Introduction to the Parser

This is designed to serve as an introduction to the work done on developing a parser for finite Akkadian verbs, with the parsing engine written in Prolog. The parser itself is written as a collection of Definite Clause Grammar (DCG) rules, which are then wrapped in standard Prolog definitions. The Prolog program itself is then interfaced to the user using HTML forms for both parsing and generation, part of the same program as Prolog programs have no definite sense of direction. The program is run locally on a dual Xeon system, 2.4GHz with 4Gig of RAM (See system stats). It can parse approximately 10 forms per second per processor. The Parser/Generator if free for use, but is copyrighted 2001-2006 by Aaron Macks and such should be noted if it is used.

The program can recognize strong and singly-weak verbs in the G, D, and N stems in the preterite, durative, precative, vetitive, imperative and perfect. It can also recognize, and ignore, the conjunctive -ma suffix and subset of dative, accusative and vantive suffixes. Due to their rarity and irregularity, it cannot recognize quadraliteral or multi-weak forms. The parser can also accept wildcard vowel lengths, testing a, ä and â for the string a*, with ä used to represent double length vowels due to the difficulty of representing the macron in HTML. It will respond with any possible valid parsing for the input string, where the length of vowels is represented by repetition, u=u, uu=ü and uuu=û..

The generator can create any of the forms, but cannot yet create suffixes. It takes input in the form of a string of the radicals, with @ representing an aleph and $ a shin, i.e., prs for parasum, qb@ for qabûm. The conjunctive suffix -ma is properly ignored by the parser, due to it commonly being present on the end of verbs.

Current progress, especially bug fixes, are detailed on the Current Activity page.

Created and maintained by Aaron Macks, aaronm(at) Last Modified Jan 30, 2006