LF_aligner_3.12_changelog.txt

New in 3.12:
- Segments with no text in one of the languages not added to bilingual TMX files
- Improved functionality in adding notes to segments in TMX files in the GUI
- GUI and command line arguments in TMX maker

New in 3.11:
- File name input bugfix

New in 3.1:
- Input files can now be in different folders; the output files are placed in a timestamped new folder created in the folder of the first input file (original input files are always left where they are)
- The aligner can now be called with command line arguments (file names, languages & other settings), providing a new and improved batch mode that also supports multilingual batch projects (see readme for details)

New in 3.01:
- Batch aligner bug fixed (no changes made to main programme)

New in 3.0:
- Graphical user interface added (enabled on Windows by default; Linux and OSX users should check the readme)
- Excel macro to make the fixing of misaligned segments quicker
- New segmentation rules (nonbreaking prefix lists) added for numerous European languages including, among others, Spanish (es), French (fr), Italian (it), Swedish (sv), Russian (ru) and Dutch (nl)
- TMX language code defaults filled in automatically based on languages picked (can be overwritten in setup)
- File naming convention changed for multilingual projects (file names often ended up being too long for Windows with the old system)
- Improved processing of presegmented TMX input files
- Cleanup question skipped by default (can be re-enabled in setup)
- Better out-of-the-box large file support (chopping mode automatically enabled for files over 15000 segments)
- Miscellaneous minor fixes and improvements

New in 2.59:
- When reviewing your files in Excel, now you don't need to transfer your changes to a txt file manually - just save the xls itself
- Xls input files now accepted by TMX maker
- Bug in the "filter out untranslated" feature fixed
- Automatic evaluation of segmenting in both the normal and batch aligner (needs to be enabled in setup)

New in 2.58:
- Bugfix for non-GUI mode

New in 2.57:
- List of language codes printed on user's request
- Minor EP report file naming bug fixed

New in 2.56:
- Bug affecting segments starting with "- " fixed

New in 2.55:
- Bug fixed (the output file got deleted halfway through when run in 2-language mode... oops)

New in 2.54:
- The main aligner now does multilingual projects (up to 100 languages); separate 3- and 4-language aligners not needed anymore
- Bug fixed in TMX maker (when taking notes from the input file, the note in the first segment was propagated to the whole file)

New in 2.53:
- Dictionary data added for 32 major languages to improve alignment results; see the two-letter language codes you need to use in the readme and at http://en.wikipedia.org/wiki/List_of_ISO_639-1_codes
- Mac and linux: segmenter bug fixed; windows: minor program termination bug related to web features fixed

New in 2.52:
- TMX input files now accepted (allows you to presegment with your CAT tool and thus align any file your CAT can process - see "Using your CAT to extract/segment text" in readme)

New in 2.51:
- Match confidence value now removed by default

New in 2.5:
- Graphical file selection window instead of drag & drop on Windows (can be disabled in setup)

New in 2.41:
- Simple Chinese/Japanese sentence segmenter added - use language codes zh and ja to apply (send feedback/suggestions to lfaligner@gmail.com)
- The TMX maker can now generate multilingual TMX files with any number of languages

New in 2.4:
- 3- and 4-language script versions added and TMX maker and batch aligner moved to "other tools" folder to make the main folder less cluttered
- The TMX maker now skips malformed lines

New in 2.302:
- Pdf bug fixed

New in 2.3:
- Support for various file formats with Abiword, including .rtf and .odt. If you're not on Windows, you'll need to install Abiword separately; visit http://www.abisource.com/download/index.phtml
- Delete duplicate and untranslated entries - see setup file

New in 2.201:
- Startup bug fixed

New in 2.2:
- Support for .doc input files using Antiword - see notes in readme
- Detect and use preexisting pdftotext and antiword installs on *nix systems

New in 2.1:
- Some character conversion functionality added (see end of setup)
- Retrieval of CELEX documents improved; now the aligner should find a lot more (all?) documents with a CELEX number
- Better Studio compatibility: creator name, creation time and Note should now be recognized by Studio (the TMX generator now spoofs the Trados o-tmf string)

New in 2.01b:
- Batch aligner now makes xls and tmx files out of entire project
- Other minor fixes in batch aligner

New in 2.01:
- Some TMX compatibility issues fixed (srclang, segtype and &)


Differences compared to aligner.bat:
- Now supports Windows, Linux and Mac OS X
- Pdf and docx source files also supported on top of txt and html
- Drag & drop source files from any folder
- Batch alignment, i.e. align any number of txt, docx, pdf or HTML file pairs in one go
- Formatted xls files created after autoalignment
- Simple download+alignment of Commission proposals and EP reports as well as CELEX documents
- No need to install Perl or any other software(applies to BAT as well, as of 1.1.4.)
- Extensive setup of defaults etc. in separate file (aligner_setup.txt)
- Time and date autorecognized
- Any new dictionaries added by the user are now autorecognized without having to change the code
- No 3 and 4 language mode
- More error checking done and more data reported by aligner in console and in log file
- Lots of other minor changes and improvements; e.g. defaults are different, better HTML segmentation etc.
- Hu-En dictionary flipped to En-Hu, now you need to go English first - get hu-en.dic from ftp://ftp.mokk.bme.hu/Hunglish/src/hunalign/latest/hunalign-1.0/data/ and save it in scripts/hunalign/data if you want the dictionary to work in either direction