Kanji Dictionary

• its various Japanese pronunciations,
• its mnemonic "meaning" as an element (radical) of other kanji characters
• all the characters that look similar
• all the characters which form the character in question, i.e. are elements of it
• a link to the group of characters having similar semantic associations, e.g. from the "spring" to the four seasons.
• some examples of its usage in compounds, especially in those cases where two characters, of which the other or both have alternative meanings, are juxtaposed to emphasize the common meaning they both have. E.g. the way the compound for "SEKAI" (the "world") is formed
• various Ad Hoc "See Also" links, for example, to the antonym of the character, or to any other character I have considered to be somehow related with the character in the question.

In addition to the navigational mode of browsing, any specific character can be looked up by its English meaning, various Japanese pronounciations, its JIS or other index, or by specifying one or more standard 'radicals' (elements) of the character. The database, which consists of over 2000 characters and their meanings, composition and usage examples, is represented with Lisp-style linked lists.

There are two implementations of this software

1. The old bare MS-DOS version coded in C (about 14,000 lines) and assembly code (1000 lines) which had fancy things like my own primitive windowing system (no space for Windows in my machine), and a limited ability to add new user commands (to the "kanji shell" the software formed) by compiling new overlays to be dynamically loaded onto the basic platform. E.g. there were an overlay module for browsing any JIS text file, showing each KANJI and KANA character as a link to that character's definition, and there were another one for converting arabic numbers to the kanji notation.

   This was all coded at 1989, when I certainly had never heard of HTML or Web, and I had only a dim inkling of the hypertext in general.

   Unfortunately, this version is nowadays very much obsolete, and I am not even sure whether I have any working set of the sources left.

   
   

2. The new Web implementation, called Kanji Server, coded mostly with the good old Franz Lisp (nowadays freely available) and with some C, running on an old MicroVax under 4.3 BSD Unix. Kanji Server is implemented as its own special purpose Web-server, listening to and parsing by itself all the HTTP-requests coming from clients.

   For this I just recoded in Lisp some C modules of the old MS-DOS version and added few C and Lisp modules for handling the HTTP protocol which is needed when communicating directly with the client browsers.

   The database is essentially the same one as used in the old MS-DOS version, being built of nested Lisp lists and expressions. The main task of Kanji Server is at the starting time (right after the boot) to build convenient, RAM-based indices of this data, and then respond to the queries sent by the clients.

   For each received valid request Kanji Server will dynamically generate a HTML-page with a standard layout but its contents reflecting the requested entry.

   Currently only a limited form of keyword queries, and queries by the index number are implemented in this version.

   The 7802 kanji and other characters of the JIS levels 1 and 2 are all collected in the GIF-format (with 24x24 pixel font) into a single huge file of about 900,000 bytes. The <IMG SRC> tags of HTML pages generated by Kanji Server have hyperlinks to a small cgi-script having as its argument the JIS-index of the required character. This script will then extract the character's GIF file out of that huge file, and send it to the client's browser, just like there had been a normal <IMG SRC> reference to an ordinary GIF file.

   History and Future Plans

   This is the first Web-application I have coded. I built the whole system in about two weeks time in late 1994, soon after which I came to Infosto R&D to build a new WWWGATE interface for the ETEP-server, software which also is mostly written with the Lisp language. There I found many features used in Kanji Server to be useful also in this new project, things like listening to and handling the HTTP requests, and dynamical generation of HTML-pages from the underlying data structures stored in wholly different format.

References

• Back to my main page
• St. Vitus Lisp (A Lisp Interpreter having common roots with the old MS-DOS Kanji Dictionary)

NOTE ADDED AT 8. May 1997