This note came with the original data:

The directory "dfki" contains the sentence test suite of DiTo, our
linguistic database for diagnostics in the syntax components of NLP
systems.

The files contain the following data:

s-all:  all sentences of our test suite (about 1270)
	the sum of s-pos and s-neg (without sentence-number)
	the sum of s-vg, s-coord, s-fvs, and s-rel (without sentence-number) 

s-all1: all sentences of our test suite with sentence-number
	(the same as s-all but with sentence-number)

s-pos:  all grammatical sentences with sentence-number (about 580)

s-neg:  all ungrammatical sentences with sentence-number  (about 690)

s-vg:   all sentences of the syntax phenomenon "verbal government"
	with sentence-number (about 430)

s-coord: all sentences of the syntax phenomenon "sentential
	 coordination" with sentence-number (about 530)

s-fvs:   all sentences of the syntax phenomenon "semi-idiomatic verbal
	 constructions" with sentence-number (Funktionsverbgefuege; about 160)

s-rel:   all sentences of the syntax phenomenon "relativ clauses" with
	 sentence number (about 150)

For the ECI, we merged all this information into one file ger01a.eci The
different kinds of sentences are now marked as attributes on the <S>,
these are:

NAME	VALUES		 DEFAULT	MEANING
gram	(yes|no)	 yes		Is sentence grammatical?
rel	(rel|notRel)	 notRel		sentences with "relative clauses" 
vg     	(vg|notVg)       notVg		sentences with "verbal government"  
coord  	(coord|notCoord) notCoord	sentences with "sentential 
					coordination " 
fvs    	(fvs|notFvs)     notFvs  	sentences with "semi-idiomatic verbal
					constructions"

This requires that the <S> element is re-defined from the ECI.DTD, in ger01.ent.

More information about the database and the data are found in:

John Nerbonne, Klaus Netter, Adbel Kader Diagne, Ludwig Dickmann,
Judith Klein: A Diagnostic Tool for German Syntax. DFKI Research
Report RR-91-18 Saarbruecken 1991

Judith Klein und Ludwig Dickmann: DiTo-Datenbank. Datendokumentation
zu Verbrektion und Koordinatin. DFKI Document D-92-04 Saarbruecken
1992

Brigitte Krenn und Martin Volk: DiTo-Datenbank. Datendokumentation zu
Funktionsverbgefuegen und Relativsaetzen. To appear as DFKI-Document
Saarbruecken 1992

For information contact:

Klaus Netter					Judith Klein
DFKI 						DFKI
Stuhlsatzenhausweg 3				Stuhlsatzenhausweg 3
6600 Saarbruecken				6600 Saarbruecken	
(+49 681) 302-5283				(+49 681) 302-5303
netter@dfki.uni-sb.de				klein@dfki.uni-sb.de