Why was Markov Dictionary Switcher created?
I was using Dao Gottwalds dictionary switcher for a while and I really loved it. However, autodetecting the language that i was typing took anywhere from 1 to 10 seconds during which Thunderbird would completely freeze. Also the Dutch language pack for thunderbird installs about 10 dictionaries while I only wanted to use 1.
I decided that I wanted to fix these issues. A quick glance at the code showed that the old dictionary switcher was cycling over all available dictionaries to find the one that gave the fewest spelling mistakes. This works, but it is very slow. Instead I created a markov chain of letter combinations and used some statistics to determine the language of a text. This is about 4000 times faster on my machine and usually completes in less than 1 millisecond.
I also added a properties dialog to choose the dictionaries that you actually want to use.
At first I thought about contributing to Dao's plugin but it was hard to get in touch with him and since my approach to language detection is so different I believe that there still is a market for his approach. Thus I decided to fork instead of sending my code to Dao.
What's next for Markov Dictionary Switcher
There will be a Firefox plugin shortly.
About the Developer
|User since||October 25, 2010|
|Number of add-ons developed||1 add-on|
|Average rating of developer's add-ons||Rated 4 out of 5 stars|