Use text similarity score when suggesting renaming or replacing
Right now the Task context suggests renaming when task parameters (start time etc) looks to be similar. This is not very good... For example, when a user mades a mistake in a word and tries to correct the task name Orangutan suggests replacing while he should suggest renaming.
So when suggesting renaming or replacing Orangutan should take into account text similarity score.
There are some implementations for Perl, for example, Text::Similarity. This module seems to be developed by scientists and seems to be what I need. But... it should be checked if it works as needed. And the speed should be tested.If one would ask me how would I do this... I would answer:
- I would remove all punctuation leaving only latin;
- I would consider first and last letter of each word and all other letters sorted (there were some research about this);
- Then I would compare similarity using words match.
I believe something like this should be written and compared to Text::Similarity.
#1 Updated by Andriy Lesyuk over 12 years ago
- Target version set to 1.07
#2 Updated by Andriy Lesyuk about 12 years ago
- Tracker changed from Feature to Improvement
- Project changed from Orangutan to Orangutan::Redmine