Features in Detail

Dictionary Formats

Babbletower assumes installed dictionaries to be in text file format, one entry per line. Text encoding is not restricted, as long as your Java runtime environment supports the encoding you're using. The same is true for displaying dictionary entries: it will only work if your particular Java runtime / OS combination allows you to install and setup the fonts required for display.

The format of a dictionary entry is logically divided into head entry, pronunciation/ reading, translation, and remarks No assumptions are made about where these fields are located in a line from a dictionary. The format for a particular dictionary can be defined in its setup file or via Babbletower's optional configuration file. Formats for a few open dictionary projects and a generic dictionary format are included.

For more elaborate formats that require some on the fly transcoding, custom formatters can be supplied by the user. This requires writing a Java class, implementing a formatter interface defined in Babbletower.


Lookup is done via full text search, i.e. the whole file is searched for the word to look up. This allows you for example to also search the translation part of a dictionary, not just the head entries. If necessary, the lookup can be restricted to only show hits in head entries or translations. This is important when working with monolingual dictionaries, e.g. looking up run in an English-English dictionary may yield a lot of unwanted hits in the 'translation' part.

This is of course also a problem when deliberately searching the translation part of a dictionary. Because a dictionary usually is a one-way mapping from keywords into translations/ explanations, trying to use it in 'reverse' doesn't always work nicely. For example, looking up talk in Jim Breen's edict Japanese-English dictionary (approx. 100'000 entries) yields 125 hits. But even when using a dictionary in the 'proper' direction, the number of found entries can still be very large. To avoid looking for the interesting entries among all the hits, lookup results can be sorted by relevance. The sorting is done based on the length of the field in which the hit was found, and its position within that field, e.g. short entries and entries where the hit appears close to the start of the entry are listed first. The sorting can be 'fine-tuned' individually for each dictionary.

The lookup uses an index file to achieve acceptable speed. A tool for creating index files for dictionaries is included. Searching can further be restricted to word beginnings, complete words, or word endings, or can be unrestricted. In the latter two cases, the lookup will be rather slow since the index information can not be used during this type of search. To narrow down lookup results, several words in one lookup are possible as well.

Multiple Dictionaries

An unrestricted number of dictionaries can be installed, the only limit of course being the storage capacity of your device. Switching between dictionaries and auto lookup in a dictionary group is uncomplicated. See the manual for details and screen shots.

Special Character Input and Custom Keyboards

Babbletower allows to plug in custom input methods. Input methods for special characters from various European languages, for Japanese Hiragana and Katakana, and for Hangeul (Korean) are provided as examples.

Vocabulary Lists

Entries from lookup results can be easily copied into flash cards and organized in vocabulary lists. No typing is required for this. The flash card will be filled in according to the format of the dictionary from which the entry was copied. Of course, a flash card can be edited later on from within the vocabulary program. It is also possible to import existing lists, and export lists gathered in Babbletower.

Vocabulary lists can be studied through quizzes, while the program keeps track of your progress, e.g. new entries will appear more frequently while entries that have been answered correctly a sufficient number of times get sorted out. In addition, a level of importance can be assigned to each lists that will determine how often vocabulary from that list may appear during a quiz.

Things to Come

Work on Babbletower is still in progress, and there are many more ideas I'm planning to implement. Given the time I'll add a list here. The reader also includes a GUI designed for desktops and notebook computers (however, implementation is still way behind the PDA GUI). If you run Babbletower from a CF memory card (which in most cases you will because of the size of most dictionary files), you can access your dictionaries and vocabulary lists also from any Java enabled desktop or notebook computer.

Back to top