id	sid	tid	token	lemma	pos
8961	1	1	The	the	DT
8961	1	2	Binary	Binary	NNP
8961	1	3	Vector	Vector	NNP
8961	1	4	as	as	IN
8961	1	5	the	the	DT
8961	1	6	Basis	basis	NN
8961	1	7	of	of	IN
8961	1	8	an	an	DT
8961	1	9	Inverted	Inverted	NNP
8961	1	10	Index	Index	NNP
8961	1	11	File	File	NNP
8961	1	12	Donald	Donald	NNP
8961	1	13	R.	R.	NNP
8961	1	14	KING	KING	NNP
8961	1	15	:	:	:
8961	1	16	Rutgers	Rutgers	NNP
8961	1	17	University	University	NNP
8961	1	18	,	,	,
8961	1	19	New	New	NNP
8961	1	20	Brunswick	Brunswick	NNP
8961	1	21	,	,	,
8961	1	22	New	New	NNP
8961	1	23	Jersey	Jersey	NNP
8961	1	24	.	.	.
8961	2	1	307	307	CD
8961	2	2	The	the	DT
8961	2	3	inverted	inverted	JJ
8961	2	4	index	index	NN
8961	2	5	file	file	NN
8961	2	6	is	be	VBZ
8961	2	7	a	a	DT
8961	2	8	frequently	frequently	RB
8961	2	9	used	use	VBN
8961	2	10	file	file	NN
8961	2	11	structure	structure	NN
8961	2	12	for	for	IN
8961	2	13	the	the	DT
8961	2	14	storage	storage	NN
8961	2	15	of	of	IN
8961	2	16	indexing	indexing	NN
8961	2	17	information	information	NN
8961	2	18	in	in	IN
8961	2	19	a	a	DT
8961	2	20	document	document	JJ
8961	2	21	retrieval	retrieval	NN
8961	2	22	system	system	NN
8961	2	23	.	.	.
8961	3	1	This	this	DT
8961	3	2	paper	paper	NN
8961	3	3	de-	de-	CC
8961	3	4	scribes	scribe	VBZ
8961	3	5	a	a	DT
8961	3	6	novel	novel	JJ
8961	3	7	method	method	NN
8961	3	8	for	for	IN
8961	3	9	the	the	DT
8961	3	10	computer	computer	NN
8961	3	11	storage	storage	NN
8961	3	12	of	of	IN
8961	3	13	such	such	PDT
8961	3	14	an	an	DT
8961	3	15	index	index	NN
8961	3	16	.	.	.
8961	4	1	The	the	DT
8961	4	2	method	method	NN
8961	4	3	not	not	RB
8961	4	4	only	only	RB
8961	4	5	offers	offer	VBZ
8961	4	6	the	the	DT
8961	4	7	possibility	possibility	NN
8961	4	8	of	of	IN
8961	4	9	reducing	reduce	VBG
8961	4	10	storage	storage	NN
8961	4	11	requirements	requirement	NNS
8961	4	12	fot	fot	NNP
8961	4	13	an	an	DT
8961	4	14	index	index	NN
8961	4	15	but	but	CC
8961	4	16	also	also	RB
8961	4	17	affords	afford	VBZ
8961	4	18	more	more	JJR
8961	4	19	mpid	mpid	JJ
8961	4	20	processing	processing	NN
8961	4	21	of	of	IN
8961	4	22	query	query	NN
8961	4	23	statements	statement	NNS
8961	4	24	ex-	ex-	RB
8961	4	25	pressed	press	VBN
8961	4	26	in	in	IN
8961	4	27	Boolean	boolean	JJ
8961	4	28	logic	logic	NN
8961	4	29	.	.	.
8961	5	1	INTRODUCTION	INTRODUCTION	NNP
8961	5	2	The	the	DT
8961	5	3	inverted	invert	VBN
8961	5	4	index	index	NN
8961	5	5	file	file	NN
8961	5	6	is	be	VBZ
8961	5	7	a	a	DT
8961	5	8	frequently	frequently	RB
8961	5	9	used	use	VBN
8961	5	10	file	file	NN
8961	5	11	structure	structure	NN
8961	5	12	for	for	IN
8961	5	13	the	the	DT
8961	5	14	storage	storage	NN
8961	5	15	of	of	IN
8961	5	16	indexing	indexing	NN
8961	5	17	information	information	NN
8961	5	18	in	in	IN
8961	5	19	document	document	JJ
8961	5	20	retrieval	retrieval	NN
8961	5	21	systems	system	NNS
8961	5	22	.	.	.
8961	6	1	An	an	DT
8961	6	2	inverted	invert	VBN
8961	6	3	index	index	NN
8961	6	4	file	file	NN
8961	6	5	may	may	MD
8961	6	6	be	be	VB
8961	6	7	used	use	VBN
8961	6	8	by	by	IN
8961	6	9	itself	-PRON-	PRP
8961	6	10	or	or	CC
8961	6	11	with	with	IN
8961	6	12	a	a	DT
8961	6	13	direct	direct	JJ
8961	6	14	file	file	NN
8961	6	15	in	in	IN
8961	6	16	a	a	DT
8961	6	17	so	so	RB
8961	6	18	-	-	HYPH
8961	6	19	called	call	VBN
8961	6	20	combined	combine	VBN
8961	6	21	file	file	NN
8961	6	22	system	system	NN
8961	6	23	.	.	.
8961	7	1	The	the	DT
8961	7	2	inverted	inverted	JJ
8961	7	3	index	index	NN
8961	7	4	file	file	NN
8961	7	5	contains	contain	VBZ
8961	7	6	a	a	DT
8961	7	7	logical	logical	JJ
8961	7	8	record	record	NN
8961	7	9	for	for	IN
8961	7	10	each	each	DT
8961	7	11	of	of	IN
8961	7	12	the	the	DT
8961	7	13	subject	subject	JJ
8961	7	14	headings	heading	NNS
8961	7	15	or	or	CC
8961	7	16	index	index	NN
8961	7	17	terms	term	NNS
8961	7	18	which	which	WDT
8961	7	19	may	may	MD
8961	7	20	be	be	VB
8961	7	21	used	use	VBN
8961	7	22	to	to	TO
8961	7	23	describe	describe	VB
8961	7	24	documents	document	NNS
8961	7	25	in	in	IN
8961	7	26	the	the	DT
8961	7	27	system	system	NN
8961	7	28	.	.	.
8961	8	1	Within	within	IN
8961	8	2	each	each	DT
8961	8	3	logical	logical	JJ
8961	8	4	record	record	NN
8961	8	5	there	there	EX
8961	8	6	is	be	VBZ
8961	8	7	a	a	DT
8961	8	8	list	list	NN
8961	8	9	of	of	IN
8961	8	10	pointers	pointer	NNS
8961	8	11	to	to	IN
8961	8	12	those	those	DT
8961	8	13	documents	document	NNS
8961	8	14	which	which	WDT
8961	8	15	have	have	VBP
8961	8	16	been	be	VBN
8961	8	17	indexed	index	VBN
8961	8	18	by	by	IN
8961	8	19	the	the	DT
8961	8	20	subject	subject	NN
8961	8	21	heading	heading	NN
8961	8	22	in	in	IN
8961	8	23	question	question	NN
8961	8	24	.	.	.
8961	9	1	The	the	DT
8961	9	2	individual	individual	JJ
8961	9	3	pointers	pointer	NNS
8961	9	4	are	be	VBP
8961	9	5	usually	usually	RB
8961	9	6	in	in	IN
8961	9	7	the	the	DT
8961	9	8	form	form	NN
8961	9	9	of	of	IN
8961	9	10	document	document	NN
8961	9	11	numbers	number	NNS
8961	9	12	stored	store	VBN
8961	9	13	in	in	IN
8961	9	14	fixed	fix	VBN
8961	9	15	-	-	HYPH
8961	9	16	length	length	NN
8961	9	17	digital	digital	JJ
8961	9	18	form	form	NN
8961	9	19	.	.	.
8961	10	1	Obviously	obviously	RB
8961	10	2	,	,	,
8961	10	3	the	the	DT
8961	10	4	length	length	NN
8961	10	5	of	of	IN
8961	10	6	the	the	DT
8961	10	7	lists	list	NNS
8961	10	8	will	will	MD
8961	10	9	vary	vary	VB
8961	10	10	from	from	IN
8961	10	11	record	record	NN
8961	10	12	to	to	IN
8961	10	13	record	record	NN
8961	10	14	.	.	.
8961	11	1	The	the	DT
8961	11	2	purpose	purpose	NN
8961	11	3	of	of	IN
8961	11	4	this	this	DT
8961	11	5	paper	paper	NN
8961	11	6	is	be	VBZ
8961	11	7	the	the	DT
8961	11	8	presentation	presentation	NN
8961	11	9	of	of	IN
8961	11	10	a	a	DT
8961	11	11	new	new	JJ
8961	11	12	technique	technique	NN
8961	11	13	for	for	IN
8961	11	14	the	the	DT
8961	11	15	storage	storage	NN
8961	11	16	of	of	IN
8961	11	17	the	the	DT
8961	11	18	lists	list	NNS
8961	11	19	of	of	IN
8961	11	20	pointers	pointer	NNS
8961	11	21	to	to	IN
8961	11	22	documents	document	NNS
8961	11	23	.	.	.
8961	12	1	It	-PRON-	PRP
8961	12	2	will	will	MD
8961	12	3	be	be	VB
8961	12	4	shown	show	VBN
8961	12	5	that	that	IN
8961	12	6	this	this	DT
8961	12	7	technique	technique	NN
8961	12	8	not	not	RB
8961	12	9	only	only	RB
8961	12	10	reduces	reduce	VBZ
8961	12	11	storage	storage	NN
8961	12	12	requirements	requirement	NNS
8961	12	13	,	,	,
8961	12	14	but	but	CC
8961	12	15	that	that	IN
8961	12	16	in	in	IN
8961	12	17	many	many	JJ
8961	12	18	cases	case	NNS
8961	12	19	the	the	DT
8961	12	20	time	time	NN
8961	12	21	required	require	VBN
8961	12	22	to	to	TO
8961	12	23	search	search	VB
8961	12	24	the	the	DT
8961	12	25	index	index	NN
8961	12	26	is	be	VBZ
8961	12	27	reduced	reduce	VBN
8961	12	28	.	.	.
8961	13	1	The	the	DT
8961	13	2	technique	technique	NN
8961	13	3	is	be	VBZ
8961	13	4	useful	useful	JJ
8961	13	5	in	in	IN
8961	13	6	systems	system	NNS
8961	13	7	which	which	WDT
8961	13	8	use	use	VBP
8961	13	9	Boolean	boolean	JJ
8961	13	10	searches	search	NNS
8961	13	11	.	.	.
8961	14	1	The	the	DT
8961	14	2	relative	relative	JJ
8961	14	3	merits	merit	NNS
8961	14	4	of	of	IN
8961	14	5	Boolean	boolean	JJ
8961	14	6	and	and	CC
8961	14	7	weighted	weighted	JJ
8961	14	8	term	term	NN
8961	14	9	searches	search	NNS
8961	14	10	are	be	VBP
8961	14	11	beyond	beyond	IN
8961	14	12	the	the	DT
8961	14	13	scope	scope	NN
8961	14	14	of	of	IN
8961	14	15	this	this	DT
8961	14	16	paper	paper	NN
8961	14	17	,	,	,
8961	14	18	as	as	IN
8961	14	19	are	be	VBP
8961	14	20	the	the	DT
8961	14	21	rela-	rela-	JJ
8961	14	22	tive	tive	JJ
8961	14	23	merits	merit	NNS
8961	14	24	of	of	IN
8961	14	25	the	the	DT
8961	14	26	various	various	JJ
8961	14	27	possible	possible	JJ
8961	14	28	file	file	NN
8961	14	29	structures	structure	NNS
8961	14	30	.	.	.
8961	15	1	THE	the	DT
8961	15	2	BINARY	BINARY	NNP
8961	15	3	VECTOR	VECTOR	NNP
8961	15	4	AS	as	IN
8961	15	5	A	a	DT
8961	15	6	STORAGE	storage	NN
8961	15	7	DEVICE	device	NN
8961	15	8	The	the	DT
8961	15	9	exact	exact	JJ
8961	15	10	form	form	NN
8961	15	11	of	of	IN
8961	15	12	each	each	DT
8961	15	13	document	document	NN
8961	15	14	pointer	pointer	NN
8961	15	15	is	be	VBZ
8961	15	16	immaterial	immaterial	JJ
8961	15	17	to	to	IN
8961	15	18	the	the	DT
8961	15	19	user	user	NN
8961	15	20	of	of	IN
8961	15	21	a	a	DT
8961	15	22	document	document	JJ
8961	15	23	retrieval	retrieval	NN
8961	15	24	system	system	NN
8961	15	25	as	as	RB
8961	15	26	long	long	RB
8961	15	27	as	as	IN
8961	15	28	he	-PRON-	PRP
8961	15	29	is	be	VBZ
8961	15	30	able	able	JJ
8961	15	31	to	to	TO
8961	15	32	obtain	obtain	VB
8961	15	33	the	the	DT
8961	15	34	document	document	NN
8961	15	35	he	-PRON-	PRP
8961	15	36	desires	desire	VBZ
8961	15	37	.	.	.
8961	16	1	The	the	DT
8961	16	2	standard	standard	JJ
8961	16	3	form	form	NN
8961	16	4	for	for	IN
8961	16	5	these	these	DT
8961	16	6	pointers	pointer	NNS
8961	16	7	in	in	IN
8961	16	8	most	most	RBS
8961	16	9	automated	automate	VBN
8961	16	10	systems	system	NNS
8961	16	11	is	be	VBZ
8961	16	12	a	a	DT
8961	16	13	document	document	NN
8961	16	14	number	number	NN
8961	16	15	.	.	.
8961	17	1	Note	note	VB
8961	17	2	that	that	IN
8961	17	3	each	each	DT
8961	17	4	pointer	pointer	NN
8961	17	5	is	be	VBZ
8961	17	6	by	by	IN
8961	17	7	itself	-PRON-	PRP
8961	17	8	a	a	DT
8961	17	9	piece	piece	NN
8961	17	10	of	of	IN
8961	17	11	infor-	infor-	JJ
8961	17	12	mation	mation	NN
8961	17	13	.	.	.
8961	18	1	However	however	RB
8961	18	2	,	,	,
8961	18	3	if	if	IN
8961	18	4	one	one	PRP
8961	18	5	thinks	think	VBZ
8961	18	6	of	of	IN
8961	18	7	a	a	DT
8961	18	8	"	"	``
8961	18	9	peek	peek	JJ
8961	18	10	-	-	HYPH
8961	18	11	a	a	DT
8961	18	12	-	-	HYPH
8961	18	13	boo	boo	NN
8961	18	14	"	"	''
8961	18	15	system	system	NN
8961	18	16	,	,	,
8961	18	17	the	the	DT
8961	18	18	document	document	NN
8961	18	19	308	308	CD
8961	18	20	Journal	Journal	NNP
8961	18	21	of	of	IN
8961	18	22	Library	Library	NNP
8961	18	23	Automation	Automation	NNP
8961	18	24	Vol	Vol	NNP
8961	18	25	.	.	.
8961	19	1	7/4	7/4	CD
8961	19	2	December	December	NNP
8961	19	3	1974	1974	CD
8961	19	4	pointer	pointer	NN
8961	19	5	becomes	become	VBZ
8961	19	6	simply	simply	RB
8961	19	7	a	a	DT
8961	19	8	hole	hole	NN
8961	19	9	punched	punch	VBN
8961	19	10	in	in	IN
8961	19	11	a	a	DT
8961	19	12	card	card	NN
8961	19	13	.	.	.
8961	20	1	In	in	IN
8961	20	2	this	this	DT
8961	20	3	case	case	NN
8961	20	4	the	the	DT
8961	20	5	position	position	NN
8961	20	6	of	of	IN
8961	20	7	the	the	DT
8961	20	8	pointer	pointer	NN
8961	20	9	,	,	,
8961	20	10	not	not	RB
8961	20	11	the	the	DT
8961	20	12	pointer	pointer	NN
8961	20	13	itself	-PRON-	PRP
8961	20	14	,	,	,
8961	20	15	conveys	convey	VBZ
8961	20	16	the	the	DT
8961	20	17	information	information	NN
8961	20	18	.	.	.
8961	21	1	The	the	DT
8961	21	2	new	new	JJ
8961	21	3	technique	technique	NN
8961	21	4	presented	present	VBN
8961	21	5	in	in	IN
8961	21	6	this	this	DT
8961	21	7	paper	paper	NN
8961	21	8	is	be	VBZ
8961	21	9	an	an	DT
8961	21	10	extension	extension	NN
8961	21	11	of	of	IN
8961	21	12	the	the	DT
8961	21	13	"	"	``
8961	21	14	peek-	peek-	JJ
8961	21	15	a	a	DT
8961	21	16	-	-	HYPH
8961	21	17	boo	boo	NN
8961	21	18	"	"	''
8961	21	19	concept	concept	NN
8961	21	20	.	.	.
8961	22	1	A	a	DT
8961	22	2	vector	vector	NN
8961	22	3	or	or	CC
8961	22	4	string	string	NN
8961	22	5	of	of	IN
8961	22	6	binary	binary	JJ
8961	22	7	zeroes	zero	NNS
8961	22	8	is	be	VBZ
8961	22	9	constructed	construct	VBN
8961	22	10	equal	equal	JJ
8961	22	11	in	in	IN
8961	22	12	length	length	NN
8961	22	13	to	to	IN
8961	22	14	the	the	DT
8961	22	15	number	number	NN
8961	22	16	of	of	IN
8961	22	17	documents	document	NNS
8961	22	18	expected	expect	VBN
8961	22	19	in	in	IN
8961	22	20	the	the	DT
8961	22	21	system	system	NN
8961	22	22	.	.	.
8961	23	1	The	the	DT
8961	23	2	position	position	NN
8961	23	3	of	of	IN
8961	23	4	each	each	DT
8961	23	5	vector	vector	NN
8961	23	6	element	element	NN
8961	23	7	corresponds	correspond	VBZ
8961	23	8	to	to	IN
8961	23	9	a	a	DT
8961	23	10	document	document	NN
8961	23	11	number	number	NN
8961	23	12	.	.	.
8961	24	1	That	that	RB
8961	24	2	is	is	RB
8961	24	3	,	,	,
8961	24	4	the	the	DT
8961	24	5	first	first	JJ
8961	24	6	position	position	NN
8961	24	7	in	in	IN
8961	24	8	a	a	DT
8961	24	9	vector	vector	NN
8961	24	10	corresponds	correspond	VBZ
8961	24	11	to	to	TO
8961	24	12	document	document	VB
8961	24	13	number	number	NN
8961	24	14	one	one	CD
8961	24	15	and	and	CC
8961	24	16	the	the	DT
8961	24	17	tenth	tenth	JJ
8961	24	18	vector	vector	NN
8961	24	19	posi-	posi-	NN
8961	24	20	tion	tion	NN
8961	24	21	corresponds	correspond	NNS
8961	24	22	to	to	TO
8961	24	23	document	document	VB
8961	24	24	number	number	NN
8961	24	25	ten	ten	CD
8961	24	26	.	.	.
8961	25	1	A	a	DT
8961	25	2	vector	vector	NN
8961	25	3	is	be	VBZ
8961	25	4	constructed	construct	VBN
8961	25	5	for	for	IN
8961	25	6	each	each	DT
8961	25	7	subject	subject	JJ
8961	25	8	heading	heading	NN
8961	25	9	in	in	IN
8961	25	10	the	the	DT
8961	25	11	system	system	NN
8961	25	12	.	.	.
8961	26	1	As	as	IN
8961	26	2	a	a	DT
8961	26	3	document	document	NN
8961	26	4	enters	enter	VBZ
8961	26	5	the	the	DT
8961	26	6	system	system	NN
8961	26	7	,	,	,
8961	26	8	ones	one	NNS
8961	26	9	are	be	VBP
8961	26	10	in-	in-	RB
8961	26	11	serted	serte	VBN
8961	26	12	in	in	IN
8961	26	13	place	place	NN
8961	26	14	of	of	IN
8961	26	15	the	the	DT
8961	26	16	zeroes	zero	NNS
8961	26	17	in	in	IN
8961	26	18	the	the	DT
8961	26	19	positions	position	NNS
8961	26	20	corresponding	correspond	VBG
8961	26	21	to	to	IN
8961	26	22	the	the	DT
8961	26	23	new	new	JJ
8961	26	24	doc-	doc-	NN
8961	26	25	ument	ument	NN
8961	26	26	number	number	NN
8961	26	27	in	in	IN
8961	26	28	the	the	DT
8961	26	29	vectors	vector	NNS
8961	26	30	for	for	IN
8961	26	31	the	the	DT
8961	26	32	subject	subject	JJ
8961	26	33	headings	heading	NNS
8961	26	34	used	use	VBN
8961	26	35	to	to	TO
8961	26	36	describe	describe	VB
8961	26	37	the	the	DT
8961	26	38	document	document	NN
8961	26	39	.	.	.
8961	27	1	As	as	IN
8961	27	2	an	an	DT
8961	27	3	example	example	NN
8961	27	4	,	,	,
8961	27	5	assume	assume	VB
8961	27	6	the	the	DT
8961	27	7	following	follow	VBG
8961	27	8	document	document	NN
8961	27	9	descriptions	description	NNS
8961	27	10	are	be	VBP
8961	27	11	presented	present	VBN
8961	27	12	to	to	IN
8961	27	13	a	a	DT
8961	27	14	system	system	NN
8961	27	15	using	use	VBG
8961	27	16	binary	binary	JJ
8961	27	17	vectors	vector	NNS
8961	27	18	:	:	:
8961	27	19	Document	document	NN
8961	27	20	Number	number	NN
8961	27	21	1	1	CD
8961	27	22	2	2	CD
8961	27	23	3	3	CD
8961	27	24	Subject	Subject	NNP
8961	27	25	Headings	Headings	NNPS
8961	27	26	A	a	NN
8961	27	27	,	,	,
8961	27	28	B	b	NN
8961	27	29	,	,	,
8961	27	30	D	D	NNP
8961	27	31	C	C	NNP
8961	27	32	,	,	,
8961	27	33	E	E	NNP
8961	27	34	A	a	NN
8961	27	35	,	,	,
8961	27	36	C	c	NN
8961	27	37	The	the	DT
8961	27	38	binary	binary	JJ
8961	27	39	vectors	vector	NNS
8961	27	40	for	for	IN
8961	27	41	terms	term	NNS
8961	27	42	A	a	NN
8961	27	43	,	,	,
8961	27	44	B	b	NN
8961	27	45	,	,	,
8961	27	46	C	C	NNP
8961	27	47	,	,	,
8961	27	48	D	d	NN
8961	27	49	,	,	,
8961	27	50	and	and	CC
8961	27	51	E	e	NN
8961	27	52	before	before	IN
8961	27	53	the	the	DT
8961	27	54	insertion	insertion	NN
8961	27	55	of	of	IN
8961	27	56	the	the	DT
8961	27	57	indexing	indexing	NN
8961	27	58	data	datum	NNS
8961	27	59	would	would	MD
8961	27	60	be	be	VB
8961	27	61	as	as	IN
8961	27	62	follows	follow	VBZ
8961	27	63	:	:	:
8961	27	64	Subject	subject	JJ
8961	27	65	Heading	head	VBG
8961	27	66	A	a	NN
8961	27	67	B	b	NN
8961	27	68	c	c	NN
8961	27	69	D	d	NN
8961	27	70	E	e	NN
8961	27	71	Vector	vector	NN
8961	27	72	000	000	CD
8961	27	73	...	...	NFP
8961	27	74	0	0	CD
8961	27	75	000	000	CD
8961	27	76	...	...	NFP
8961	27	77	0	0	CD
8961	27	78	000	000	CD
8961	27	79	...	...	NFP
8961	27	80	0	0	CD
8961	27	81	000	000	CD
8961	27	82	...	...	NFP
8961	27	83	0	0	NFP
8961	27	84	ooo	ooo	NN
8961	27	85	...	...	NFP
8961	27	86	·	·	NFP
8961	27	87	o	o	XX
8961	27	88	After	after	IN
8961	27	89	the	the	DT
8961	27	90	insertion	insertion	NN
8961	27	91	of	of	IN
8961	27	92	the	the	DT
8961	27	93	indexing	indexing	NN
8961	27	94	information	information	NN
8961	27	95	,	,	,
8961	27	96	the	the	DT
8961	27	97	same	same	JJ
8961	27	98	vectors	vector	NNS
8961	27	99	would	would	MD
8961	27	100	appear	appear	VB
8961	27	101	as	as	IN
8961	27	102	follows	follow	VBZ
8961	27	103	:	:	:
8961	27	104	Subject	subject	JJ
8961	27	105	Heading	head	VBG
8961	27	106	A	a	NN
8961	27	107	B	b	NN
8961	27	108	c	c	NN
8961	27	109	D	d	NN
8961	27	110	E	e	NN
8961	27	111	Vector	vector	NN
8961	27	112	101	101	CD
8961	27	113	...	...	SYM
8961	27	114	0	0	CD
8961	27	115	100	100	CD
8961	27	116	...	...	SYM
8961	27	117	0	0	CD
8961	27	118	011	011	CD
8961	27	119	...	...	NFP
8961	27	120	0	0	CD
8961	27	121	100	100	CD
8961	27	122	...	...	SYM
8961	27	123	0	0	CD
8961	27	124	010	010	CD
8961	27	125	...	...	NFP
8961	27	126	0	0	NFP
8961	27	127	The	the	DT
8961	27	128	binary	binary	JJ
8961	27	129	vector	vector	NN
8961	27	130	seems	seem	VBZ
8961	27	131	to	to	TO
8961	27	132	have	have	VB
8961	27	133	several	several	JJ
8961	27	134	advantages	advantage	NNS
8961	27	135	over	over	IN
8961	27	136	the	the	DT
8961	27	137	standard	standard	JJ
8961	27	138	form	form	NN
8961	27	139	of	of	IN
8961	27	140	storage	storage	NN
8961	27	141	of	of	IN
8961	27	142	document	document	NN
8961	27	143	numbers	number	NNS
8961	27	144	in	in	IN
8961	27	145	an	an	DT
8961	27	146	inverted	invert	VBN
8961	27	147	file	file	NN
8961	27	148	.	.	.
8961	28	1	First	first	RB
8961	28	2	,	,	,
8961	28	3	the	the	DT
8961	28	4	rec-	rec-	NN
8961	28	5	ords	ord	NNS
8961	28	6	are	be	VBP
8961	28	7	of	of	IN
8961	28	8	fixed	fix	VBN
8961	28	9	length	length	NN
8961	28	10	since	since	IN
8961	28	11	the	the	DT
8961	28	12	vectors	vector	NNS
8961	28	13	are	be	VBP
8961	28	14	all	all	RB
8961	28	15	equal	equal	JJ
8961	28	16	in	in	IN
8961	28	17	length	length	NN
8961	28	18	to	to	IN
8961	28	19	the	the	DT
8961	28	20	ex-	ex-	XX
8961	28	21	pected	pecte	VBN
8961	28	22	number	number	NN
8961	28	23	of	of	IN
8961	28	24	documents	document	NNS
8961	28	25	in	in	IN
8961	28	26	the	the	DT
8961	28	27	system	system	NN
8961	28	28	.	.	.
8961	29	1	Space	space	NN
8961	29	2	may	may	MD
8961	29	3	be	be	VB
8961	29	4	left	leave	VBN
8961	29	5	at	at	IN
8961	29	6	the	the	DT
8961	29	7	end	end	NN
8961	29	8	of	of	IN
8961	29	9	each	each	DT
8961	29	10	vector	vector	NN
8961	29	11	for	for	IN
8961	29	12	the	the	DT
8961	29	13	addition	addition	NN
8961	29	14	of	of	IN
8961	29	15	new	new	JJ
8961	29	16	documents	document	NNS
8961	29	17	.	.	.
8961	30	1	Periodic	periodic	JJ
8961	30	2	copying	copying	NN
8961	30	3	of	of	IN
8961	30	4	the	the	DT
8961	30	5	file	file	NN
8961	30	6	may	may	MD
8961	30	7	be	be	VB
8961	30	8	used	use	VBN
8961	30	9	to	to	TO
8961	30	10	expand	expand	VB
8961	30	11	the	the	DT
8961	30	12	index	index	NN
8961	30	13	records	record	NNS
8961	30	14	with	with	IN
8961	30	15	additional	additional	JJ
8961	30	16	zeroes	zero	NNS
8961	30	17	added	add	VBN
8961	30	18	at	at	IN
8961	30	19	the	the	DT
8961	30	20	end	end	NN
8961	30	21	of	of	IN
8961	30	22	each	each	DT
8961	30	23	record	record	NN
8961	30	24	during	during	IN
8961	30	25	the	the	DT
8961	30	26	process	process	NN
8961	30	27	.	.	.
8961	31	1	Consequently	consequently	RB
8961	31	2	,	,	,
8961	31	3	unless	unless	IN
8961	31	4	Binary	Binary	NNP
8961	31	5	Vector	Vector	NNP
8961	31	6	/	/	SYM
8961	31	7	KING	KING	NNP
8961	31	8	309	309	CD
8961	31	9	there	there	EX
8961	31	10	are	be	VBP
8961	31	11	limitations	limitation	NNS
8961	31	12	of	of	IN
8961	31	13	size	size	NN
8961	31	14	imposed	impose	VBN
8961	31	15	by	by	IN
8961	31	16	the	the	DT
8961	31	17	equipment	equipment	NN
8961	31	18	,	,	,
8961	31	19	only	only	RB
8961	31	20	one	one	CD
8961	31	21	access	access	NN
8961	31	22	to	to	IN
8961	31	23	the	the	DT
8961	31	24	storage	storage	NN
8961	31	25	device	device	NN
8961	31	26	will	will	MD
8961	31	27	be	be	VB
8961	31	28	needed	need	VBN
8961	31	29	to	to	TO
8961	31	30	retrieve	retrieve	VB
8961	31	31	the	the	DT
8961	31	32	index	index	NN
8961	31	33	record	record	NN
8961	31	34	for	for	IN
8961	31	35	a	a	DT
8961	31	36	term	term	NN
8961	31	37	.	.	.
8961	32	1	The	the	DT
8961	32	2	second	second	JJ
8961	32	3	advantage	advantage	NN
8961	32	4	offered	offer	VBN
8961	32	5	by	by	IN
8961	32	6	the	the	DT
8961	32	7	binary	binary	JJ
8961	32	8	vector	vector	NN
8961	32	9	method	method	NN
8961	32	10	appears	appear	VBZ
8961	32	11	in	in	IN
8961	32	12	the	the	DT
8961	32	13	search	search	NN
8961	32	14	process	process	NN
8961	32	15	.	.	.
8961	33	1	Most	Most	JJS
8961	33	2	modern	modern	JJ
8961	33	3	computers	computer	NNS
8961	33	4	have	have	VBP
8961	33	5	a	a	DT
8961	33	6	built	build	VBN
8961	33	7	-	-	HYPH
8961	33	8	in	in	RP
8961	33	9	capability	capability	NN
8961	33	10	of	of	IN
8961	33	11	per-	per-	NN
8961	33	12	forming	form	VBG
8961	33	13	Boolean	boolean	JJ
8961	33	14	logical	logical	JJ
8961	33	15	manipulations	manipulation	NNS
8961	33	16	on	on	IN
8961	33	17	binary	binary	JJ
8961	33	18	digit	digit	NN
8961	33	19	vectors	vector	NNS
8961	33	20	or	or	CC
8961	33	21	strings	string	NNS
8961	33	22	.	.	.
8961	34	1	Thus	thus	RB
8961	34	2	,	,	,
8961	34	3	when	when	WRB
8961	34	4	Boolean	boolean	JJ
8961	34	5	operations	operation	NNS
8961	34	6	are	be	VBP
8961	34	7	specified	specify	VBN
8961	34	8	as	as	IN
8961	34	9	part	part	NN
8961	34	10	of	of	IN
8961	34	11	a	a	DT
8961	34	12	query	query	NN
8961	34	13	,	,	,
8961	34	14	the	the	DT
8961	34	15	imple-	imple-	NN
8961	34	16	mentation	mentation	NN
8961	34	17	of	of	IN
8961	34	18	the	the	DT
8961	34	19	operations	operation	NNS
8961	34	20	within	within	IN
8961	34	21	the	the	DT
8961	34	22	·	·	NFP
8961	34	23	computer	computer	NN
8961	34	24	is	be	VBZ
8961	34	25	considerably	considerably	RB
8961	34	26	easier	easy	JJR
8961	34	27	and	and	CC
8961	34	28	faster	fast	RBR
8961	34	29	for	for	IN
8961	34	30	binary	binary	JJ
8961	34	31	vectors	vector	NNS
8961	34	32	than	than	IN
8961	34	33	for	for	IN
8961	34	34	the	the	DT
8961	34	35	standard	standard	JJ
8961	34	36	form	form	NN
8961	34	37	of	of	IN
8961	34	38	inverted	invert	VBN
8961	34	39	files	file	NNS
8961	34	40	.	.	.
8961	35	1	Other	other	JJ
8961	35	2	investigators	investigator	NNS
8961	35	3	of	of	IN
8961	35	4	the	the	DT
8961	35	5	use	use	NN
8961	35	6	of	of	IN
8961	35	7	the	the	DT
8961	35	8	binary	binary	JJ
8961	35	9	digit	digit	NN
8961	35	10	patterns	pattern	NNS
8961	35	11	or	or	CC
8961	35	12	vectors	vector	NNS
8961	35	13	have	have	VBP
8961	35	14	not	not	RB
8961	35	15	fully	fully	RB
8961	35	16	explored	explore	VBN
8961	35	17	its	-PRON-	PRP$
8961	35	18	advantages	advantage	NNS
8961	35	19	and	and	CC
8961	35	20	disadvantages	disadvantage	NNS
8961	35	21	.	.	.
8961	36	1	Bloom	bloom	NN
8961	36	2	suggests	suggest	VBZ
8961	36	3	,	,	,
8961	36	4	without	without	IN
8961	36	5	an	an	DT
8961	36	6	explanation	explanation	NN
8961	36	7	or	or	CC
8961	36	8	evaluation	evaluation	NN
8961	36	9	,	,	,
8961	36	10	the	the	DT
8961	36	11	use	use	NN
8961	36	12	of	of	IN
8961	36	13	bit	bit	NN
8961	36	14	patterns	pattern	NNS
8961	36	15	as	as	IN
8961	36	16	the	the	DT
8961	36	17	storage	storage	NN
8961	36	18	technique	technique	NN
8961	36	19	for	for	IN
8961	36	20	inverted	inverted	JJ
8961	36	21	files	file	NNS
8961	36	22	in	in	IN
8961	36	23	large	large	JJ
8961	36	24	data	data	NN
8961	36	25	bases	basis	NNS
8961	36	26	in	in	IN
8961	36	27	the	the	DT
8961	36	28	area	area	NN
8961	36	29	of	of	IN
8961	36	30	management	management	NN
8961	36	31	information	information	NN
8961	36	32	systems.1	systems.1	CD
8961	36	33	Davis	Davis	NNP
8961	36	34	and	and	CC
8961	36	35	Lin	Lin	NNP
8961	36	36	,	,	,
8961	36	37	again	again	RB
8961	36	38	in	in	IN
8961	36	39	the	the	DT
8961	36	40	area	area	NN
8961	36	41	of	of	IN
8961	36	42	management	management	NN
8961	36	43	in-	in-	NNP
8961	36	44	formation	formation	NN
8961	36	45	systems	system	NNS
8961	36	46	,	,	,
8961	36	47	propose	propose	VB
8961	36	48	bit	bit	NN
8961	36	49	patterns	pattern	NNS
8961	36	50	as	as	IN
8961	36	51	the	the	DT
8961	36	52	means	mean	NNS
8961	36	53	of	of	IN
8961	36	54	locating	locate	VBG
8961	36	55	pertinent	pertinent	JJ
8961	36	56	records	record	NNS
8961	36	57	in	in	IN
8961	36	58	a	a	DT
8961	36	59	master	master	NN
8961	36	60	file	file	NN
8961	36	61	.	.	.
8961	37	1	2	2	LS
8961	37	2	They	-PRON-	PRP
8961	37	3	do	do	VBP
8961	37	4	not	not	RB
8961	37	5	compare	compare	VB
8961	37	6	the	the	DT
8961	37	7	method	method	NN
8961	37	8	with	with	IN
8961	37	9	other	other	JJ
8961	37	10	pos-	pos-	JJ
8961	37	11	sible	sible	JJ
8961	37	12	techniques	technique	NNS
8961	37	13	.	.	.
8961	38	1	Sammon	sammon	JJ
8961	38	2	discusses	discusse	NNS
8961	38	3	briefly	briefly	RB
8961	38	4	the	the	DT
8961	38	5	use	use	NN
8961	38	6	of	of	IN
8961	38	7	binary	binary	JJ
8961	38	8	vectors	vector	NNS
8961	38	9	as	as	IN
8961	38	10	a	a	DT
8961	38	11	storage	storage	NN
8961	38	12	technique	technique	NN
8961	38	13	,	,	,
8961	38	14	but	but	CC
8961	38	15	dismisses	dismiss	VBZ
8961	38	16	it	-PRON-	PRP
8961	38	17	on	on	IN
8961	38	18	the	the	DT
8961	38	19	basis	basis	NN
8961	38	20	that	that	IN
8961	38	21	the	the	DT
8961	38	22	two	two	CD
8961	38	23	-	-	HYPH
8961	38	24	valued	value	VBN
8961	38	25	ap-	ap-	CD
8961	38	26	proach	proach	NN
8961	38	27	obviates	obviate	NNS
8961	38	28	the	the	DT
8961	38	29	possible	possible	JJ
8961	38	30	assignment	assignment	NN
8961	38	31	of	of	IN
8961	38	32	weights	weight	NNS
8961	38	33	to	to	IN
8961	38	34	index	index	NN
8961	38	35	terms	term	NNS
8961	38	36	in	in	IN
8961	38	37	de-	de-	RB
8961	38	38	scribing	scribe	VBG
8961	38	39	documents	document	NNS
8961	38	40	.	.	.
8961	39	1	3	3	CD
8961	39	2	Gorokhov	Gorokhov	NNP
8961	39	3	discusses	discuss	VBZ
8961	39	4	the	the	DT
8961	39	5	use	use	NN
8961	39	6	of	of	IN
8961	39	7	a	a	DT
8961	39	8	modified	modify	VBN
8961	39	9	binary	binary	NN
8961	39	10	vec-	vec-	NN
8961	39	11	tor	tor	NN
8961	39	12	approach	approach	NN
8961	39	13	in	in	IN
8961	39	14	a	a	DT
8961	39	15	document	document	JJ
8961	39	16	retrieval	retrieval	NN
8961	39	17	system	system	NN
8961	39	18	implemented	implement	VBN
8961	39	19	on	on	IN
8961	39	20	a	a	DT
8961	39	21	small	small	JJ
8961	39	22	Soviet	soviet	JJ
8961	39	23	computer.4	computer.4	NN
8961	39	24	Faced	face	VBN
8961	39	25	with	with	IN
8961	39	26	the	the	DT
8961	39	27	need	need	NN
8961	39	28	to	to	TO
8961	39	29	minimize	minimize	VB
8961	39	30	storage	storage	NN
8961	39	31	requirements	requirement	NNS
8961	39	32	for	for	IN
8961	39	33	his	-PRON-	PRP$
8961	39	34	inverted	invert	VBN
8961	39	35	file	file	NN
8961	39	36	,	,	,
8961	39	37	Gorokhov	Gorokhov	NNP
8961	39	38	concentrated	concentrate	VBD
8961	39	39	on	on	IN
8961	39	40	developing	develop	VBG
8961	39	41	a	a	DT
8961	39	42	technique	technique	NN
8961	39	43	for	for	IN
8961	39	44	lo-	lo-	NNP
8961	39	45	cating	cating	NNP
8961	39	46	and	and	CC
8961	39	47	removing	remove	VBG
8961	39	48	strings	string	NNS
8961	39	49	of	of	IN
8961	39	50	zeroes	zero	NNS
8961	39	51	occurring	occur	VBG
8961	39	52	in	in	IN
8961	39	53	the	the	DT
8961	39	54	binary	binary	JJ
8961	39	55	vectors	vector	NNS
8961	39	56	used	use	VBN
8961	39	57	within	within	IN
8961	39	58	the	the	DT
8961	39	59	system	system	NN
8961	39	60	.	.	.
8961	40	1	Since	since	IN
8961	40	2	these	these	DT
8961	40	3	zeroes	zero	NNS
8961	40	4	represent	represent	VBP
8961	40	5	the	the	DT
8961	40	6	absence	absence	NN
8961	40	7	of	of	IN
8961	40	8	information	information	NN
8961	40	9	they	-PRON-	PRP
8961	40	10	could	could	MD
8961	40	11	be	be	VB
8961	40	12	removed	remove	VBN
8961	40	13	if	if	IN
8961	40	14	there	there	EX
8961	40	15	were	be	VBD
8961	40	16	a	a	DT
8961	40	17	way	way	NN
8961	40	18	to	to	TO
8961	40	19	indicate	indicate	VB
8961	40	20	the	the	DT
8961	40	21	position	position	NN
8961	40	22	in	in	IN
8961	40	23	the	the	DT
8961	40	24	original	original	JJ
8961	40	25	vector	vector	NN
8961	40	26	of	of	IN
8961	40	27	the	the	DT
8961	40	28	ones	one	NNS
8961	40	29	that	that	WDT
8961	40	30	remained	remain	VBD
8961	40	31	.	.	.
8961	41	1	He	-PRON-	PRP
8961	41	2	proposed	propose	VBD
8961	41	3	the	the	DT
8961	41	4	removal	removal	NN
8961	41	5	of	of	IN
8961	41	6	strings	string	NNS
8961	41	7	of	of	IN
8961	41	8	zeroes	zero	NNS
8961	41	9	and	and	CC
8961	41	10	the	the	DT
8961	41	11	inclusion	inclusion	NN
8961	41	12	of	of	IN
8961	41	13	numeric	numeric	JJ
8961	41	14	place	place	NN
8961	41	15	values	value	NNS
8961	41	16	with	with	IN
8961	41	17	the	the	DT
8961	41	18	re-	re-	JJ
8961	41	19	maining	maining	NN
8961	41	20	vector	vector	NN
8961	41	21	elements	element	NNS
8961	41	22	.	.	.
8961	42	1	His	-PRON-	PRP$
8961	42	2	result	result	NN
8961	42	3	is	be	VBZ
8961	42	4	a	a	DT
8961	42	5	file	file	NN
8961	42	6	with	with	IN
8961	42	7	variable	variable	JJ
8961	42	8	-	-	HYPH
8961	42	9	length	length	NN
8961	42	10	index	index	NN
8961	42	11	rec-	rec-	NN
8961	42	12	ords	ord	NNS
8961	42	13	.	.	.
8961	43	1	The	the	DT
8961	43	2	abandoning	abandoning	NN
8961	43	3	of	of	IN
8961	43	4	the	the	DT
8961	43	5	pure	pure	JJ
8961	43	6	binary	binary	JJ
8961	43	7	vector	vector	NN
8961	43	8	obviates	obviate	VBZ
8961	43	9	the	the	DT
8961	43	10	process	process	NN
8961	43	11	,	,	,
8961	43	12	and	and	CC
8961	43	13	Gorokhov	Gorokhov	NNP
8961	43	14	found	find	VBD
8961	43	15	it	-PRON-	PRP
8961	43	16	necessary	necessary	JJ
8961	43	17	to	to	TO
8961	43	18	expand	expand	VB
8961	43	19	the	the	DT
8961	43	20	vector	vector	NN
8961	43	21	elements	element	NNS
8961	43	22	into	into	IN
8961	43	23	the	the	DT
8961	43	24	orig-	orig-	JJ
8961	43	25	inal	inal	JJ
8961	43	26	vector	vector	NN
8961	43	27	before	before	IN
8961	43	28	logical	logical	JJ
8961	43	29	operations	operation	NNS
8961	43	30	could	could	MD
8961	43	31	be	be	VB
8961	43	32	applied	apply	VBN
8961	43	33	.	.	.
8961	44	1	Even	even	RB
8961	44	2	though	though	IN
8961	44	3	he	-PRON-	PRP
8961	44	4	does	do	VBZ
8961	44	5	not	not	RB
8961	44	6	state	state	VB
8961	44	7	so	so	RB
8961	44	8	explicitly	explicitly	RB
8961	44	9	,	,	,
8961	44	10	Gorokhov	Gorokhov	NNP
8961	44	11	seems	seem	VBZ
8961	44	12	to	to	TO
8961	44	13	have	have	VB
8961	44	14	found	find	VBN
8961	44	15	his	-PRON-	PRP$
8961	44	16	method	method	NN
8961	44	17	more	more	RBR
8961	44	18	efficient	efficient	JJ
8961	44	19	than	than	IN
8961	44	20	the	the	DT
8961	44	21	standard	standard	JJ
8961	44	22	inverted	invert	VBN
8961	44	23	file	file	NN
8961	44	24	.	.	.
8961	45	1	Gorokhov	Gorokhov	NNP
8961	45	2	'	'	POS
8961	45	3	s	s	POS
8961	45	4	suggestion	suggestion	NN
8961	45	5	has	have	VBZ
8961	45	6	led	lead	VBN
8961	45	7	to	to	IN
8961	45	8	the	the	DT
8961	45	9	development	development	NN
8961	45	10	of	of	IN
8961	45	11	an	an	DT
8961	45	12	algorithm	algorithm	NN
8961	45	13	for	for	IN
8961	45	14	the	the	DT
8961	45	15	compression	compression	NN
8961	45	16	of	of	IN
8961	45	17	binary	binary	JJ
8961	45	18	vectors	vector	NNS
8961	45	19	.	.	.
8961	46	1	Heaps	Heaps	NNP
8961	46	2	and	and	CC
8961	46	3	Thiel	Thiel	NNP
8961	46	4	have	have	VBP
8961	46	5	also	also	RB
8961	46	6	discussed	discuss	VBN
8961	46	7	the	the	DT
8961	46	8	use	use	NN
8961	46	9	of	of	IN
8961	46	10	compressed	compress	VBN
8961	46	11	binary	binary	JJ
8961	46	12	vec-	vec-	NN
8961	46	13	tors	tor	NNS
8961	46	14	as	as	IN
8961	46	15	the	the	DT
8961	46	16	basis	basis	NN
8961	46	17	of	of	IN
8961	46	18	an	an	DT
8961	46	19	inverted	invert	VBN
8961	46	20	index	index	NN
8961	46	21	file	file	NN
8961	46	22	.	.	.
8961	47	1	5•	5•	CD
8961	47	2	6	6	CD
8961	47	3	Aside	aside	RB
8961	47	4	from	from	IN
8961	47	5	a	a	DT
8961	47	6	brief	brief	JJ
8961	47	7	descrip-	descrip-	NNS
8961	47	8	tion	tion	NN
8961	47	9	of	of	IN
8961	47	10	the	the	DT
8961	47	11	method	method	NN
8961	47	12	for	for	IN
8961	47	13	implementing	implement	VBG
8961	47	14	the	the	DT
8961	47	15	concept	concept	NN
8961	47	16	,	,	,
8961	47	17	they	-PRON-	PRP
8961	47	18	offer	offer	VBP
8961	47	19	no	no	DT
8961	47	20	compari-	compari-	JJ
8961	47	21	son	son	NN
8961	47	22	of	of	IN
8961	47	23	the	the	DT
8961	47	24	binary	binary	JJ
8961	47	25	vector	vector	NN
8961	47	26	with	with	IN
8961	47	27	the	the	DT
8961	47	28	standard	standard	JJ
8961	47	29	inverted	invert	VBN
8961	47	30	file	file	NN
8961	47	31	.	.	.
8961	48	1	STORAGE	storage	NN
8961	48	2	REQUIREMENTS	requirement	NNS
8961	48	3	An	an	DT
8961	48	4	immediate	immediate	JJ
8961	48	5	reaction	reaction	NN
8961	48	6	to	to	IN
8961	48	7	the	the	DT
8961	48	8	concept	concept	NN
8961	48	9	of	of	IN
8961	48	10	binary	binary	JJ
8961	48	11	vectors	vector	NNS
8961	48	12	is	be	VBZ
8961	48	13	to	to	TO
8961	48	14	state	state	VB
8961	48	15	that	that	IN
8961	48	16	they	-PRON-	PRP
8961	48	17	will	will	MD
8961	48	18	obviously	obviously	RB
8961	48	19	take	take	VB
8961	48	20	more	more	JJR
8961	48	21	storage	storage	NN
8961	48	22	space	space	NN
8961	48	23	than	than	IN
8961	48	24	the	the	DT
8961	48	25	standard	standard	JJ
8961	48	26	inverted	invert	VBN
8961	48	27	file	file	NN
8961	48	28	.	.	.
8961	49	1	A	a	DT
8961	49	2	closer	close	JJR
8961	49	3	study	study	NN
8961	49	4	shows	show	VBZ
8961	49	5	that	that	IN
8961	49	6	this	this	DT
8961	49	7	is	be	VBZ
8961	49	8	not	not	RB
8961	49	9	always	always	RB
8961	49	10	the	the	DT
8961	49	11	case	case	NN
8961	49	12	.	.	.
8961	50	1	The	the	DT
8961	50	2	storage	storage	NN
8961	50	3	require-	require-	VBD
8961	50	4	ments	ment	NNS
8961	50	5	for	for	IN
8961	50	6	the	the	DT
8961	50	7	two	two	CD
8961	50	8	types	type	NNS
8961	50	9	of	of	IN
8961	50	10	files	file	NNS
8961	50	11	may	may	MD
8961	50	12	be	be	VB
8961	50	13	calculated	calculate	VBN
8961	50	14	as	as	IN
8961	50	15	follows	follow	VBZ
8961	50	16	:	:	:
8961	50	17	310	310	CD
8961	50	18	Journal	Journal	NNP
8961	50	19	of	of	IN
8961	50	20	Library	Library	NNP
8961	50	21	Automation	Automation	NNP
8961	50	22	Vol	Vol	NNP
8961	50	23	.	.	.
8961	51	1	7/4	7/4	CD
8961	51	2	December	December	NNP
8961	51	3	1974	1974	CD
8961	51	4	D·N	D·N	NNP
8961	51	5	1	1	CD
8961	51	6	.	.	.
8961	52	1	MBv	MBv	NNS
8961	52	2	=	=	SYM
8961	52	3	8	8	CD
8961	52	4	bytes	bytes	NN
8961	52	5	2	2	CD
8961	52	6	.	.	.
8961	53	1	Msr	msr	JJ
8961	53	2	=	=	SYM
8961	53	3	D	d	NN
8961	53	4	·	·	NFP
8961	53	5	I	-PRON-	PRP
8961	53	6	·	·	NFP
8961	53	7	K	K	NNP
8961	53	8	where	where	WRB
8961	53	9	:	:	:
8961	53	10	(	(	-LRB-
8961	53	11	binary	binary	JJ
8961	53	12	vector	vector	NN
8961	53	13	file	file	NN
8961	53	14	)	)	-RRB-
8961	53	15	(	(	-LRB-
8961	53	16	standard	standard	JJ
8961	53	17	inverted	invert	VBN
8961	53	18	file	file	NN
8961	53	19	)	)	-RRB-
8961	53	20	M	m	NN
8961	53	21	=	=	SYM
8961	53	22	Storage	storage	NN
8961	53	23	requirements	requirement	NNS
8961	53	24	in	in	IN
8961	53	25	bytes	bytes	NN
8961	53	26	D	d	NN
8961	53	27	=	=	SYM
8961	53	28	Number	number	NN
8961	53	29	of	of	IN
8961	53	30	documents	document	NNS
8961	53	31	in	in	IN
8961	53	32	the	the	DT
8961	53	33	system	system	NN
8961	53	34	N	n	NN
8961	53	35	=	=	SYM
8961	53	36	Number	number	NN
8961	53	37	of	of	IN
8961	53	38	index	index	NN
8961	53	39	terms	term	NNS
8961	53	40	in	in	IN
8961	53	41	the	the	DT
8961	53	42	system	system	NN
8961	53	43	I	-PRON-	PRP
8961	53	44	=	=	NFP
8961	53	45	Average	average	JJ
8961	53	46	depth	depth	NN
8961	53	47	of	of	IN
8961	53	48	indexing	indexing	NN
8961	53	49	in	in	IN
8961	53	50	the	the	DT
8961	53	51	system	system	NN
8961	53	52	K	k	NN
8961	53	53	=	=	SYM
8961	53	54	Size	size	NN
8961	53	55	in	in	IN
8961	53	56	bytes	byte	NNS
8961	53	57	of	of	IN
8961	53	58	a	a	DT
8961	53	59	document	document	NN
8961	53	60	number	number	NN
8961	53	61	stored	store	VBN
8961	53	62	in	in	IN
8961	53	63	the	the	DT
8961	53	64	file	file	NN
8961	53	65	Using	use	VBG
8961	53	66	equations	equation	NNS
8961	53	67	1	1	CD
8961	53	68	and	and	CC
8961	53	69	2	2	CD
8961	53	70	we	-PRON-	PRP
8961	53	71	find	find	VBP
8961	53	72	that	that	IN
8961	53	73	the	the	DT
8961	53	74	storage	storage	NN
8961	53	75	requirements	requirement	NNS
8961	53	76	for	for	IN
8961	53	77	the	the	DT
8961	53	78	binary	binary	JJ
8961	53	79	vector	vector	NN
8961	53	80	file	file	NN
8961	53	81	are	be	VBP
8961	53	82	,	,	,
8961	53	83	in	in	IN
8961	53	84	fact	fact	NN
8961	53	85	,	,	,
8961	53	86	less	less	JJR
8961	53	87	than	than	IN
8961	53	88	the	the	DT
8961	53	89	requirements	requirement	NNS
8961	53	90	for	for	IN
8961	53	91	the	the	DT
8961	53	92	standard	standard	JJ
8961	53	93	inverted	invert	VBN
8961	53	94	file	file	NN
8961	53	95	if	if	IN
8961	53	96	N	N	NNP
8961	53	97	<	<	XX
8961	53	98	8	8	CD
8961	53	99	•	•	NNP
8961	53	100	]	]	-RRB-
8961	53	101	•	•	NNP
8961	53	102	K.	K.	NNP
8961	53	103	It	-PRON-	PRP
8961	53	104	is	be	VBZ
8961	53	105	well	well	RB
8961	53	106	lmown	lmown	VBN
8961	53	107	that	that	IN
8961	53	108	the	the	DT
8961	53	109	distribution	distribution	NN
8961	53	110	of	of	IN
8961	53	111	the	the	DT
8961	53	112	use	use	NN
8961	53	113	of	of	IN
8961	53	114	index	index	NN
8961	53	115	terms	term	NNS
8961	53	116	follows	follow	VBZ
8961	53	117	a	a	DT
8961	53	118	logarithmic	logarithmic	JJ
8961	53	119	curve	curve	NN
8961	53	120	.	.	.
8961	54	1	In	in	IN
8961	54	2	simple	simple	JJ
8961	54	3	terms	term	NNS
8961	54	4	,	,	,
8961	54	5	one	one	PRP
8961	54	6	might	may	MD
8961	54	7	say	say	VB
8961	54	8	that	that	IN
8961	54	9	a	a	DT
8961	54	10	few	few	JJ
8961	54	11	terms	term	NNS
8961	54	12	are	be	VBP
8961	54	13	used	use	VBN
8961	54	14	very	very	RB
8961	54	15	frequently	frequently	RB
8961	54	16	and	and	CC
8961	54	17	many	many	JJ
8961	54	18	terms	term	NNS
8961	54	19	are	be	VBP
8961	54	20	used	use	VBN
8961	54	21	infrequently	infrequently	RB
8961	54	22	.	.	.
8961	55	1	This	this	DT
8961	55	2	condi-	condi-	NN
8961	55	3	tion	tion	NN
8961	55	4	implies	imply	VBZ
8961	55	5	that	that	IN
8961	55	6	in	in	IN
8961	55	7	a	a	DT
8961	55	8	binary	binary	JJ
8961	55	9	vector	vector	NN
8961	55	10	file	file	NN
8961	55	11	the	the	DT
8961	55	12	records	record	NNS
8961	55	13	for	for	IN
8961	55	14	many	many	JJ
8961	55	15	terms	term	NNS
8961	55	16	will	will	MD
8961	55	17	contain	contain	VB
8961	55	18	segments	segment	NNS
8961	55	19	in	in	IN
8961	55	20	which	which	WDT
8961	55	21	there	there	EX
8961	55	22	are	be	VBP
8961	55	23	no	no	DT
8961	55	24	"	"	``
8961	55	25	ones	one	NNS
8961	55	26	"	"	''
8961	55	27	in	in	IN
8961	55	28	any	any	DT
8961	55	29	byte	byte	NN
8961	55	30	.	.	.
8961	56	1	A	a	DT
8961	56	2	method	method	NN
8961	56	3	for	for	IN
8961	56	4	re-	re-	RB
8961	56	5	moving	move	VBG
8961	56	6	these	these	DT
8961	56	7	"	"	``
8961	56	8	zero	zero	CD
8961	56	9	"	"	''
8961	56	10	bytes	byte	NNS
8961	56	11	is	be	VBZ
8961	56	12	called	call	VBN
8961	56	13	compression	compression	NN
8961	56	14	.	.	.
8961	57	1	COMPRESSION	COMPRESSION	VBN
8961	57	2	ALGORITHM	ALGORITHM	NNP
8961	57	3	The	the	DT
8961	57	4	technique	technique	NN
8961	57	5	for	for	IN
8961	57	6	the	the	DT
8961	57	7	compression	compression	NN
8961	57	8	of	of	IN
8961	57	9	binary	binary	JJ
8961	57	10	vectors	vector	NNS
8961	57	11	as	as	IN
8961	57	12	described	describe	VBN
8961	57	13	here	here	RB
8961	57	14	is	be	VBZ
8961	57	15	designed	design	VBN
8961	57	16	specifically	specifically	RB
8961	57	17	for	for	IN
8961	57	18	the	the	DT
8961	57	19	IBM	IBM	NNP
8961	57	20	360	360	CD
8961	57	21	family	family	NN
8961	57	22	of	of	IN
8961	57	23	computers	computer	NNS
8961	57	24	and	and	CC
8961	57	25	similar	similar	JJ
8961	57	26	machines	machine	NNS
8961	57	27	.	.	.
8961	58	1	The	the	DT
8961	58	2	extension	extension	NN
8961	58	3	to	to	IN
8961	58	4	other	other	JJ
8961	58	5	machines	machine	NNS
8961	58	6	should	should	MD
8961	58	7	be	be	VB
8961	58	8	obvious	obvious	JJ
8961	58	9	.	.	.
8961	59	1	Within	within	IN
8961	59	2	the	the	DT
8961	59	3	IBM	IBM	NNP
8961	59	4	360	360	CD
8961	59	5	the	the	DT
8961	59	6	byte	byte	NN
8961	59	7	,	,	,
8961	59	8	which	which	WDT
8961	59	9	contains	contain	VBZ
8961	59	10	eight	eight	CD
8961	59	11	binary	binary	NN
8961	59	12	digits	digit	NNS
8961	59	13	,	,	,
8961	59	14	is	be	VBZ
8961	59	15	the	the	DT
8961	59	16	basic	basic	JJ
8961	59	17	storage	storage	NN
8961	59	18	unit	unit	NN
8961	59	19	,	,	,
8961	59	20	and	and	CC
8961	59	21	with	with	IN
8961	59	22	the	the	DT
8961	59	23	eight	eight	CD
8961	59	24	binary	binary	NN
8961	59	25	digits	digit	NNS
8961	59	26	it	-PRON-	PRP
8961	59	27	is	be	VBZ
8961	59	28	possible	possible	JJ
8961	59	29	to	to	TO
8961	59	30	store	store	VB
8961	59	31	a	a	DT
8961	59	32	maximum	maximum	JJ
8961	59	33	integer	integer	NN
8961	59	34	value	value	NN
8961	59	35	of	of	IN
8961	59	36	255	255	CD
8961	59	37	.	.	.
8961	60	1	For	for	IN
8961	60	2	the	the	DT
8961	60	3	purpose	purpose	NN
8961	60	4	of	of	IN
8961	60	5	describing	describe	VBG
8961	60	6	a	a	DT
8961	60	7	proposed	propose	VBN
8961	60	8	compression	compression	NN
8961	60	9	algorithm	algorithm	NN
8961	60	10	for	for	IN
8961	60	11	the	the	DT
8961	60	12	binary	binary	JJ
8961	60	13	vector	vector	NN
8961	60	14	in	in	IN
8961	60	15	the	the	DT
8961	60	16	IBM	IBM	NNP
8961	60	17	360	360	CD
8961	60	18	,	,	,
8961	60	19	the	the	DT
8961	60	20	term	term	NN
8961	60	21	sub-	sub-	JJ
8961	60	22	vector	vector	NN
8961	60	23	will	will	MD
8961	60	24	be	be	VB
8961	60	25	defined	define	VBN
8961	60	26	as	as	IN
8961	60	27	a	a	DT
8961	60	28	string	string	NN
8961	60	29	of	of	IN
8961	60	30	contiguous	contiguous	JJ
8961	60	31	bytes	byte	NNS
8961	60	32	chosen	choose	VBN
8961	60	33	from	from	IN
8961	60	34	within	within	IN
8961	60	35	the	the	DT
8961	60	36	binary	binary	JJ
8961	60	37	vector	vector	NN
8961	60	38	.	.	.
8961	61	1	A	a	DT
8961	61	2	zero	zero	CD
8961	61	3	subvector	subvector	NN
8961	61	4	will	will	MD
8961	61	5	be	be	VB
8961	61	6	a	a	DT
8961	61	7	subvector	subvector	NN
8961	61	8	each	each	DT
8961	61	9	of	of	IN
8961	61	10	whose	whose	WP$
8961	61	11	bytes	byte	NNS
8961	61	12	contains	contain	VBZ
8961	61	13	eight	eight	CD
8961	61	14	binary	binary	JJ
8961	61	15	zeroes	zero	NNS
8961	61	16	.	.	.
8961	62	1	A	a	DT
8961	62	2	nonzero	nonzero	NN
8961	62	3	subvecto1	subvecto1	NN
8961	62	4	·	·	NFP
8961	62	5	will	will	MD
8961	62	6	be	be	VB
8961	62	7	a	a	DT
8961	62	8	subvector	subvector	NN
8961	62	9	each	each	DT
8961	62	10	of	of	IN
8961	62	11	whose	whose	WP$
8961	62	12	bytes	byte	NNS
8961	62	13	contains	contain	VBZ
8961	62	14	at	at	IN
8961	62	15	least	least	JJS
8961	62	16	one	one	CD
8961	62	17	binary	binary	JJ
8961	62	18	one	one	NN
8961	62	19	.	.	.
8961	63	1	To	to	TO
8961	63	2	compress	compress	VB
8961	63	3	a	a	DT
8961	63	4	binary	binary	JJ
8961	63	5	vec-	vec-	NN
8961	63	6	tor	tor	NN
8961	63	7	in	in	IN
8961	63	8	the	the	DT
8961	63	9	IBM	IBM	NNP
8961	63	10	360	360	CD
8961	63	11	the	the	DT
8961	63	12	following	follow	VBG
8961	63	13	steps	step	NNS
8961	63	14	may	may	MD
8961	63	15	be	be	VB
8961	63	16	taken	take	VBN
8961	63	17	:	:	:
8961	63	18	1	1	CD
8961	63	19	.	.	.
8961	64	1	Divide	divide	VB
8961	64	2	the	the	DT
8961	64	3	binary	binary	JJ
8961	64	4	vector	vector	NN
8961	64	5	into	into	IN
8961	64	6	a	a	DT
8961	64	7	series	series	NN
8961	64	8	of	of	IN
8961	64	9	zero	zero	CD
8961	64	10	subvectors	subvector	NNS
8961	64	11	and	and	CC
8961	64	12	nonzero	nonzero	NNP
8961	64	13	subvectors	subvector	NNS
8961	64	14	.	.	.
8961	65	1	Subvectors	subvector	NNS
8961	65	2	of	of	IN
8961	65	3	either	either	DT
8961	65	4	type	type	NN
8961	65	5	may	may	MD
8961	65	6	have	have	VB
8961	65	7	a	a	DT
8961	65	8	maximum	maximum	JJ
8961	65	9	length	length	NN
8961	65	10	of	of	IN
8961	65	11	255	255	CD
8961	65	12	bytes	byte	NNS
8961	65	13	.	.	.
8961	66	1	For	for	IN
8961	66	2	zero	zero	CD
8961	66	3	subvectors	subvector	NNS
8961	66	4	longer	long	JJR
8961	66	5	than	than	IN
8961	66	6	255	255	CD
8961	66	7	bytes	byte	NNS
8961	66	8	,	,	,
8961	66	9	the	the	DT
8961	66	10	256th	256th	JJ
8961	66	11	byte	byte	NN
8961	66	12	is	be	VBZ
8961	66	13	to	to	TO
8961	66	14	be	be	VB
8961	66	15	treated	treat	VBN
8961	66	16	as	as	IN
8961	66	17	a	a	DT
8961	66	18	nonzero	nonzero	NN
8961	66	19	byte	byte	NN
8961	66	20	,	,	,
8961	66	21	thus	thus	RB
8961	66	22	dividing	divide	VBG
8961	66	23	the	the	DT
8961	66	24	long	long	JJ
8961	66	25	zero	zero	CD
8961	66	26	subvector	subvector	NN
8961	66	27	.	.	.
8961	67	1	2	2	LS
8961	67	2	.	.	.
8961	68	1	Each	each	DT
8961	68	2	nonzero	nonzero	NN
8961	68	3	subvector	subvector	NN
8961	68	4	is	be	VBZ
8961	68	5	prefixed	prefix	VBN
8961	68	6	with	with	IN
8961	68	7	two	two	CD
8961	68	8	bytes	byte	NNS
8961	68	9	.	.	.
8961	69	1	The	the	DT
8961	69	2	first	first	JJ
8961	69	3	of	of	IN
8961	69	4	the	the	DT
8961	69	5	prefix	prefix	JJ
8961	69	6	bytes	byte	NNS
8961	69	7	contains	contain	VBZ
8961	69	8	the	the	DT
8961	69	9	count	count	NN
8961	69	10	of	of	IN
8961	69	11	zero	zero	CD
8961	69	12	bytes	byte	NNS
8961	69	13	which	which	WDT
8961	69	14	precede	precede	VBP
8961	69	15	the	the	DT
8961	69	16	non-	non-	NNP
8961	69	17	zero	zero	CD
8961	69	18	subvector	subvector	NN
8961	69	19	in	in	IN
8961	69	20	the	the	DT
8961	69	21	uncompressed	uncompressed	JJ
8961	69	22	vector	vector	NN
8961	69	23	.	.	.
8961	70	1	The	the	DT
8961	70	2	second	second	JJ
8961	70	3	prefix	prefix	NN
8961	70	4	byte	byte	NN
8961	70	5	contains	contain	VBZ
8961	70	6	a	a	DT
8961	70	7	count	count	NN
8961	70	8	of	of	IN
8961	70	9	the	the	DT
8961	70	10	bytes	byte	NNS
8961	70	11	in	in	IN
8961	70	12	the	the	DT
8961	70	13	nonzero	nonzero	NN
8961	70	14	subvector	subvector	NN
8961	70	15	.	.	.
8961	71	1	3	3	LS
8961	71	2	.	.	.
8961	72	1	The	the	DT
8961	72	2	compressed	compressed	JJ
8961	72	3	vector	vector	NN
8961	72	4	then	then	RB
8961	72	5	consists	consist	VBZ
8961	72	6	of	of	IN
8961	72	7	only	only	RB
8961	72	8	the	the	DT
8961	72	9	nonzero	nonzero	NN
8961	72	10	subvectors	subvector	NNS
8961	72	11	together	together	RB
8961	72	12	with	with	IN
8961	72	13	their	-PRON-	PRP$
8961	72	14	prefix	prefix	JJ
8961	72	15	bytes	byte	NNS
8961	72	16	.	.	.
8961	73	1	4	4	LS
8961	73	2	.	.	.
8961	74	1	A	a	DT
8961	74	2	two	two	CD
8961	74	3	byte	byte	NN
8961	74	4	field	field	NN
8961	74	5	of	of	IN
8961	74	6	binary	binary	JJ
8961	74	7	zeroes	zero	NNS
8961	74	8	will	will	MD
8961	74	9	end	end	VB
8961	74	10	the	the	DT
8961	74	11	compressed	compressed	JJ
8961	74	12	vector	vector	NN
8961	74	13	.	.	.
8961	75	1	Binmy	Binmy	NNP
8961	75	2	Vector	Vector	NNP
8961	75	3	/	/	SYM
8961	75	4	KING	KING	NNP
8961	75	5	311	311	CD
8961	75	6	The	the	DT
8961	75	7	compression	compression	NN
8961	75	8	of	of	IN
8961	75	9	the	the	DT
8961	75	10	vectors	vector	NNS
8961	75	11	creates	create	VBZ
8961	75	12	variable	variable	JJ
8961	75	13	-	-	HYPH
8961	75	14	length	length	NN
8961	75	15	records	record	NNS
8961	75	16	and	and	CC
8961	75	17	re-	re-	RB
8961	75	18	moves	move	VBZ
8961	75	19	the	the	DT
8961	75	20	advantage	advantage	NN
8961	75	21	of	of	IN
8961	75	22	having	have	VBG
8961	75	23	records	record	NNS
8961	75	24	which	which	WDT
8961	75	25	are	be	VBP
8961	75	26	directly	directly	RB
8961	75	27	amenable	amenable	JJ
8961	75	28	to	to	IN
8961	75	29	Boolean	boolean	JJ
8961	75	30	manipulation	manipulation	NN
8961	75	31	.	.	.
8961	76	1	The	the	DT
8961	76	2	effect	effect	NN
8961	76	3	of	of	IN
8961	76	4	file	file	NN
8961	76	5	compression	compression	NN
8961	76	6	on	on	IN
8961	76	7	such	such	JJ
8961	76	8	manipula-	manipula-	JJ
8961	76	9	tion	tion	NN
8961	76	10	in	in	IN
8961	76	11	the	the	DT
8961	76	12	search	search	NN
8961	76	13	process	process	NN
8961	76	14	is	be	VBZ
8961	76	15	not	not	RB
8961	76	16	as	as	RB
8961	76	17	severe	severe	JJ
8961	76	18	as	as	IN
8961	76	19	it	-PRON-	PRP
8961	76	20	might	may	MD
8961	76	21	appear	appear	VB
8961	76	22	.	.	.
8961	77	1	For	for	IN
8961	77	2	the	the	DT
8961	77	3	search	search	NN
8961	77	4	process	process	NN
8961	77	5	,	,	,
8961	77	6	the	the	DT
8961	77	7	compressed	compressed	JJ
8961	77	8	vector	vector	NN
8961	77	9	may	may	MD
8961	77	10	be	be	VB
8961	77	11	expanded	expand	VBN
8961	77	12	into	into	IN
8961	77	13	its	-PRON-	PRP$
8961	77	14	original	original	JJ
8961	77	15	form	form	NN
8961	77	16	.	.	.
8961	78	1	The	the	DT
8961	78	2	process	process	NN
8961	78	3	of	of	IN
8961	78	4	expansion	expansion	NN
8961	78	5	of	of	IN
8961	78	6	the	the	DT
8961	78	7	binary	binary	JJ
8961	78	8	vectors	vector	NNS
8961	78	9	is	be	VBZ
8961	78	10	relatively	relatively	RB
8961	78	11	simple	simple	JJ
8961	78	12	,	,	,
8961	78	13	and	and	CC
8961	78	14	since	since	IN
8961	78	15	only	only	RB
8961	78	16	those	those	DT
8961	78	17	index	index	NN
8961	78	18	term	term	NN
8961	78	19	records	record	NNS
8961	78	20	which	which	WDT
8961	78	21	are	be	VBP
8961	78	22	used	use	VBN
8961	78	23	in	in	IN
8961	78	24	a	a	DT
8961	78	25	query	query	NN
8961	78	26	need	need	NN
8961	78	27	to	to	TO
8961	78	28	be	be	VB
8961	78	29	expanded	expand	VBN
8961	78	30	at	at	IN
8961	78	31	the	the	DT
8961	78	32	search	search	NN
8961	78	33	time	time	NN
8961	78	34	,	,	,
8961	78	35	the	the	DT
8961	78	36	search	search	NN
8961	78	37	time	time	NN
8961	78	38	is	be	VBZ
8961	78	39	not	not	RB
8961	78	40	significantly	significantly	RB
8961	78	41	affected	affect	VBN
8961	78	42	.	.	.
8961	79	1	As	as	IN
8961	79	2	an	an	DT
8961	79	3	example	example	NN
8961	79	4	of	of	IN
8961	79	5	the	the	DT
8961	79	6	use	use	NN
8961	79	7	of	of	IN
8961	79	8	the	the	DT
8961	79	9	compression	compression	NN
8961	79	10	algorithm	algorithm	NNP
8961	79	11	consider	consider	VBP
8961	79	12	the	the	DT
8961	79	13	fol-	fol-	NN
8961	79	14	lowing	low	VBG
8961	79	15	binary	binary	JJ
8961	79	16	vector	vector	NN
8961	79	17	.	.	.
8961	80	1	01100000/10000000/	01100000/10000000/	NFP
8961	80	2	seven	seven	CD
8961	80	3	zero	zero	CD
8961	80	4	bytes	bytes	NN
8961	80	5	j00000001j10000000j	j00000001j10000000j	CD
8961	80	6	...	...	.
8961	81	1	The	the	DT
8961	81	2	slashes	slash	NNS
8961	81	3	indicate	indicate	VBP
8961	81	4	the	the	DT
8961	81	5	division	division	NN
8961	81	6	of	of	IN
8961	81	7	the	the	DT
8961	81	8	vector	vector	NN
8961	81	9	into	into	IN
8961	81	10	bytes	byte	NNS
8961	81	11	.	.	.
8961	82	1	The	the	DT
8961	82	2	vector	vector	NN
8961	82	3	might	may	MD
8961	82	4	be	be	VB
8961	82	5	read	read	VBN
8961	82	6	as	as	IN
8961	82	7	indicating	indicate	VBG
8961	82	8	the	the	DT
8961	82	9	following	follow	VBG
8961	82	10	list	list	NN
8961	82	11	of	of	IN
8961	82	12	document	document	NN
8961	82	13	numbers	number	NNS
8961	82	14	:	:	:
8961	82	15	2	2	CD
8961	82	16	,	,	,
8961	82	17	3	3	CD
8961	82	18	,	,	,
8961	82	19	9	9	CD
8961	82	20	,	,	,
8961	82	21	80	80	CD
8961	82	22	,	,	,
8961	82	23	and	and	CC
8961	82	24	81	81	CD
8961	82	25	.	.	.
8961	83	1	In	in	IN
8961	83	2	a	a	DT
8961	83	3	standard	standard	JJ
8961	83	4	inverted	invert	VBN
8961	83	5	file	file	NN
8961	83	6	with	with	IN
8961	83	7	each	each	DT
8961	83	8	document	document	NN
8961	83	9	number	number	NN
8961	83	10	assigned	assign	VBN
8961	83	11	three	three	CD
8961	83	12	bytes	byte	NNS
8961	83	13	of	of	IN
8961	83	14	storage	storage	NN
8961	83	15	,	,	,
8961	83	16	fifteen	fifteen	CD
8961	83	17	bytes	byte	NNS
8961	83	18	would	would	MD
8961	83	19	be	be	VB
8961	83	20	required	require	VBN
8961	83	21	to	to	TO
8961	83	22	store	store	VB
8961	83	23	these	these	DT
8961	83	24	numbers	number	NNS
8961	83	25	.	.	.
8961	84	1	The	the	DT
8961	84	2	compressed	compressed	JJ
8961	84	3	vector	vector	NN
8961	84	4	which	which	WDT
8961	84	5	results	result	VBZ
8961	84	6	from	from	IN
8961	84	7	the	the	DT
8961	84	8	application	application	NN
8961	84	9	of	of	IN
8961	84	10	the	the	DT
8961	84	11	algo-	algo-	JJ
8961	84	12	rithm	rithm	NN
8961	84	13	is	be	VBZ
8961	84	14	the	the	DT
8961	84	15	following	follow	VBG
8961	84	16	:	:	:
8961	84	17	00000000j00000010j01100000/10000000j00000111/00000010/	00000000j00000010j01100000/10000000j00000111/00000010/	CD
8961	84	18	00000001/10000000/	00000001/10000000/	CD
8961	84	19	...	...	:
8961	84	20	Again	again	RB
8961	84	21	the	the	DT
8961	84	22	slashes	slash	NNS
8961	84	23	separate	separate	VBP
8961	84	24	the	the	DT
8961	84	25	vector	vector	NN
8961	84	26	into	into	IN
8961	84	27	bytes	byte	NNS
8961	84	28	.	.	.
8961	85	1	For	for	IN
8961	85	2	the	the	DT
8961	85	3	purpose	purpose	NN
8961	85	4	of	of	IN
8961	85	5	the	the	DT
8961	85	6	fol-	fol-	NN
8961	85	7	lowing	low	VBG
8961	85	8	discussion	discussion	NN
8961	85	9	consider	consider	VB
8961	85	10	each	each	DT
8961	85	11	byte	byte	NN
8961	85	12	in	in	IN
8961	85	13	a	a	DT
8961	85	14	vector	vector	NN
8961	85	15	to	to	TO
8961	85	16	be	be	VB
8961	85	17	numbered	number	VBN
8961	85	18	sequential-	sequential-	NN
8961	85	19	ly	ly	XX
8961	85	20	beginning	begin	VBG
8961	85	21	with	with	IN
8961	85	22	byte	byte	NN
8961	85	23	one	one	CD
8961	85	24	at	at	IN
8961	85	25	the	the	DT
8961	85	26	left	left	NN
8961	85	27	.	.	.
8961	86	1	In	in	IN
8961	86	2	the	the	DT
8961	86	3	uncompressed	uncompressed	JJ
8961	86	4	vector	vector	NN
8961	86	5	bytes	byte	VBZ
8961	86	6	one	one	CD
8961	86	7	and	and	CC
8961	86	8	two	two	CD
8961	86	9	form	form	NN
8961	86	10	a	a	DT
8961	86	11	nonzero	nonzero	NN
8961	86	12	subvector	subvector	NN
8961	86	13	.	.	.
8961	87	1	Consequently	consequently	RB
8961	87	2	,	,	,
8961	87	3	the	the	DT
8961	87	4	first	first	JJ
8961	87	5	four	four	CD
8961	87	6	bytes	byte	NNS
8961	87	7	in	in	IN
8961	87	8	the	the	DT
8961	87	9	compressed	compress	VBN
8961	87	10	vector	vector	NN
8961	87	11	can	can	MD
8961	87	12	be	be	VB
8961	87	13	inter-	inter-	NN
8961	87	14	preted	prete	VBN
8961	87	15	as	as	IN
8961	87	16	follows	follow	VBZ
8961	87	17	:	:	:
8961	87	18	Byte	byte	VB
8961	87	19	one	one	CD
8961	87	20	.	.	.
8961	88	1	Binary	binary	JJ
8961	88	2	zero	zero	CD
8961	88	3	indicating	indicate	VBG
8961	88	4	that	that	IN
8961	88	5	no	no	DT
8961	88	6	zero	zero	CD
8961	88	7	bytes	byte	NNS
8961	88	8	were	be	VBD
8961	88	9	re-	re-	RB
8961	88	10	moved	move	VBN
8961	88	11	preceding	precede	VBG
8961	88	12	this	this	DT
8961	88	13	subvector	subvector	NN
8961	88	14	.	.	.
8961	89	1	Byte	byte	NN
8961	89	2	two	two	CD
8961	89	3	.	.	.
8961	90	1	Binary	binary	JJ
8961	90	2	two	two	CD
8961	90	3	indicating	indicate	VBG
8961	90	4	that	that	IN
8961	90	5	the	the	DT
8961	90	6	following	follow	VBG
8961	90	7	nonzero	nonzero	CD
8961	90	8	sub-	sub-	JJ
8961	90	9	vector	vector	NN
8961	90	10	is	be	VBZ
8961	90	11	two	two	CD
8961	90	12	bytes	byte	NNS
8961	90	13	long	long	JJ
8961	90	14	.	.	.
8961	91	1	Bytes	byte	NNS
8961	91	2	three	three	CD
8961	91	3	,	,	,
8961	91	4	four	four	CD
8961	91	5	.	.	.
8961	92	1	Bytes	byte	NNS
8961	92	2	one	one	CD
8961	92	3	and	and	CC
8961	92	4	two	two	CD
8961	92	5	of	of	IN
8961	92	6	the	the	DT
8961	92	7	original	original	JJ
8961	92	8	vector	vector	NN
8961	92	9	.	.	.
8961	93	1	Bytes	byte	NNS
8961	93	2	three	three	CD
8961	93	3	through	through	IN
8961	93	4	nine	nine	CD
8961	93	5	of	of	IN
8961	93	6	the	the	DT
8961	93	7	original	original	JJ
8961	93	8	vector	vector	NN
8961	93	9	are	be	VBP
8961	93	10	a	a	DT
8961	93	11	zero	zero	CD
8961	93	12	subvector	subvector	NN
8961	93	13	,	,	,
8961	93	14	and	and	CC
8961	93	15	bytes	byte	VBZ
8961	93	16	ten	ten	CD
8961	93	17	and	and	CC
8961	93	18	eleven	eleven	CD
8961	93	19	form	form	NN
8961	93	20	a	a	DT
8961	93	21	second	second	JJ
8961	93	22	nonzero	nonzero	NN
8961	93	23	subvector	subvector	NN
8961	93	24	.	.	.
8961	94	1	Consequently	consequently	RB
8961	94	2	,	,	,
8961	94	3	the	the	DT
8961	94	4	second	second	JJ
8961	94	5	four	four	CD
8961	94	6	bytes	byte	NNS
8961	94	7	of	of	IN
8961	94	8	the	the	DT
8961	94	9	compressed	compressed	JJ
8961	94	10	vector	vector	NN
8961	94	11	are	be	VBP
8961	94	12	interpreted	interpret	VBN
8961	94	13	as	as	IN
8961	94	14	follows	follow	VBZ
8961	94	15	:	:	:
8961	94	16	Byte	byte	NN
8961	94	17	five	five	CD
8961	94	18	.	.	.
8961	95	1	Binary	binary	JJ
8961	95	2	seven	seven	CD
8961	95	3	indicating	indicate	VBG
8961	95	4	that	that	IN
8961	95	5	a	a	DT
8961	95	6	zero	zero	CD
8961	95	7	subvector	subvector	NN
8961	95	8	of	of	IN
8961	95	9	seven	seven	CD
8961	95	10	bytes	byte	NNS
8961	95	11	has	have	VBZ
8961	95	12	been	be	VBN
8961	95	13	removed	remove	VBN
8961	95	14	.	.	.
8961	96	1	Byte	byte	NN
8961	96	2	six	six	CD
8961	96	3	.	.	.
8961	97	1	Binary	binary	JJ
8961	97	2	two	two	CD
8961	97	3	indicating	indicate	VBG
8961	97	4	that	that	IN
8961	97	5	the	the	DT
8961	97	6	following	follow	VBG
8961	97	7	two	two	CD
8961	97	8	bytes	byte	NNS
8961	97	9	are	be	VBP
8961	97	10	a	a	DT
8961	97	11	nonzero	nonzero	NN
8961	97	12	subvector	subvector	NN
8961	97	13	.	.	.
8961	98	1	Bytes	byte	NNS
8961	98	2	seven	seven	CD
8961	98	3	,	,	,
8961	98	4	eight	eight	CD
8961	98	5	.	.	.
8961	99	1	Bytes	byte	NNS
8961	99	2	ten	ten	CD
8961	99	3	and	and	CC
8961	99	4	eleven	eleven	CD
8961	99	5	of	of	IN
8961	99	6	the	the	DT
8961	99	7	original	original	JJ
8961	99	8	vector	vector	NN
8961	99	9	.	.	.
8961	100	1	Thus	thus	RB
8961	100	2	the	the	DT
8961	100	3	binary	binary	JJ
8961	100	4	vector	vector	NN
8961	100	5	has	have	VBZ
8961	100	6	been	be	VBN
8961	100	7	reduced	reduce	VBN
8961	100	8	from	from	IN
8961	100	9	eleven	eleven	CD
8961	100	10	bytes	byte	NNS
8961	100	11	to	to	IN
8961	100	12	eight	eight	CD
8961	100	13	312	312	CD
8961	100	14	Journal	Journal	NNP
8961	100	15	of	of	IN
8961	100	16	Library	Library	NNP
8961	100	17	Automation	Automation	NNP
8961	100	18	Vol	Vol	NNP
8961	100	19	.	.	.
8961	101	1	7/4	7/4	CD
8961	101	2	December	December	NNP
8961	101	3	1974	1974	CD
8961	101	4	bytes	byte	NNS
8961	101	5	while	while	IN
8961	101	6	the	the	DT
8961	101	7	space	space	NN
8961	101	8	required	require	VBN
8961	101	9	to	to	TO
8961	101	10	record	record	VB
8961	101	11	the	the	DT
8961	101	12	document	document	NN
8961	101	13	numbers	number	NNS
8961	101	14	in	in	IN
8961	101	15	the	the	DT
8961	101	16	stan-	stan-	NNP
8961	101	17	dard	dard	NNP
8961	101	18	inverted	invert	VBN
8961	101	19	file	file	NN
8961	101	20	remains	remain	VBZ
8961	101	21	fifteen	fifteen	CD
8961	101	22	bytes	byte	NNS
8961	101	23	.	.	.
8961	102	1	MEMORY	memory	NN
8961	102	2	REQUIREMENTS	requirement	NNS
8961	102	3	FOR	for	IN
8961	102	4	THE	the	DT
8961	102	5	STANDARD	STANDARD	NNP
8961	102	6	INVERTED	INVERTED	NNP
8961	102	7	FILE	FILE	NNP
8961	102	8	AND	and	CC
8961	102	9	THE	the	DT
8961	102	10	BINARY	BINARY	NNP
8961	102	11	VECTOR	vector	NN
8961	102	12	FILE	FILE	VBD
8961	102	13	To	to	TO
8961	102	14	compare	compare	VB
8961	102	15	memory	memory	NN
8961	102	16	requirements	requirement	NNS
8961	102	17	for	for	IN
8961	102	18	the	the	DT
8961	102	19	standard	standard	JJ
8961	102	20	inverted	invert	VBN
8961	102	21	file	file	NN
8961	102	22	and	and	CC
8961	102	23	the	the	DT
8961	102	24	compressed	compressed	JJ
8961	102	25	binary	binary	JJ
8961	102	26	vector	vector	NN
8961	102	27	file	file	NN
8961	102	28	,	,	,
8961	102	29	we	-PRON-	PRP
8961	102	30	base	base	VBP
8961	102	31	our	-PRON-	PRP$
8961	102	32	comparison	comparison	NN
8961	102	33	on	on	IN
8961	102	34	the	the	DT
8961	102	35	total	total	JJ
8961	102	36	number	number	NN
8961	102	37	of	of	IN
8961	102	38	postings	posting	NNS
8961	102	39	in	in	IN
8961	102	40	the	the	DT
8961	102	41	file	file	NN
8961	102	42	.	.	.
8961	103	1	In	in	IN
8961	103	2	the	the	DT
8961	103	3	standard	standard	JJ
8961	103	4	inverted	invert	VBN
8961	103	5	file	file	NN
8961	103	6	the	the	DT
8961	103	7	storage	storage	NN
8961	103	8	space	space	NN
8961	103	9	for	for	IN
8961	103	10	the	the	DT
8961	103	11	postings	posting	NNS
8961	103	12	is	be	VBZ
8961	103	13	equal	equal	JJ
8961	103	14	to	to	IN
8961	103	15	the	the	DT
8961	103	16	number	number	NN
8961	103	17	of	of	IN
8961	103	18	postings	posting	NNS
8961	103	19	times	time	NNS
8961	103	20	the	the	DT
8961	103	21	length	length	NN
8961	103	22	of	of	IN
8961	103	23	a	a	DT
8961	103	24	sin-	sin-	RB
8961	103	25	gle	gle	JJ
8961	103	26	posting	posting	NN
8961	103	27	,	,	,
8961	103	28	which	which	WDT
8961	103	29	is	be	VBZ
8961	103	30	usually	usually	RB
8961	103	31	two	two	CD
8961	103	32	,	,	,
8961	103	33	three	three	CD
8961	103	34	,	,	,
8961	103	35	or	or	CC
8961	103	36	five	five	CD
8961	103	37	bytes	byte	NNS
8961	103	38	.	.	.
8961	104	1	Memory	memory	NN
8961	104	2	requirements	requirement	NNS
8961	104	3	for	for	IN
8961	104	4	the	the	DT
8961	104	5	compressed	compressed	JJ
8961	104	6	binary	binary	JJ
8961	104	7	vector	vector	NN
8961	104	8	file	file	NN
8961	104	9	are	be	VBP
8961	104	10	more	more	RBR
8961	104	11	difficult	difficult	JJ
8961	104	12	to	to	TO
8961	104	13	estimate	estimate	VB
8961	104	14	because	because	IN
8961	104	15	the	the	DT
8961	104	16	distribution	distribution	NN
8961	104	17	of	of	IN
8961	104	18	document	document	NN
8961	104	19	numbers	number	NNS
8961	104	20	within	within	IN
8961	104	21	the	the	DT
8961	104	22	record	record	NN
8961	104	23	for	for	IN
8961	104	24	each	each	DT
8961	104	25	index	index	NN
8961	104	26	term	term	NN
8961	104	27	is	be	VBZ
8961	104	28	not	not	RB
8961	104	29	known	know	VBN
8961	104	30	.	.	.
8961	105	1	The	the	DT
8961	105	2	fact	fact	NN
8961	105	3	that	that	IN
8961	105	4	a	a	DT
8961	105	5	single	single	JJ
8961	105	6	byte	byte	NN
8961	105	7	in	in	IN
8961	105	8	the	the	DT
8961	105	9	binary	binary	JJ
8961	105	10	vector	vector	NN
8961	105	11	file	file	NN
8961	105	12	may	may	MD
8961	105	13	contain	contain	VB
8961	105	14	between	between	IN
8961	105	15	zero	zero	CD
8961	105	16	and	and	CC
8961	105	17	eight	eight	CD
8961	105	18	postings	posting	NNS
8961	105	19	is	be	VBZ
8961	105	20	extremely	extremely	RB
8961	105	21	important	important	JJ
8961	105	22	.	.	.
8961	106	1	The	the	DT
8961	106	2	worst	bad	JJS
8961	106	3	possible	possible	JJ
8961	106	4	case	case	NN
8961	106	5	occurs	occur	VBZ
8961	106	6	if	if	IN
8961	106	7	the	the	DT
8961	106	8	postings	posting	NNS
8961	106	9	in	in	IN
8961	106	10	the	the	DT
8961	106	11	binary	binary	JJ
8961	106	12	vector	vector	NN
8961	106	13	are	be	VBP
8961	106	14	spaced	space	VBN
8961	106	15	in	in	IN
8961	106	16	such	such	PDT
8961	106	17	a	a	DT
8961	106	18	way	way	NN
8961	106	19	that	that	IN
8961	106	20	each	each	DT
8961	106	21	nonzero	nonzero	NN
8961	106	22	byte	byte	NN
8961	106	23	contains	contain	VBZ
8961	106	24	only	only	RB
8961	106	25	one	one	CD
8961	106	26	posting	posting	NN
8961	106	27	,	,	,
8961	106	28	and	and	CC
8961	106	29	these	these	DT
8961	106	30	bytes	byte	NNS
8961	106	31	are	be	VBP
8961	106	32	separated	separate	VBN
8961	106	33	by	by	IN
8961	106	34	zero	zero	CD
8961	106	35	bytes	byte	NNS
8961	106	36	.	.	.
8961	107	1	Consider	consider	VB
8961	107	2	the	the	DT
8961	107	3	following	follow	VBG
8961	107	4	example	example	NN
8961	107	5	:	:	:
8961	107	6	...	...	NFP
8961	107	7	/00000000/00010000/00000000/00000100/	/00000000/00010000/00000000/00000100/	NFP
8961	107	8	...	...	.
8961	108	1	In	in	IN
8961	108	2	this	this	DT
8961	108	3	case	case	NN
8961	108	4	the	the	DT
8961	108	5	compression	compression	NN
8961	108	6	algorithm	algorithm	NNP
8961	108	7	will	will	MD
8961	108	8	remove	remove	VB
8961	108	9	the	the	DT
8961	108	10	zero	zero	CD
8961	108	11	bytes	byte	NNS
8961	108	12	,	,	,
8961	108	13	but	but	CC
8961	108	14	will	will	MD
8961	108	15	add	add	VB
8961	108	16	two	two	CD
8961	108	17	bytes	byte	NNS
8961	108	18	(	(	-LRB-
8961	108	19	the	the	DT
8961	108	20	prefix	prefix	JJ
8961	108	21	bytes	byte	VBZ
8961	108	22	)	)	-RRB-
8961	108	23	for	for	IN
8961	108	24	each	each	DT
8961	108	25	nonzero	nonzero	NN
8961	108	26	byte	byte	NN
8961	108	27	.	.	.
8961	109	1	The	the	DT
8961	109	2	resulting	result	VBG
8961	109	3	com-	com-	NN
8961	109	4	pressed	press	VBN
8961	109	5	vector	vector	NN
8961	109	6	will	will	MD
8961	109	7	be	be	VB
8961	109	8	essentially	essentially	RB
8961	109	9	the	the	DT
8961	109	10	same	same	JJ
8961	109	11	length	length	NN
8961	109	12	as	as	IN
8961	109	13	the	the	DT
8961	109	14	standard	standard	JJ
8961	109	15	inverted	invert	VBN
8961	109	16	file	file	NN
8961	109	17	record	record	NN
8961	109	18	if	if	IN
8961	109	19	each	each	DT
8961	109	20	posting	post	VBG
8961	109	21	is	be	VBZ
8961	109	22	three	three	CD
8961	109	23	bytes	byte	NNS
8961	109	24	long	long	JJ
8961	109	25	in	in	IN
8961	109	26	the	the	DT
8961	109	27	standard	standard	JJ
8961	109	28	inverted	invert	VBN
8961	109	29	file	file	NN
8961	109	30	.	.	.
8961	110	1	It	-PRON-	PRP
8961	110	2	might	may	MD
8961	110	3	seem	seem	VB
8961	110	4	that	that	IN
8961	110	5	the	the	DT
8961	110	6	distribution	distribution	NN
8961	110	7	of	of	IN
8961	110	8	one	one	CD
8961	110	9	posting	post	VBG
8961	110	10	per	per	IN
8961	110	11	byte	byte	NN
8961	110	12	for	for	IN
8961	110	13	the	the	DT
8961	110	14	entire	entire	JJ
8961	110	15	vector	vector	NN
8961	110	16	represents	represent	VBZ
8961	110	17	an	an	DT
8961	110	18	even	even	RB
8961	110	19	worse	bad	JJR
8961	110	20	situation	situation	NN
8961	110	21	.	.	.
8961	111	1	It	-PRON-	PRP
8961	111	2	is	be	VBZ
8961	111	3	clear	clear	JJ
8961	111	4	that	that	IN
8961	111	5	the	the	DT
8961	111	6	compression	compression	NN
8961	111	7	algorithm	algorithm	NNP
8961	111	8	will	will	MD
8961	111	9	,	,	,
8961	111	10	in	in	IN
8961	111	11	this	this	DT
8961	111	12	case	case	NN
8961	111	13	,	,	,
8961	111	14	not	not	RB
8961	111	15	reduce	reduce	VB
8961	111	16	the	the	DT
8961	111	17	size	size	NN
8961	111	18	of	of	IN
8961	111	19	the	the	DT
8961	111	20	vector	vector	NN
8961	111	21	.	.	.
8961	112	1	However	however	RB
8961	112	2	,	,	,
8961	112	3	it	-PRON-	PRP
8961	112	4	must	must	MD
8961	112	5	be	be	VB
8961	112	6	remembered	remember	VBN
8961	112	7	that	that	IN
8961	112	8	in	in	IN
8961	112	9	the	the	DT
8961	112	10	standard	standard	JJ
8961	112	11	inverted	invert	VBN
8961	112	12	file	file	NN
8961	112	13	each	each	DT
8961	112	14	posting	posting	NN
8961	112	15	will	will	MD
8961	112	16	re-	re-	RB
8961	112	17	quire	quire	VB
8961	112	18	at	at	RB
8961	112	19	least	least	RBS
8961	112	20	two	two	CD
8961	112	21	bytes	byte	NNS
8961	112	22	and	and	CC
8961	112	23	perhaps	perhaps	RB
8961	112	24	three	three	CD
8961	112	25	bytes	byte	NNS
8961	112	26	.	.	.
8961	113	1	Thus	thus	RB
8961	113	2	,	,	,
8961	113	3	the	the	DT
8961	113	4	length	length	NN
8961	113	5	of	of	IN
8961	113	6	the	the	DT
8961	113	7	record	record	NN
8961	113	8	in	in	IN
8961	113	9	the	the	DT
8961	113	10	standard	standard	JJ
8961	113	11	inverted	invert	VBN
8961	113	12	file	file	NN
8961	113	13	is	be	VBZ
8961	113	14	two	two	CD
8961	113	15	or	or	CC
8961	113	16	three	three	CD
8961	113	17	times	time	NNS
8961	113	18	longer	long	JJR
8961	113	19	than	than	IN
8961	113	20	the	the	DT
8961	113	21	corresponding	correspond	VBG
8961	113	22	binary	binary	JJ
8961	113	23	vector	vector	NN
8961	113	24	regardless	regardless	RB
8961	113	25	of	of	IN
8961	113	26	compression	compression	NN
8961	113	27	.	.	.
8961	114	1	In	in	IN
8961	114	2	data	datum	NNS
8961	114	3	used	use	VBN
8961	114	4	in	in	IN
8961	114	5	two	two	CD
8961	114	6	model	model	NN
8961	114	7	retrieval	retrieval	NN
8961	114	8	systems	system	NNS
8961	114	9	prepared	prepare	VBD
8961	114	10	to	to	TO
8961	114	11	compare	compare	VB
8961	114	12	the	the	DT
8961	114	13	standard	standard	JJ
8961	114	14	inverted	invert	VBN
8961	114	15	file	file	NN
8961	114	16	and	and	CC
8961	114	17	the	the	DT
8961	114	18	binary	binary	JJ
8961	114	19	vector	vector	NN
8961	114	20	file	file	NN
8961	114	21	there	there	EX
8961	114	22	are	be	VBP
8961	114	23	6,121	6,121	CD
8961	114	24	documents	document	NNS
8961	114	25	with	with	IN
8961	114	26	a	a	DT
8961	114	27	total	total	NN
8961	114	28	of	of	IN
8961	114	29	94,542	94,542	CD
8961	114	30	postings	posting	NNS
8961	114	31	.	.	.
8961	115	1	An	an	DT
8961	115	2	examination	examination	NN
8961	115	3	of	of	IN
8961	115	4	the	the	DT
8961	115	5	binary	binary	JJ
8961	115	6	inverted	invert	VBN
8961	115	7	file	file	NN
8961	115	8	for	for	IN
8961	115	9	the	the	DT
8961	115	10	model	model	NN
8961	115	11	systems	system	NNS
8961	115	12	discloses	disclose	VBZ
8961	115	13	that	that	IN
8961	115	14	there	there	EX
8961	115	15	are	be	VBP
8961	115	16	only	only	RB
8961	115	17	55,311	55,311	CD
8961	115	18	nonzero	nonzero	NN
8961	115	19	bytes	byte	NNS
8961	115	20	in	in	IN
8961	115	21	the	the	DT
8961	115	22	binary	binary	JJ
8961	115	23	vector	vector	NN
8961	115	24	file	file	NN
8961	115	25	.	.	.
8961	116	1	Thus	thus	RB
8961	116	2	there	there	EX
8961	116	3	seems	seem	VBZ
8961	116	4	to	to	TO
8961	116	5	be	be	VB
8961	116	6	some	some	DT
8961	116	7	form	form	NN
8961	116	8	of	of	IN
8961	116	9	clustering	clustering	NN
8961	116	10	of	of	IN
8961	116	11	the	the	DT
8961	116	12	document	document	NN
8961	116	13	numbers	number	NNS
8961	116	14	in	in	IN
8961	116	15	each	each	DT
8961	116	16	index	index	NN
8961	116	17	term	term	NN
8961	116	18	record	record	NN
8961	116	19	.	.	.
8961	117	1	If	if	IN
8961	117	2	each	each	DT
8961	117	3	nonzero	nonzero	NN
8961	117	4	byte	byte	NN
8961	117	5	in	in	IN
8961	117	6	this	this	DT
8961	117	7	binary	binary	JJ
8961	117	8	vector	vector	NN
8961	117	9	is	be	VBZ
8961	117	10	isolated	isolate	VBN
8961	117	11	by	by	IN
8961	117	12	zero	zero	CD
8961	117	13	bytes	byte	NNS
8961	117	14	,	,	,
8961	117	15	two	two	CD
8961	117	16	prefix	prefix	JJ
8961	117	17	bytes	byte	NNS
8961	117	18	would	would	MD
8961	117	19	be	be	VB
8961	117	20	added	add	VBN
8961	117	21	for	for	IN
8961	117	22	each	each	DT
8961	117	23	byte	byte	NN
8961	117	24	.	.	.
8961	118	1	Thus	thus	RB
8961	118	2	the	the	DT
8961	118	3	total	total	JJ
8961	118	4	memory	memory	NN
8961	118	5	requirements	requirement	NNS
8961	118	6	for	for	IN
8961	118	7	the	the	DT
8961	118	8	postings	posting	NNS
8961	118	9	in	in	IN
8961	118	10	the	the	DT
8961	118	11	compressed	compress	VBN
8961	118	12	file	file	NN
8961	118	13	would	would	MD
8961	118	14	be	be	VB
8961	118	15	165,933	165,933	CD
8961	118	16	bytes	byte	NNS
8961	118	17	.	.	.
8961	119	1	Less	Less	JJR
8961	119	2	storage	storage	NN
8961	119	3	space	space	NN
8961	119	4	is	be	VBZ
8961	119	5	required	require	VBN
8961	119	6	if	if	IN
8961	119	7	some	some	DT
8961	119	8	nonzero	nonzero	NN
8961	119	9	bytes	byte	NNS
8961	119	10	are	be	VBP
8961	119	11	contiguous	contiguous	JJ
8961	119	12	.	.	.
8961	120	1	On	on	IN
8961	120	2	the	the	DT
8961	120	3	other	other	JJ
8961	120	4	hand	hand	NN
8961	120	5	,	,	,
8961	120	6	the	the	DT
8961	120	7	standard	standard	NN
8961	120	8	in-	in-	NNP
8961	120	9	verted	verte	VBN
8961	120	10	file	file	NN
8961	120	11	will	will	MD
8961	120	12	require	require	VB
8961	120	13	189,084	189,084	CD
8961	120	14	bytes	byte	NNS
8961	120	15	if	if	IN
8961	120	16	a	a	DT
8961	120	17	two	two	CD
8961	120	18	-	-	HYPH
8961	120	19	byte	byte	NN
8961	120	20	posting	posting	NN
8961	120	21	is	be	VBZ
8961	120	22	used	use	VBN
8961	120	23	,	,	,
8961	120	24	or	or	CC
8961	120	25	283,626	283,626	CD
8961	120	26	bytes	byte	NNS
8961	120	27	if	if	IN
8961	120	28	a	a	DT
8961	120	29	three	three	CD
8961	120	30	-	-	HYPH
8961	120	31	byte	byte	NN
8961	120	32	posting	posting	NN
8961	120	33	is	be	VBZ
8961	120	34	used	use	VBN
8961	120	35	.	.	.
8961	121	1	Further	further	JJ
8961	121	2	study	study	NN
8961	121	3	of	of	IN
8961	121	4	the	the	DT
8961	121	5	cluster-	cluster-	XX
8961	121	6	ing	ing	NNP
8961	121	7	phenomenon	phenomenon	NN
8961	121	8	is	be	VBZ
8961	121	9	needed	need	VBN
8961	121	10	.	.	.
8961	122	1	Binary	Binary	NNP
8961	122	2	Vector	Vector	NNP
8961	122	3	/KING	/KING	.
8961	122	4	313	313	CD
8961	122	5	MODEL	MODEL	NNP
8961	122	6	RETRIEVAL	RETRIEVAL	NNP
8961	122	7	SYSTEMS	systems	NN
8961	122	8	To	to	TO
8961	122	9	test	test	VB
8961	122	10	some	some	DT
8961	122	11	of	of	IN
8961	122	12	the	the	DT
8961	122	13	conjectures	conjecture	NNS
8961	122	14	about	about	IN
8961	122	15	the	the	DT
8961	122	16	differences	difference	NNS
8961	122	17	between	between	IN
8961	122	18	the	the	DT
8961	122	19	stan-	stan-	NNP
8961	122	20	dard	dard	NNP
8961	122	21	inverted	invert	VBN
8961	122	22	file	file	NN
8961	122	23	and	and	CC
8961	122	24	the	the	DT
8961	122	25	binary	binary	JJ
8961	122	26	vector	vector	NN
8961	122	27	file	file	NN
8961	122	28	,	,	,
8961	122	29	two	two	CD
8961	122	30	model	model	NN
8961	122	31	systems	system	NNS
8961	122	32	were	be	VBD
8961	122	33	pre-	pre-	RB
8961	122	34	pared	pare	VBN
8961	122	35	for	for	IN
8961	122	36	operation	operation	NN
8961	122	37	on	on	IN
8961	122	38	an	an	DT
8961	122	39	IBM	IBM	NNP
8961	122	40	360/67	360/67	CD
8961	122	41	.	.	.
8961	123	1	Details	detail	NNS
8961	123	2	of	of	IN
8961	123	3	the	the	DT
8961	123	4	systems	system	NNS
8961	123	5	and	and	CC
8961	123	6	PL/1	PL/1	NNP
8961	123	7	program	program	NN
8961	123	8	listings	listing	NNS
8961	123	9	are	be	VBP
8961	123	10	available	available	JJ
8961	123	11	elsewhere	elsewhere	RB
8961	123	12	.	.	.
8961	124	1	7	7	CD
8961	124	2	The	the	DT
8961	124	3	data	datum	NNS
8961	124	4	base	base	NN
8961	124	5	used	use	VBN
8961	124	6	was	be	VBD
8961	124	7	obtained	obtain	VBN
8961	124	8	from	from	IN
8961	124	9	the	the	DT
8961	124	10	Institute	Institute	NNP
8961	124	11	of	of	IN
8961	124	12	Animal	Animal	NNP
8961	124	13	Behavior	Behavior	NNP
8961	124	14	at	at	IN
8961	124	15	Rutgers	Rutgers	NNP
8961	124	16	University	University	NNP
8961	124	17	.	.	.
8961	125	1	In	in	IN
8961	125	2	the	the	DT
8961	125	3	data	datum	NNS
8961	125	4	base	base	NN
8961	125	5	6,121	6,121	CD
8961	125	6	documents	document	NNS
8961	125	7	were	be	VBD
8961	125	8	indexed	index	VBN
8961	125	9	by	by	IN
8961	125	10	1,484	1,484	CD
8961	125	11	index	index	NN
8961	125	12	terms	term	NNS
8961	125	13	.	.	.
8961	126	1	A	a	DT
8961	126	2	total	total	NN
8961	126	3	of	of	IN
8961	126	4	94,542	94,542	CD
8961	126	5	postings	posting	NNS
8961	126	6	in	in	IN
8961	126	7	the	the	DT
8961	126	8	system	system	NN
8961	126	9	gives	give	VBZ
8961	126	10	an	an	DT
8961	126	11	average	average	JJ
8961	126	12	depth	depth	NN
8961	126	13	of	of	IN
8961	126	14	indexing	indexing	NN
8961	126	15	of	of	IN
8961	126	16	15.4	15.4	CD
8961	126	17	terms	term	NNS
8961	126	18	per	per	IN
8961	126	19	document	document	NN
8961	126	20	.	.	.
8961	127	1	Both	both	DT
8961	127	2	inverted	invert	VBN
8961	127	3	files	file	NNS
8961	127	4	were	be	VBD
8961	127	5	stored	store	VBN
8961	127	6	on	on	IN
8961	127	7	IBM	IBM	NNP
8961	127	8	2314	2314	CD
8961	127	9	disc	disc	NN
8961	127	10	storage	storage	NN
8961	127	11	devices	device	NNS
8961	127	12	.	.	.
8961	128	1	To	to	TO
8961	128	2	ease	ease	VB
8961	128	3	the	the	DT
8961	128	4	problem	problem	NN
8961	128	5	of	of	IN
8961	128	6	handling	handle	VBG
8961	128	7	variable	variable	JJ
8961	128	8	-	-	HYPH
8961	128	9	length	length	NN
8961	128	10	records	record	NNS
8961	128	11	in	in	IN
8961	128	12	both	both	DT
8961	128	13	files	file	NNS
8961	128	14	the	the	DT
8961	128	15	logical	logical	JJ
8961	128	16	records	record	NNS
8961	128	17	for	for	IN
8961	128	18	each	each	DT
8961	128	19	index	index	NN
8961	128	20	term	term	NN
8961	128	21	were	be	VBD
8961	128	22	divided	divide	VBN
8961	128	23	into	into	IN
8961	128	24	chains	chain	NNS
8961	128	25	of	of	IN
8961	128	26	fixed~	fixed~	NNP
8961	128	27	lehgth	lehgth	JJ
8961	128	28	physical	physical	JJ
8961	128	29	records	record	NNS
8961	128	30	.	.	.
8961	129	1	For	for	IN
8961	129	2	the	the	DT
8961	129	3	standard	standard	JJ
8961	129	4	inverted	invert	VBN
8961	129	5	file	file	NN
8961	129	6	a	a	DT
8961	129	7	physical	physical	JJ
8961	129	8	record	record	NN
8961	129	9	size	size	NN
8961	129	10	of	of	IN
8961	129	11	331	331	CD
8961	129	12	bytes	byte	NNS
8961	129	13	was	be	VBD
8961	129	14	chosen	choose	VBN
8961	129	15	.	.	.
8961	130	1	The	the	DT
8961	130	2	entire	entire	JJ
8961	130	3	file	file	NN
8961	130	4	required	require	VBD
8961	130	5	702,713	702,713	CD
8961	130	6	bytes	byte	NNS
8961	130	7	including	include	VBG
8961	130	8	record	record	NN
8961	130	9	overhead	overhead	RB
8961	130	10	.	.	.
8961	131	1	For	for	IN
8961	131	2	the	the	DT
8961	131	3	uncompressed	uncompressed	JJ
8961	131	4	binary	binary	JJ
8961	131	5	vector	vector	NN
8961	131	6	file	file	NN
8961	131	7	a	a	DT
8961	131	8	physical	physical	JJ
8961	131	9	record	record	NN
8961	131	10	size	size	NN
8961	131	11	of	of	IN
8961	131	12	1,286	1,286	CD
8961	131	13	bytes	byte	NNS
8961	131	14	was	be	VBD
8961	131	15	chosen	choose	VBN
8961	131	16	to	to	TO
8961	131	17	include	include	VB
8961	131	18	overhead	overhead	NN
8961	131	19	and	and	CC
8961	131	20	space	space	VB
8961	131	21	for	for	IN
8961	131	22	up	up	IN
8961	131	23	to	to	TO
8961	131	24	10,216	10,216	CD
8961	131	25	document	document	NN
8961	131	26	numbers	number	NNS
8961	131	27	.	.	.
8961	132	1	When	when	WRB
8961	132	2	the	the	DT
8961	132	3	compression	compression	NN
8961	132	4	algorithm	algorithm	NNP
8961	132	5	was	be	VBD
8961	132	6	applied	apply	VBN
8961	132	7	,	,	,
8961	132	8	with	with	IN
8961	132	9	a	a	DT
8961	132	10	physical	physical	JJ
8961	132	11	record	record	NN
8961	132	12	length	length	NN
8961	132	13	of	of	IN
8961	132	14	130	130	CD
8961	132	15	bytes	byte	NNS
8961	132	16	,	,	,
8961	132	17	the	the	DT
8961	132	18	memory	memory	NN
8961	132	19	requirements	requirement	NNS
8961	132	20	for	for	IN
8961	132	21	the	the	DT
8961	132	22	binary	binary	JJ
8961	132	23	vector	vector	NN
8961	132	24	file	file	NN
8961	132	25	were	be	VBD
8961	132	26	reduced	reduce	VBN
8961	132	27	to	to	IN
8961	132	28	281,450	281,450	CD
8961	132	29	bytes	byte	NNS
8961	132	30	,	,	,
8961	132	31	or	or	CC
8961	132	32	41	41	CD
8961	132	33	percent	percent	NN
8961	132	34	of	of	IN
8961	132	35	the	the	DT
8961	132	36	space	space	NN
8961	132	37	required	require	VBN
8961	132	38	to	to	TO
8961	132	39	store	store	VB
8961	132	40	the	the	DT
8961	132	41	standard	standard	JJ
8961	132	42	inverted	invert	VBN
8961	132	43	file	file	NN
8961	132	44	.	.	.
8961	133	1	A	a	DT
8961	133	2	series	series	NN
8961	133	3	of	of	IN
8961	133	4	forty	forty	CD
8961	133	5	searches	search	NNS
8961	133	6	of	of	IN
8961	133	7	varying	varying	NN
8961	133	8	complexities	complexity	NNS
8961	133	9	were	be	VBD
8961	133	10	run	run	VBN
8961	133	11	against	against	IN
8961	133	12	both	both	DT
8961	133	13	files	file	NNS
8961	133	14	.	.	.
8961	134	1	The	the	DT
8961	134	2	"	"	``
8961	134	3	TIME	time	NN
8961	134	4	"	"	''
8961	134	5	function	function	NN
8961	134	6	of	of	IN
8961	134	7	PL/1	PL/1	NNP
8961	134	8	made	make	VBD
8961	134	9	it	-PRON-	PRP
8961	134	10	possible	possible	JJ
8961	134	11	to	to	TO
8961	134	12	accumulate	accumulate	VB
8961	134	13	tim-	tim-	NNS
8961	134	14	ing	ing	NNP
8961	134	15	statistics	statistic	NNS
8961	134	16	which	which	WDT
8961	134	17	excluded	exclude	VBD
8961	134	18	input	input	NN
8961	134	19	/	/	SYM
8961	134	20	output	output	NN
8961	134	21	functions	function	NNS
8961	134	22	.	.	.
8961	135	1	Search	search	NN
8961	135	2	times	time	NNS
8961	135	3	for	for	IN
8961	135	4	the	the	DT
8961	135	5	binary	binary	JJ
8961	135	6	vector	vector	NN
8961	135	7	file	file	NN
8961	135	8	include	include	VBP
8961	135	9	expansion	expansion	NN
8961	135	10	of	of	IN
8961	135	11	the	the	DT
8961	135	12	compressed	compress	VBN
8961	135	13	vectors	vector	NNS
8961	135	14	,	,	,
8961	135	15	Boolean	boolean	JJ
8961	135	16	manipulation	manipulation	NN
8961	135	17	of	of	IN
8961	135	18	the	the	DT
8961	135	19	vectors	vector	NNS
8961	135	20	,	,	,
8961	135	21	and	and	CC
8961	135	22	conversion	conversion	NN
8961	135	23	of	of	IN
8961	135	24	the	the	DT
8961	135	25	resultant	resultant	JJ
8961	135	26	vector	vector	NN
8961	135	27	into	into	IN
8961	135	28	digital	digital	JJ
8961	135	29	document	document	NN
8961	135	30	numbers	number	NNS
8961	135	31	.	.	.
8961	136	1	The	the	DT
8961	136	2	times	time	NNS
8961	136	3	for	for	IN
8961	136	4	the	the	DT
8961	136	5	standard	standard	JJ
8961	136	6	inverted	invert	VBN
8961	136	7	file	file	NN
8961	136	8	are	be	VBP
8961	136	9	for	for	IN
8961	136	10	the	the	DT
8961	136	11	Boolean	boolean	JJ
8961	136	12	manipulation	manipulation	NN
8961	136	13	of	of	IN
8961	136	14	the	the	DT
8961	136	15	lists	list	NNS
8961	136	16	.	.	.
8961	137	1	The	the	DT
8961	137	2	following	follow	VBG
8961	137	3	points	point	NNS
8961	137	4	were	be	VBD
8961	137	5	noted	note	VBN
8961	137	6	in	in	IN
8961	137	7	the	the	DT
8961	137	8	analysis	analysis	NN
8961	137	9	of	of	IN
8961	137	10	the	the	DT
8961	137	11	times	time	NNS
8961	137	12	:	:	:
8961	137	13	1	1	CD
8961	137	14	.	.	.
8961	138	1	In	in	IN
8961	138	2	twenty	twenty	CD
8961	138	3	-	-	HYPH
8961	138	4	two	two	CD
8961	138	5	of	of	IN
8961	138	6	the	the	DT
8961	138	7	forty	forty	CD
8961	138	8	queries	query	NNS
8961	138	9	for	for	IN
8961	138	10	which	which	WDT
8961	138	11	comparative	comparative	JJ
8961	138	12	timings	timing	NNS
8961	138	13	were	be	VBD
8961	138	14	obtained	obtain	VBN
8961	138	15	,	,	,
8961	138	16	the	the	DT
8961	138	17	search	search	NN
8961	138	18	of	of	IN
8961	138	19	the	the	DT
8961	138	20	binary	binary	JJ
8961	138	21	vector	vector	NN
8961	138	22	file	file	NN
8961	138	23	was	be	VBD
8961	138	24	faster	fast	RBR
8961	138	25	,	,	,
8961	138	26	in	in	IN
8961	138	27	one	one	CD
8961	138	28	case	case	NN
8961	138	29	by	by	IN
8961	138	30	a	a	DT
8961	138	31	factor	factor	NN
8961	138	32	of	of	IN
8961	138	33	thirty	thirty	CD
8961	138	34	-	-	HYPH
8961	138	35	five	five	CD
8961	138	36	.	.	.
8961	139	1	In	in	IN
8961	139	2	the	the	DT
8961	139	3	eighteen	eighteen	CD
8961	139	4	cases	case	NNS
8961	139	5	in	in	IN
8961	139	6	which	which	WDT
8961	139	7	the	the	DT
8961	139	8	search	search	NN
8961	139	9	of	of	IN
8961	139	10	the	the	DT
8961	139	11	standard	standard	JJ
8961	139	12	inverted	invert	VBN
8961	139	13	file	file	NN
8961	139	14	was	be	VBD
8961	139	15	faster	fast	JJR
8961	139	16	,	,	,
8961	139	17	the	the	DT
8961	139	18	search	search	NN
8961	139	19	of	of	IN
8961	139	20	the	the	DT
8961	139	21	stan-	stan-	NNP
8961	139	22	dard	dard	NNP
8961	139	23	inverted	invert	VBN
8961	139	24	file	file	NN
8961	139	25	was	be	VBD
8961	139	26	at	at	IN
8961	139	27	most	most	JJS
8961	139	28	6.17	6.17	CD
8961	139	29	times	time	NNS
8961	139	30	faster	fast	RBR
8961	139	31	.	.	.
8961	140	1	2	2	LS
8961	140	2	.	.	.
8961	141	1	The	the	DT
8961	141	2	range	range	NN
8961	141	3	of	of	IN
8961	141	4	the	the	DT
8961	141	5	total	total	JJ
8961	141	6	times	time	NNS
8961	141	7	for	for	IN
8961	141	8	the	the	DT
8961	141	9	binary	binary	JJ
8961	141	10	vector	vector	NN
8961	141	11	file	file	NN
8961	141	12	was	be	VBD
8961	141	13	.79	.79	CD
8961	141	14	seconds	second	NNS
8961	141	15	to	to	IN
8961	141	16	9.72	9.72	CD
8961	141	17	seconds	second	NNS
8961	141	18	.	.	.
8961	142	1	The	the	DT
8961	142	2	range	range	NN
8961	142	3	for	for	IN
8961	142	4	searching	search	VBG
8961	142	5	the	the	DT
8961	142	6	standard	standard	JJ
8961	142	7	inverted	invert	VBN
8961	142	8	file	file	NN
8961	142	9	was	be	VBD
8961	142	10	.15	.15	CD
8961	142	11	seconds	second	NNS
8961	142	12	to	to	IN
8961	142	13	202.98	202.98	CD
8961	142	14	seconds	second	NNS
8961	142	15	.	.	.
8961	143	1	The	the	DT
8961	143	2	fact	fact	NN
8961	143	3	that	that	IN
8961	143	4	the	the	DT
8961	143	5	search	search	NN
8961	143	6	times	time	NNS
8961	143	7	for	for	IN
8961	143	8	the	the	DT
8961	143	9	binary	binary	JJ
8961	143	10	vector	vector	NN
8961	143	11	file	file	NN
8961	143	12	are	be	VBP
8961	143	13	within	within	IN
8961	143	14	a	a	DT
8961	143	15	fairly	fairly	RB
8961	143	16	narrow	narrow	JJ
8961	143	17	range	range	NN
8961	143	18	,	,	,
8961	143	19	in	in	IN
8961	143	20	contrast	contrast	NN
8961	143	21	to	to	IN
8961	143	22	the	the	DT
8961	143	23	wider	wide	JJR
8961	143	24	range	range	NN
8961	143	25	of	of	IN
8961	143	26	times	time	NNS
8961	143	27	for	for	IN
8961	143	28	searching	search	VBG
8961	143	29	the	the	DT
8961	143	30	standard	standard	JJ
8961	143	31	inverted	invert	VBN
8961	143	32	file	file	NN
8961	143	33	,	,	,
8961	143	34	has	have	VBZ
8961	143	35	im-	im-	DT
8961	143	36	portant	portant	JJ
8961	143	37	implications	implication	NNS
8961	143	38	for	for	IN
8961	143	39	the	the	DT
8961	143	40	design	design	NN
8961	143	41	of	of	IN
8961	143	42	an	an	DT
8961	143	43	on	on	IN
8961	143	44	-	-	HYPH
8961	143	45	line	line	NN
8961	143	46	interactive	interactive	JJ
8961	143	47	docu-	docu-	NN
8961	143	48	ment	ment	JJ
8961	143	49	retrieval	retrieval	NN
8961	143	50	system	system	NN
8961	143	51	.	.	.
8961	144	1	In	in	IN
8961	144	2	such	such	PDT
8961	144	3	a	a	DT
8961	144	4	system	system	NN
8961	144	5	it	-PRON-	PRP
8961	144	6	is	be	VBZ
8961	144	7	important	important	JJ
8961	144	8	that	that	IN
8961	144	9	the	the	DT
8961	144	10	com-	com-	NN
8961	144	11	puter	puter	NN
8961	144	12	respond	respond	NN
8961	144	13	to	to	IN
8961	144	14	users	user	NNS
8961	144	15	'	'	POS
8961	144	16	requests	request	NNS
8961	144	17	not	not	RB
8961	144	18	only	only	RB
8961	144	19	rapidly	rapidly	RB
8961	144	20	but	but	CC
8961	144	21	consistently	consistently	RB
8961	144	22	.	.	.
8961	145	1	The	the	DT
8961	145	2	narrower	narrow	JJR
8961	145	3	range	range	NN
8961	145	4	of	of	IN
8961	145	5	the	the	DT
8961	145	6	search	search	NN
8961	145	7	times	time	NNS
8961	145	8	provided	provide	VBN
8961	145	9	by	by	IN
8961	145	10	the	the	DT
8961	145	11	binary	binary	JJ
8961	145	12	vector	vector	NN
8961	145	13	file	file	NN
8961	145	14	will	will	MD
8961	145	15	assist	assist	VB
8961	145	16	in	in	IN
8961	145	17	producing	produce	VBG
8961	145	18	consistent	consistent	JJ
8961	145	19	times	time	NNS
8961	145	20	.	.	.
8961	146	1	3	3	LS
8961	146	2	.	.	.
8961	147	1	The	the	DT
8961	147	2	search	search	NN
8961	147	3	times	time	NNS
8961	147	4	for	for	IN
8961	147	5	the	the	DT
8961	147	6	binary	binary	JJ
8961	147	7	vector	vector	NN
8961	147	8	file	file	NN
8961	147	9	,	,	,
8961	147	10	exclusive	exclusive	JJ
8961	147	11	of	of	IN
8961	147	12	expansion	expansion	NN
8961	147	13	and	and	CC
8961	147	14	conversion	conversion	NN
8961	147	15	times	time	NNS
8961	147	16	,	,	,
8961	147	17	are	be	VBP
8961	147	18	unaffected	unaffected	JJ
8961	147	19	by	by	IN
8961	147	20	the	the	DT
8961	147	21	number	number	NN
8961	147	22	of	of	IN
8961	147	23	postings	posting	NNS
8961	147	24	con-	con-	NN
8961	147	25	314	314	CD
8961	147	26	Journal	Journal	NNP
8961	147	27	of	of	IN
8961	147	28	Library	Library	NNP
8961	147	29	Automation	Automation	NNP
8961	147	30	Vol	Vol	NNP
8961	147	31	.	.	.
8961	148	1	7/4	7/4	CD
8961	148	2	December	December	NNP
8961	148	3	1974	1974	CD
8961	148	4	tained	taine	VBD
8961	148	5	in	in	IN
8961	148	6	the	the	DT
8961	148	7	index	index	NN
8961	148	8	terms	term	NNS
8961	148	9	used	use	VBN
8961	148	10	in	in	IN
8961	148	11	a	a	DT
8961	148	12	query	query	NN
8961	148	13	.	.	.
8961	149	1	On	on	IN
8961	149	2	the	the	DT
8961	149	3	other	other	JJ
8961	149	4	hand	hand	NN
8961	149	5	,	,	,
8961	149	6	the	the	DT
8961	149	7	number	number	NN
8961	149	8	of	of	IN
8961	149	9	postings	posting	NNS
8961	149	10	in	in	IN
8961	149	11	the	the	DT
8961	149	12	records	record	NNS
8961	149	13	used	use	VBN
8961	149	14	from	from	IN
8961	149	15	the	the	DT
8961	149	16	standard	standard	JJ
8961	149	17	inverted	invert	VBN
8961	149	18	file	file	NN
8961	149	19	appears	appear	VBZ
8961	149	20	to	to	TO
8961	149	21	cause	cause	VB
8961	149	22	the	the	DT
8961	149	23	differences	difference	NNS
8961	149	24	in	in	IN
8961	149	25	search	search	NN
8961	149	26	times	time	NNS
8961	149	27	for	for	IN
8961	149	28	that	that	DT
8961	149	29	file	file	NN
8961	149	30	.	.	.
8961	150	1	To	to	TO
8961	150	2	test	test	VB
8961	150	3	the	the	DT
8961	150	4	conjectures	conjecture	NNS
8961	150	5	!	!	.
8961	151	1	that	that	IN
8961	151	2	1	1	CD
8961	151	3	.	.	.
8961	151	4	search	search	NN
8961	151	5	times	time	NNS
8961	151	6	for	for	IN
8961	151	7	the	the	DT
8961	151	8	binary	binary	JJ
8961	151	9	vector	vector	NN
8961	151	10	file	file	NN
8961	151	11	are	be	VBP
8961	151	12	related	relate	VBN
8961	151	13	to	to	IN
8961	151	14	the	the	DT
8961	151	15	number	number	NN
8961	151	16	of	of	IN
8961	151	17	index	index	NN
8961	151	18	terms	term	NNS
8961	151	19	in	in	IN
8961	151	20	the	the	DT
8961	151	21	query	query	NN
8961	151	22	,	,	,
8961	151	23	and	and	CC
8961	151	24	2	2	CD
8961	151	25	.	.	.
8961	151	26	search	search	NN
8961	151	27	times	time	NNS
8961	151	28	for	for	IN
8961	151	29	the	the	DT
8961	151	30	standard	standard	JJ
8961	151	31	inverted	invert	VBN
8961	151	32	file	file	NN
8961	151	33	are	be	VBP
8961	151	34	related	relate	VBN
8961	151	35	to	to	IN
8961	151	36	the	the	DT
8961	151	37	num-	num-	NNP
8961	151	38	ber	ber	NN
8961	151	39	of	of	IN
8961	151	40	postings	posting	NNS
8961	151	41	in	in	IN
8961	151	42	the	the	DT
8961	151	43	index	index	NN
8961	151	44	terms	term	NNS
8961	151	45	in	in	IN
8961	151	46	the	the	DT
8961	151	47	query	query	NN
8961	151	48	,	,	,
8961	151	49	a	a	DT
8961	151	50	correlation	correlation	NN
8961	151	51	analysis	analysis	NN
8961	151	52	was	be	VBD
8961	151	53	performed	perform	VBN
8961	151	54	.	.	.
8961	152	1	The	the	DT
8961	152	2	following	following	JJ
8961	152	3	correlation	correlation	NN
8961	152	4	co-	co-	JJ
8961	152	5	efficients	efficient	NNS
8961	152	6	were	be	VBD
8961	152	7	obtained	obtain	VBN
8961	152	8	:	:	:
8961	152	9	V	v	NN
8961	152	10	a1'iables	a1'iable	NNS
8961	152	11	1	1	CD
8961	152	12	'	'	''
8961	152	13	Number	number	NN
8961	152	14	of	of	IN
8961	152	15	terms	term	NNS
8961	152	16	in	in	IN
8961	152	17	query	query	NN
8961	152	18	and	and	CC
8961	152	19	search	search	NN
8961	152	20	.960	.960	CD
8961	152	21	times	time	NNS
8961	152	22	for	for	IN
8961	152	23	the	the	DT
8961	152	24	binary	binary	JJ
8961	152	25	vector	vector	NN
8961	152	26	file	file	NN
8961	152	27	.	.	.
8961	153	1	Number	number	NN
8961	153	2	of	of	IN
8961	153	3	postings	posting	NNS
8961	153	4	in	in	IN
8961	153	5	query	query	NN
8961	153	6	terms	term	NNS
8961	153	7	and	and	CC
8961	153	8	.979	.979	CD
8961	153	9	search	search	NN
8961	153	10	times	time	NNS
8961	153	11	for	for	IN
8961	153	12	standard	standard	JJ
8961	153	13	inverted	invert	VBN
8961	153	14	file	file	NN
8961	153	15	.	.	.
8961	154	1	The	the	DT
8961	154	2	relationships	relationship	NNS
8961	154	3	indicated	indicate	VBD
8961	154	4	above	above	RB
8961	154	5	are	be	VBP
8961	154	6	significant	significant	JJ
8961	154	7	at	at	IN
8961	154	8	the	the	DT
8961	154	9	.001	.001	CD
8961	154	10	level	level	NN
8961	154	11	.	.	.
8961	155	1	No	no	DT
8961	155	2	attempt	attempt	NN
8961	155	3	was	be	VBD
8961	155	4	made	make	VBN
8961	155	5	to	to	TO
8961	155	6	compute	compute	VB
8961	155	7	an	an	DT
8961	155	8	average	average	JJ
8961	155	9	search	search	NN
8961	155	10	time	time	NN
8961	155	11	per	per	IN
8961	155	12	term	term	NN
8961	155	13	for	for	IN
8961	155	14	the	the	DT
8961	155	15	binary	binary	JJ
8961	155	16	vector	vector	NN
8961	155	17	file	file	NN
8961	155	18	or	or	CC
8961	155	19	average	average	JJ
8961	155	20	search	search	NN
8961	155	21	time	time	NN
8961	155	22	per	per	IN
8961	155	23	posting	post	VBG
8961	155	24	for	for	IN
8961	155	25	the	the	DT
8961	155	26	standard	standard	JJ
8961	155	27	in-	in-	NNP
8961	155	28	verted	verted	JJ
8961	155	29	file	file	NN
8961	155	30	.	.	.
8961	156	1	Such	such	JJ
8961	156	2	times	time	NNS
8961	156	3	would	would	MD
8961	156	4	have	have	VB
8961	156	5	meaning	meaning	NN
8961	156	6	only	only	RB
8961	156	7	for	for	IN
8961	156	8	the	the	DT
8961	156	9	model	model	NN
8961	156	10	systems	system	NNS
8961	156	11	.	.	.
8961	157	1	SUMMARY	summary	VB
8961	157	2	The	the	DT
8961	157	3	binary	binary	JJ
8961	157	4	vector	vector	NN
8961	157	5	is	be	VBZ
8961	157	6	suggested	suggest	VBN
8961	157	7	as	as	IN
8961	157	8	an	an	DT
8961	157	9	alternative	alternative	NN
8961	157	10	to	to	IN
8961	157	11	the	the	DT
8961	157	12	usual	usual	JJ
8961	157	13	method	method	NN
8961	157	14	of	of	IN
8961	157	15	storing	store	VBG
8961	157	16	document	document	NN
8961	157	17	pointers	pointer	NNS
8961	157	18	in	in	IN
8961	157	19	an	an	DT
8961	157	20	inverted	invert	VBN
8961	157	21	index	index	NN
8961	157	22	file	file	NN
8961	157	23	.	.	.
8961	158	1	The	the	DT
8961	158	2	binary	binary	JJ
8961	158	3	vector	vector	NN
8961	158	4	file	file	NN
8961	158	5	can	can	MD
8961	158	6	provide	provide	VB
8961	158	7	savings	saving	NNS
8961	158	8	in	in	IN
8961	158	9	storage	storage	NN
8961	158	10	space	space	NN
8961	158	11	,	,	,
8961	158	12	search	search	NN
8961	158	13	times	time	NNS
8961	158	14	,	,	,
8961	158	15	and	and	CC
8961	158	16	programming	programming	NN
8961	158	17	effort	effort	NN
8961	158	18	.	.	.
8961	159	1	REFERENCES	reference	NNS
8961	159	2	1	1	CD
8961	159	3	.	.	.
8961	160	1	Burton	Burton	NNP
8961	160	2	H.	H.	NNP
8961	160	3	Bloom	Bloom	NNP
8961	160	4	,	,	,
8961	160	5	"	"	''
8961	160	6	Some	some	DT
8961	160	7	Techniques	Techniques	NNPS
8961	160	8	and	and	CC
8961	160	9	Trade	Trade	NNP
8961	160	10	-	-	HYPH
8961	160	11	Offs	Offs	NNP
8961	160	12	Affecting	Affecting	NNP
8961	160	13	Large	Large	NNP
8961	160	14	Data	Data	NNP
8961	160	15	Base	Base	NNP
8961	160	16	Retrieval	Retrieval	NNP
8961	160	17	Times	Times	NNP
8961	160	18	,	,	,
8961	160	19	"	"	``
8961	160	20	Proceedings	proceeding	NNS
8961	160	21	of	of	IN
8961	160	22	the	the	DT
8961	160	23	ACM	ACM	NNP
8961	160	24	24	24	CD
8961	160	25	(	(	-LRB-
8961	160	26	1969	1969	CD
8961	160	27	)	)	-RRB-
8961	160	28	.	.	.
8961	161	1	2	2	LS
8961	161	2	.	.	.
8961	162	1	D.	D.	NNP
8961	162	2	R.	R.	NNP
8961	162	3	Davis	Davis	NNP
8961	162	4	and	and	CC
8961	162	5	A.	A.	NNP
8961	162	6	D.	D.	NNP
8961	162	7	Lin	Lin	NNP
8961	162	8	,	,	,
8961	162	9	"	"	``
8961	162	10	Secondary	Secondary	NNP
8961	162	11	Key	Key	NNP
8961	162	12	Retrieval	Retrieval	NNP
8961	162	13	Using	use	VBG
8961	162	14	an	an	DT
8961	162	15	IBM	IBM	NNP
8961	162	16	7090	7090	CD
8961	162	17	-	-	HYPH
8961	162	18	1310	1310	CD
8961	162	19	System	system	NN
8961	162	20	,	,	,
8961	162	21	"	"	''
8961	162	22	Communications	communication	NNS
8961	162	23	of	of	IN
8961	162	24	the	the	DT
8961	162	25	ACM	ACM	NNP
8961	162	26	8:243	8:243	CD
8961	162	27	-	-	SYM
8961	162	28	46	46	CD
8961	162	29	(	(	-LRB-
8961	162	30	April1965	April1965	NNP
8961	162	31	)	)	-RRB-
8961	162	32	.	.	.
8961	163	1	3	3	LS
8961	163	2	.	.	.
8961	164	1	John	John	NNP
8961	164	2	W.	W.	NNP
8961	164	3	Sammon	Sammon	NNP
8961	164	4	,	,	,
8961	164	5	Some	some	DT
8961	164	6	Mathematics	Mathematics	NNPS
8961	164	7	of	of	IN
8961	164	8	Information	Information	NNP
8961	164	9	Storage	Storage	NNP
8961	164	10	and	and	CC
8961	164	11	Retrieval	Retrieval	NNP
8961	164	12	(	(	-LRB-
8961	164	13	Tech-	Tech-	NNP
8961	164	14	nical	nical	JJ
8961	164	15	Report	Report	NNP
8961	164	16	RADC	RADC	NNP
8961	164	17	-	-	HYPH
8961	164	18	Tr-68	tr-68	WP
8961	164	19	-	-	HYPH
8961	164	20	178	178	CD
8961	164	21	[	[	-LRB-
8961	164	22	Rome	Rome	NNP
8961	164	23	,	,	,
8961	164	24	New	New	NNP
8961	164	25	York	York	NNP
8961	164	26	:	:	:
8961	164	27	Rome	Rome	NNP
8961	164	28	Air	Air	NNP
8961	164	29	Development	Development	NNP
8961	164	30	Center	Center	NNP
8961	164	31	,	,	,
8961	164	32	1968	1968	CD
8961	164	33	]	]	-RRB-
8961	164	34	)	)	-RRB-
8961	164	35	.	.	.
8961	165	1	4	4	LS
8961	165	2	.	.	.
8961	166	1	S.	S.	NNP
8961	166	2	A.	A.	NNP
8961	166	3	Gorokhov	Gorokhov	NNP
8961	166	4	,	,	,
8961	166	5	"	"	``
8961	166	6	The	the	DT
8961	166	7	'	'	``
8961	166	8	Setka-3	Setka-3	NNP
8961	166	9	'	'	''
8961	166	10	Automated	Automated	NNP
8961	166	11	IRS	IRS	NNP
8961	166	12	on	on	IN
8961	166	13	the	the	DT
8961	166	14	'	'	``
8961	166	15	Minsk-22	minsk-22	NN
8961	166	16	'	'	''
8961	166	17	with	with	IN
8961	166	18	the	the	DT
8961	166	19	Use	Use	NNP
8961	166	20	of	of	IN
8961	166	21	the	the	DT
8961	166	22	Socket	Socket	NNP
8961	166	23	Associative	Associative	NNP
8961	166	24	-	-	HYPH
8961	166	25	Address	Address	NNP
8961	166	26	Method	Method	NNP
8961	166	27	of	of	IN
8961	166	28	Organization	Organization	NNP
8961	166	29	of	of	IN
8961	166	30	Information	Information	NNP
8961	166	31	"	"	''
8961	166	32	(	(	-LRB-
8961	166	33	Paper	paper	NN
8961	166	34	presented	present	VBD
8961	166	35	at	at	IN
8961	166	36	the	the	DT
8961	166	37	All	All	NNP
8961	166	38	-	-	HYPH
8961	166	39	Union	Union	NNP
8961	166	40	Conference	Conference	NNP
8961	166	41	on	on	IN
8961	166	42	Information	Information	NNP
8961	166	43	Retrieval	Retrieval	NNP
8961	166	44	Systems	Systems	NNP
8961	166	45	and	and	CC
8961	166	46	Auto-	Auto-	NNPS
8961	166	47	matic	matic	JJ
8961	166	48	Processing	Processing	NNP
8961	166	49	of	of	IN
8961	166	50	Scientific	Scientific	NNP
8961	166	51	and	and	CC
8961	166	52	Technical	Technical	NNP
8961	166	53	Information	Information	NNP
8961	166	54	,	,	,
8961	166	55	Moscow	Moscow	NNP
8961	166	56	,	,	,
8961	166	57	1967	1967	CD
8961	166	58	.	.	.
8961	167	1	Translated	translate	VBN
8961	167	2	and	and	CC
8961	167	3	published	publish	VBN
8961	167	4	as	as	IN
8961	167	5	part	part	NN
8961	167	6	of	of	IN
8961	167	7	AD	ad	NN
8961	167	8	697	697	CD
8961	167	9	687	687	CD
8961	167	10	,	,	,
8961	167	11	National	National	NNP
8961	167	12	Technical	Technical	NNP
8961	167	13	Information	Information	NNP
8961	167	14	Service	Service	NNP
8961	167	15	)	)	-RRB-
8961	167	16	.	.	.
8961	168	1	5	5	CD
8961	168	2	.	.	.
8961	169	1	H.	H.	NNP
8961	169	2	S.	S.	NNP
8961	169	3	Heaps	Heaps	NNP
8961	169	4	and	and	CC
8961	169	5	L.	L.	NNP
8961	169	6	H.	H.	NNP
8961	169	7	Thiel	Thiel	NNP
8961	169	8	,	,	,
8961	169	9	"	"	``
8961	169	10	Optimum	optimum	JJ
8961	169	11	Procedures	procedure	NNS
8961	169	12	for	for	IN
8961	169	13	Economic	Economic	NNP
8961	169	14	Information	Information	NNP
8961	169	15	Re-	Re-	NNP
8961	169	16	trieval	trieval	NN
8961	169	17	,	,	,
8961	169	18	"	"	``
8961	169	19	Information	Information	NNP
8961	169	20	Storage	Storage	NNP
8961	169	21	&	&	CC
8961	169	22	Retrieval6:131	Retrieval6:131	NNP
8961	169	23	-	-	HYPH
8961	169	24	53	53	CD
8961	169	25	(	(	-LRB-
8961	169	26	1970	1970	CD
8961	169	27	)	)	-RRB-
8961	169	28	.	.	.
8961	170	1	6	6	CD
8961	170	2	.	.	.
8961	171	1	L.	L.	NNP
8961	171	2	H.	H.	NNP
8961	171	3	Thiel	Thiel	NNP
8961	171	4	and	and	CC
8961	171	5	H.	H.	NNP
8961	171	6	S.	S.	NNP
8961	171	7	Heaps	Heaps	NNP
8961	171	8	,	,	,
8961	171	9	"	"	''
8961	171	10	Program	Program	NNP
8961	171	11	Design	Design	NNP
8961	171	12	for	for	IN
8961	171	13	Retrospective	Retrospective	NNP
8961	171	14	Searches	Searches	NNPS
8961	171	15	on	on	IN
8961	171	16	Large	Large	NNP
8961	171	17	Data	datum	NNS
8961	171	18	Bases	basis	NNS
8961	171	19	,	,	,
8961	171	20	"	"	``
8961	171	21	Information	Information	NNP
8961	171	22	Storage	Storage	NNP
8961	171	23	&	&	CC
8961	171	24	Retrieval8:1	Retrieval8:1	NNP
8961	171	25	-	-	HYPH
8961	171	26	20	20	CD
8961	171	27	(	(	-LRB-
8961	171	28	1972	1972	CD
8961	171	29	)	)	-RRB-
8961	171	30	.	.	.
8961	172	1	7	7	LS
8961	172	2	.	.	.
8961	173	1	D.	D.	NNP
8961	173	2	R.	R.	NNP
8961	173	3	King	King	NNP
8961	173	4	,	,	,
8961	173	5	"	"	''
8961	173	6	An	an	DT
8961	173	7	Inverted	invert	VBN
8961	173	8	File	File	NNP
8961	173	9	Structure	structure	NN
8961	173	10	for	for	IN
8961	173	11	an	an	DT
8961	173	12	Interactive	Interactive	NNP
8961	173	13	Document	Document	NNP
8961	173	14	Retrieval	Retrieval	NNP
8961	173	15	System	System	NNP
8961	173	16	"	"	''
8961	173	17	(	(	-LRB-
8961	173	18	Ph.D.	Ph.D.	NNP
8961	173	19	dissertation	dissertation	NN
8961	173	20	,	,	,
8961	173	21	Rutgers	Rutgers	NNP
8961	173	22	University	University	NNP
8961	173	23	,	,	,
8961	173	24	1971	1971	CD
8961	173	25	)	)	-RRB-
8961	173	26	.	.	.