment empty, saving valuable time for post-editors.
  • Better evaluation results for files where segmentation is changed during post-editing.
  • Improved use of bilingual glossaries.