The original condition is evaluated to false.
The substring is shorter than "uni".

Perhaps the assumption should be that the letters are ASCII.
If all letters are ASCII checking 'A' <= k.charAt(0) <= 'Z'
might make more sense.
I am not convinced about doing this with a regex.

Both test cases involves an author name containing department or school
 without university or institute of technology.

Corporate authors without university/institute of technology

# Conflicts:
#	src/main/java/org/jabref/logic/bibtexkeypattern/BracketedPattern.java

If the name of an institution can't be split, assume that
"School"/"Department" is part of the name.

Minor typos and the "rest" part is now created differently.

There are some examples that will turn out wrong if only capital letters
 are kept, e.g., "iOS Developer University Program".
 The original problem with "The School of Life" becoming too short is
 avoided by only removing school/department for names containing two or
  more ',' separated strings. This will still produce an unexpected
   result if the address of the institution is part of the authors field

First name first and abbreviated first name first

Last name first and abbreviated first names

The deprecated static methods BibtexKeyGenerator.generateKey are moved
to the test file as a similar convenience method is required for
the test cases. Suppress warning has been added for some methods.

generateKey was not copy-pasted properly.

The author list parsing is moved outside of the if/else statements

Move them close to other parse tests

Same as for all others

# Conflicts:
#	src/main/java/org/jabref/gui/maintable/BibEntryTableViewModel.java
#	src/main/java/org/jabref/gui/maintable/MainTableNameFormatter.java
#	src/main/java/org/jabref/model/entry/AuthorList.java
#	src/test/java/org/jabref/model/entry/AuthorListTest.java

This reverts commit 94a46e2

Since the name of an institution can be used to create an `Author`
directly it is the most likely `Author` to contain a reference to the
original String object and create a memory leak.

Updated JavaDoc and now checks if the other object is itself

The parser is intended to deal with a format that is not BibTeX but
appears to be used in practice

Make the unique names more realistic

This reverts commit 557ffc9

Moving the responsibility of getting the unformatted latex-free name to
aid in caching and keeping it consistent with other getter functions.

Move the fields loop from BibEntryTableViewModel to
MainTableFieldValueFormatter.

Store a reference to the used BibDatabase in
MainTableFieldValueFormatter with the constructor.