%
%        (C)1992 Institute for New Generation Computer Technology
%        $BG[I[$=$NB>$O(BCOPYRIGHT$B%U%!%$%k$r;2>H$7$F2<$5$$(B
%        (Read COPYRIGHT for detailed information.)
%
A Sentence Dividing Tool

1: Abstract

  A long sentence has many ambigiuties. SDI can resolve  ambiguity of a
long compilcated  Japanese   sentence by dividing  it at   a conjunctive
particle into some simple sentences.


*SDI can divide a sentence at 13 conjunctive particles.
*When the sentence has some conjunctive particles,
  preference rule decides at which conjunctive particle it should be divided.
*Tense rule decides tenses of the devided sentences.
*SDI can divide surface sentence by using LAX for morphological analysis.

2: Environment

	OS version/ language version: UNIX, CESP-V3.2
	machine: Sparc station
	LAX system % ICOT Free Software
	
3: Files

  All files contain Japanease character codes(EUC).

  ./sdiv/doc

	README
	README.j
	INSTALL
	INSTALL.j
	class.sho
	man0.sho
	man1.sho
	man2.sho
	man3.sho

  ./sdiv/src

     (main programs)
	convertVer1.1_r.esp
	convert_dict_r.esp
	division_r.esp
	priolity_r.esp
	sent_check_r.esp
	tense_r.esp
	tense_rule_r.esp
	tstdat_r.esp

     (utility programs to use LAX system)
	conv_tblV3.esp
	execV3.esp
	laxsdiV3.esp
	patternV3.esp
	ppoutV3.esp

4: Brief Explanation on Install

	(1) Invoke CESP system
	(2) Load (or install) LAX system
	(3) Catalogue all source files with CESP librarian
	(4) Save all files with CESP librarian

5: Information on Transportation

	This software is the trasported one.

6: Reference

	Nothing

7: Operation

7.1 Execution

(1) If you use LAX standard dictionary:

  (a) get LAX instance

	??- :init(#exec,O).  
	O = $lex_engine

  (b) Input class method :division/4

	:division(#laxsdi,LAX,Sent,Rslt).

     (example)

	:division(#laxsdi,O,"$BEXNO$7$?7k2L@.@S$,>e$,$C$?!#(B",D03),
	D03 = {"$BEXNO$7$?7k2L@.@S$,>e$,$C$?!#(B",
               "$BEXNO$7$?!#$=$N7k2L!"@.@S$,>e$,$C$?!#(B"}

(2) If you use SDIV standalone:

	please see SDIV tool manual.

7.2 At present, SDIV (with LAX) can divide the following sentences:


(1) Only the following 11 sentences can be divided:

	$B$N$G(B(node)
	$B$?7k2L(B(ta-kekka)
	$B$1$l$I$b(B(keredomo)
	$B$H$-$K(B(toki-ni)
	$B$N$G$O$J$/(B(no-de-ha-naku)
	$B$@$1$G$J$/(B(dake-de-naku)
	$B$K$D$l$F(B(ni-ture-te)
	$B$F$+$i(B(te-kara)
	$B$?>e$G(B(ta-ue-de)
	$B$H(B(to)
	$B$H$$$&$h$&$K(B(to-iu-you-ni)

(2) The following two senetnces cannot be divided
     because of the LAX system:

	$B$"$^$j(B(amari)
	$B$+$o$j$K(B(kawari-ni)
				
	

