Version 4.3.5.1 is a minor upgrade, fixes a few bugs (one quite important, but affecting Mac PCs only) and includes a handful of enhancements.

Bug Fixes:

On Apple machines, the (relatively new) "Key Terms" highlighting feature (in the "Document details" window) was able to grind EPPI-Reviewer to a halt. This happened only when users had at least a few terms in their list, but didn't require to click on "Show key Terms". As a result, Mac users found that opening certain items "crashed" EPPI-Reviewer. This wasn't technically the case: in fact, the client application just slowed down to the point that it took minutes to load the item details. To fix this, we have completely re-written the code that generates the highlights (slightly changing its behaviour, see below) and the result allowed also to add new features.

The "Key Terms" highlighting feature wasn't working in "Coding Only" mode, this is now fixed.

Code Types. EPPI-Reviewer allows to use special code-types, in order to support different data extraction activities. The list includes "Selectable", "Not Selectable", "Outcome" and so forth. Some of these code-types are place holders for future functionalities (one that should become available fairly soon is "Numeric": to enter numeric data instead of "Info" additional text). However, these special types were not disabled, allowing people to use them, even if to all effects these types were equivalent to the "Selectable" one. Starting with this version, the forthcoming Code-types remain visible, but disabled.

New Feature:

In the "Document details" window, it is now possible to view and edit the full list of highlighted terms. To do so, click on "Show Key Terms" and then on "Show/Hide Terms List". The list will appear on the left side, (temporarily) hiding the code-trees. You can use the list to edit the terms, change their relevant/irrelevant status and delete them. Adding a new term is still done as before. The changes will immediately propagate to the Title and Abstract field.
Because the text-highlighting engine has been re-written, its behaviour changed slightly. Specifically, we have changed how it handles nested and overlapping terms. This is a tricky little problem, and one that doesn't really allow for a unique "always right" solution.
The current system works as follows: in the case of nested terms, such as "and" and "standard", the system will always show the highlight for "standard", this is because it finds the larger term first (the "and" part comes after "st" in the word "standard"). In case both recognised texts start at the same point, such as with the terms "colour" and "colourful" applied to text like "this abstract is colourful" but only one is set to irrelevant, the highlight for the "relevant" term will be shown. Other rules may be chosen, but we think the system above is a good compromise between the performance constraints (see fixed bug above) and the need to minimise risk: for example, the second rule should not lead to incorrectly excluding an item while screening.

Minor Enhancements:

• When right clicking a code, and selecting "Report: all text coded with this code (all PDFs)" the corresponding report will now include all contents of "Info" boxes (Additional text), even if a coded item does not have any text coded directly. We added this little tweak because many people use the "Info" box to copy and paste text from an uploaded document: in this way it will appear in the report, even if direct full-text coding is not being used.
• The “Description” text area for a code properties is now bigger. This field is used to describe the full meaning of a code, for example a code called "Regional Information" could have a "Description" such as: "Please indicate the precise geographical information as specified in the full text: some studies may apply to Greater London, some others to Wales or Scotland". The current area was very small, discouraging users from entering longer descriptions (that may sometimes be necessary).

Note on V 4.3.5.2:
This quick release fixes a bug on the the new Highlighting system: some special character combinations were able to crash EPPI-Reviewer without even generating an error: the "Document Details" window simply stopped to work properly. This is now solved.