Release Notes for version 1.3.1.21

A maintenance and bug fixes release, this version extends the new Readability Analysis tool.

New Features:

• Readability Analysis: added support for French texts.
  Supported readability scores are:
  "Kandel-Moles adaptation of Flesch Reading Ease formula for French language" and "Szigriszt's Lisibilité score for French texts".
• Readability Analysis: added support for Spanish texts.
  Supported readability scores are:
  "Fernandez Huerta index" and "Szigriszt Perspicuidad score for Spanish texts".
• Readability Analysis: added "Vacca-Franchina 1986" score for Italian texts.
  Less known than the 1972 score with the same name, we waited adding it because the Italian Wikipedia page documenting it was clearly badly wrong. After more research we finally found some documentation and results to compare to.

User Experience / Usability:

• Readability Analysis: when user inserts an invalid XPath expression, the validation error message was delivered by an obstrusive message box; now it as a much more discrete error pointer within the field.
• The splash window is no longer "top most". This was especially a pain for users with slow PCs as it prevented working with other programs in the meanwhile.
• The splash windows now has a tiny black border, making it easier to distinguish from the background and making a more pleasant user experience.

Performances:

• Readability Analysis: minor speed-up in computing number of syllables in Italian texts.

Various:

• Minor corrections in Credits notes.

Fixes:

• HTML5 new HTML entities were left unescaped when getting the plain text from the HTML. Affected parsing titles, meta data, and searching the plain text or extracting to compute readability score or show it in the Plain Text view.
  1778 out of 2031 HTML5 named entities were not recognized when transforming HTML to plain text and were left unascaped.
  Now they all are correctly handled. Furthermore, the transformation is faster.
• Fixed real-world crashing condition in Readability Analysis occurring when trying to extract the plain text out of an HTML section containing some uncommon HTML5 named entities. Two users experienced it.
• Readability Analysis: abbreviations caused erroneous sentences count. They are now handled for all supported languages.