This tool helps Tibetan to English translators partially automate the translation process. In the Tibetan language, the boundaries of individual words are not marked in any manner such as the way in which spaces separate and mark words in English. Instead, there is a punctuation mark called a "tsheg" which separates each syllable. Thus while syllabic boundaries are utterly explicit, word boundaries are often unclear. One of the main difficulties beginning students thus have with translating Tibetan texts is figuring out where each word ends and the next word starts, and determining what series of syllables to look up in the dictionary either as constituting a single word or a larger compound phrase. This entails a very time consuming process of looking up multiple combinations of syllables to determine which are found within a given dictionary.

It partially automates that process by breaking up a sentence/paragraph entered in Extended Wylie or Tibetan script into the biggest component parts it can find in multiple dictionary databases. Then for each component part found, it displays its stored definitions and relevant information. This will thus often yield only the definition of a long phrase, rather than its component words, but one can also search for the syllables of that phrase one by one separately.

The tool can run on-line through a:

• Java servlet (using Roman script for input and Tibetan script for output) directly on a browser

  The text is typed (or pasted) using Extended Wylie in a text box within a form. All of the processing is done on the server, and the results are returned in plain HTML. This allows the user to run this version on even the most basic browser without needing any additional software installed. Also, because the results are returned in HTML, features of HTML like hyperlinks, tables, and text formatting allow it to be skimmed more easily. The user can choose between seeing the Tibetan within the results in Extended Wylie or in Tibetan script (using Tibetan Machine Web font now available for free).
   

• Java applet & application (using Tibetan script for both input and output) communicating to a servlet

  The text is typed in Extended Wylie, but with the added value that optionally the user can choose to see it directly in the Tibetan script (using Tibetan Machine Web font now available for free) as he types. We eventually plan to support other keyboard methods of entry as well. Here all the processing is also done on the server side, and the results are displayed interactively within the program's window. Again the user can choose to see the results in Extended Wylie or in the Tibetan script. 

  Even though the application runs as a stand-alone application in the desktop's user, connection to the Internet is still necessary to access the dictionary databases. Easy launching of the application can be done over the Internet using Java Web Start, which comes with Sun's Java Runtime Environment version 1.4 or higher. This is the recommended way to run the tool.

  The applet runs within a browser. The browser not only needs to support Java, but since the classes that handle the Tibetan font use Swing, Sun's Java Runtime Environment version 1.4 or higher must additionally be installed.
   

The tool can also run off-line in:

• Desktop & laptop computers supporting the Sun's Java Runtime Environment version 1.2 or higher; although version 1.4 or higher is recommended. This is distributed as DictionarySearchStandalone.jar.
• Handheld devices supporting PersonalJava Application Environment version 1.2a or higher. This is distributed as DictionarySearchHandheld.jar.

The classes designed to be run from the command-line are:

• BinaryFileGenerator (included only in DictionarySearchStandalone.jar)
• AcipToWylie (included only in DictionarySearchStandalone.jar)
• WindowScannerFilter (included in both DictionarySearchStandalone.jar and DictionarySearchHandheld.jar)
• ConsoleScannerFilter (included in both DictionarySearchStandalone.jar and DictionarySearchHandheld.jar)

Notes on Input:

• For the "punctuation marks", the tool assumes that
  • '  ' (tsheg), '_' (space), <enter>, <tab>: function as syllable separators and may show up in between component word or phrases.
  • '/' (shad), ';', '|', '!', ':', '[', ']', '^', '@', '#', '$', '%', '=', '<', '>', '(', ')', '{', '}', blank line (two enters in a row): may not show up in between component word or phrases (and hence is interpreted as marking the end of a component word or phrase). See Extended Wylie documentation for the corresponding symbols in the Tibetan script.
  • all other characters are part of the syllable
    
• To force the parser to "break up" a component word or phrase into its individual syllables, use any character of the second set in between the syllables. For example, if the entry is:

  chos nyid

  or

  chos
  nyid

  the parser will recognize it as a single word "chos nyid". But if the entry is:

  chos / nyid

  or

  chos

  nyid

  the parser will assume "chos" and "nyid" are independent, and will be looked up separately.

Author: Andrés Montano Pellegrini

Related Documentation