Lars Marius Garshol

_From [[email protected]](https://code.google.com/u/106380900043315593284/) on February 04, 2013 19:12:00_ We can implement a simpler one that assumes the database is clean every time Duke is run. This allows us to avoid lots...

_From [[email protected]](https://code.google.com/u/106380900043315593284/) on February 04, 2013 19:10:05_ We need some proper tests for this before we can make an 0.7 release. _Original issue: http://code.google.com/p/duke/issues/detail?id=109_

_From [[email protected]](https://code.google.com/u/106380900043315593284/) on January 15, 2013 14:59:47_ This comparator is supposed to be very good, so we should support it. _Original issue: http://code.google.com/p/duke/issues/detail?id=103_

_From [[email protected]](https://code.google.com/u/106380900043315593284/) on January 15, 2013 14:59:13_ The ability to reduce scores for insertions longer than one character should be useful. _Original issue: http://code.google.com/p/duke/issues/detail?id=102_

Add a longest common subsequence comparator

2

_From [[email protected]](https://code.google.com/u/106380900043315593284/) on January 12, 2013 12:39:54_ Might be good for comparing phone numbers, for example. _Original issue: http://code.google.com/p/duke/issues/detail?id=99_

_From [[email protected]](https://code.google.com/u/106380900043315593284/) on January 10, 2013 08:38:38_ It should be possible to say "ignore", in case one doesn't want to commit to y/n. In addition, it should be possible to...

Speed up Levenshtein comparison with automata

1

_From [[email protected]](https://code.google.com/u/106380900043315593284/) on October 02, 2012 19:08:50_ It's possible to generate a Levenshtein automaton in O(n), and to run comparison with another string in O(n), for a fixed max edit...

_From [[email protected]](https://code.google.com/u/106380900043315593284/) on March 26, 2012 20:33:32_ Might, for example, use jgeocoder. http://jgeocoder.sourceforge.net/quick.html This would involve an additional dependency, so may want to move this out into a separate contribs...

Support for deployment-independent configurations

2

_From [[email protected]](https://code.google.com/u/106380900043315593284/) on March 09, 2012 13:44:46_ It's a problem that the current Duke configuration mixes deployment-specific details in with the more general mapping and comparison setup. The actual server...

_From [[email protected]](https://code.google.com/u/106380900043315593284/) on March 08, 2012 11:31:27_ This is _not_ the same as Q-gram difference (even if Q-grams and N-grams are the same thing). Described in this paper: http://www.cs.ualberta.ca/~kondrak/papers/spire05.pdf Appears...