Better visible page & text overlay layer combinations
Resizing word boxes is possible in shrunk/zoom view too
Keyboard shortcuts (F2 to edit, cursor move on word boxes, +/-: zoom, 1,2,3,4: layers, 0: toggle zoom, see [HotKeys])
Right-click on page area brings up the jsTree menu for the clicked word
Click selects words on page area (with CTRL: toggle, or SHIFT: ranges, multiple disjoint ranges possible)
Cut & Paste allows to paste to a different tree levels for easy structure fixes (e.g. paste two LINES "After" a PARAGRAPH and a new paragraph with those two lines is auto-created).
Ability to Split two columns if the OCR created "long lines" (see on same page as [MergeTwoColumns]
Split text and auto-create new word boxes if you write space characters into the text
Spellchecker supports division of words, fixes numbers with o, O and l letters, added an "-edit-" option so that you can fix the word if none of the suggestions is good
Added a symbol table to insert special characters
Added a bash shell script to help DjVu -> XML & PNG conversion on Linux
Usability improvements in file open, especially in image name patterns and better error messages
Technical changes:
Changed page scaling code from CSS Tranform to recalculations for less CPU use and more possiblities
jsTree and some other third party libraries are upgraded to newer version
jsTree "wholerow" plugin for better visiblity of selection
Image name pattern supports partial books exports, where first page of the XML is in e.g. 012.png (%%%+12.png) See [FileOpenSave]
XML output closely matches that of djvutoxml except that there is no indentation. If you delete identation before starting to edit, you can track all your changes if you keep your XML in a source code repository
Number of bugfixes, see SVN history for complete changes