COCTOBER 1, 1990

Niet ru govi
na ag oi
ie

k di
di
4...4
ab
Zaii ila.
7

ča ie s

m kj Ii r
K i ij 1 h di! upi
ji a s . - ix, i Ke 14 i. | m i ? ši SHE ži > | s) "NI i i dokaj k. ur mi, h k upal ' J
h na ' ž i ? odi LA r SI | as dl | l s tagih k: a, a Fe i ni va : a dke ba ; pi - - ud es ; 'u j - zi (ČE Hi :
i dije JK] sme a dka ae de. ci Prave. y " Ki) di ii ba a viša, sb pih eb k ki bak Ha a" ae ' lei, re di v: No, nM [ea de oe
h Z s ze pov na teh dli B; down k | adi pa de bra Sa b. Re eb pa F. še la ka a zi rih A 48. 1 h si ta x ' Pari prilo, š | 1 '
ee ke Me me eh 4 itev PN iv. MAT m ; | Ha be! 4 nej |
piv. gif: ie "ua | div b | nč |
kel a. si spit da » i is — -t 9 j h
ga | AA iji aa em ZA
es, 4 sed kr Ka em: yi: re, 4 adi ijari je Ji i 1 ST h z " s tai re F
L' ori zdi, k vami | A Eti radi ' ai J eh ta ii 4! ia: 4 ze ML, ve s sa i ha zav nas
; ž 41 Ja LJ ej : PA Mili. pade ej liki sin zd Pa t a ki ME dedi TI Me. saj:
bič a k moto a ds, ME i ; " aš

ii —a mi

ij h r : ali s
| ta bali: gt iN | | a s
TE avi € .p
N hi a a a ka iki Le .

Mu —

1

ey | [ 4. k be EP sa, pri

J

PA Puh t ž K kij Us rea si Pp E
; ad ni zet | ;
ka siY. a sit vra zd nini x ai € oča uk za! ha: a nim m (ke 4 plo t3: bo ii am

serie nji sika JA Poz bo EI OKJa Ma PESE Ta a | ie papo

Š a ea mesti a na renci ; et m edium., This appro«
Oil telo pa VE oris location in the document," in iza A A ee RE ka ičon elated pieces of text, opi oe a. ok: Ves
5; This,is the. dije Also, most full-text retrieval s oro ey'can be compared

jušna | | V Supports w zid card and Di Jart | raotunitjue capability—readers can browse
| ca series on text. zaueries to helo Make the search process <: and switched to a preferred term si

Po NA ja po Milana Pi Nm vame
ož. via o o nn nn ro m v | rani ma k sn mod 8 : :

POMETE mai
mi

ič — ega
i
TRENJ h mi ke Umek] aaa sj: —
4 mi daj a m pet 7 % ne zi izi raie ai! Ma z 1
s m z d m ž n z " č se za al m Ne
o Po a nn a PE a NO » ea SA proba ab di mi! NA ES a
- i kosa ud 4 la Ja poi d
a ! o A oi ! - 41 [ i :
: bid ! z ne
k A zni kosih it a 1 sš m: pek x. . m"
a... ga
z j : 5—-oše R
- P k L nail Z k ij
K! L di de - 4 h h | Ča k
h ut m a Mure - m. , ša,
ši gi a M.J ka B A ker a ud
ta: sei s K H
a % " - ! je " r ' % si P K s seg
- kiji .Č E irioa ., "-i : uč j ai 3
m. " " - Ha k ča
ži . ) z i
| i z s h
P ' Li PA
8 m B. zcni dai ad Z acai dol
za " ' bot mai ij i h - šk Piše ginij Moj ni " mi dd
- 1. ul 4 y s NA S ni Ia MI AM nu i K) 1
z i ; j i i
h z Kaiov zo J s s — a. m uč | me
| ' | du p- le: Mana fil
mi a t h JE den ? ke sl - aa
m - o ea i di j ; ih; ok. SI - Vač
ši di ij pa H li f Bi JMM PE ar m A
v . na kiji m mite « 1 bota 7 f "a
L. 5ME ti kai -e- Ha KE ;
" ne! 4 Kl Y em pe
4 1 %,: 4 - pi u, iu KE , .
| s NE. ur E be ge k. k K s pi iu h 4 ze - LE
k s - h uš ž
Li Ji J ta | h pro h l j
mr - ka Pe: ? a h
m mio > k, hi kje z k Ha ni dei i zi
; | ae, h ; š gi ii h ;
4 ee oa ne" ME z ; ai a v h
«a. k zd HI 4 s LM Di f peča i z Tae —] .% h k H mi
MI m : Ha j ? 1 "yi ri a 1..
pi x z ke. te ai sta h si i 1 - |
JOE z k ur m
are MS bi ič, a bče
a s ; paje i i
ia ko. h fi. gi EL na gi B "
ih 2 s
nim ; m
k
ki
Pi

o a mni sš

FI či ki ikoni,

through the text collection by follow-
V management. ' more effective. Other commonly used sconsistency b zj ing links of interest. Their guest for in-

upi | bu žime, si ČNE

Wu č Me] and ts inf  technigues include phonetic nea Topic,-a text-retrieval product from. a lt jupoo vama: be as niče ic or as gen
1 avli ile, vnos, the zgiel where, ior example; a search on. :/:..:.:Verity. od is otin ama de; ojrov š | za he atlons reguire that...

| ste ods vertsa gre:

| bark ekoa x S tai o UM TOPNE RA pori USE mete daš bi Ma jine ea iz ze someone.with knowledge. of the do-

< maln identify the location.of: the. ina
In addition, the links must; be manual: 5
J Jy. built, into the application, They.are' pe:
Avanalogous.to 'GOTO. ) stafements in uh-' |
SR... /structured programming code, with all.
paj a »the attendant, maintenance problems.
df. —ičuenikoo.. The assigning of key words, the' - '
BA VE VEŽI ejuding friend: " h ' building of topics and any other labor- .
ee user A and: the ability to': MM i:. —.. a "lesjons oka a AI RE MIM ; intensive front-end work regulres a
| generate answers that do not reguire Ii dv! ba ANO (jas. tini Fin on Eh SEE EIN ' major commitment from an organiza:
| additional verification. They-will as-' "4 ge eki Ai | Pare AM io ea az UKRO | MM ie Paine ENE AH sa mEči
".| sume that the information returned »/ty:" exing, can A | cal-
| will fit well within the para metere of. ii | "pesPe—7., ly and may be beneficial when a large
HB their search reguest. ::.2:/.:, ETA KAKO ct] resi kin raš: ž ; o Te tslici text collection is involved. However,

" . Today, however, most as; applica: čni an jleS a er ome | this approach reguires that the user
| tions do not live up to these expecta-. di? . — rele! RMAN understand the conceptual categories
| tions. Many employ an indexing tech-'' HO Uf ČI peri?.. | mi" Seas | Sym m and relationships that sreči in ke em
nology that peča] pyepedlsti This ap ap- : Na — za slv ii Elba |] RR ec, SANJE ti RAV tjaka se: ah Doze aciliden!
kliše | redto! asa dictionary file) tine ; ei TU Mlak OL ENA nim As | i should understand natural language, .,
sli the words that comprise a text collec» s: [Of yo Oms] | HE - However, to date, interpreting the a
se sie -| ton, omitting 'noise?words such asis) akti SIH [INA meaning of text reguires human inter. |

pije mete aand and, him let ostati ite | | pe-- | nie - RI O m stvention to supply intuitive Gi domain. |
7] To search this data. :pool,:'users must: o ' knowledge.

| apply Boolean logic:to 'phrase. gueries. lisi Next week I will i— te isr CD

However, because the: gueries are not, ;. : ROM plays in text management. H.
'specific,'a: single search: phrase might: ije |

'| retrieve unrelated or pveneco na ar bje

NE Boolean seje | — 'Technigues that Reduce
an v gega grd] Dkjsaa 3 "v Mismatched Responses

mir nini A S na o omeji — a ai A nani

JU: management: se ; 30E s,
| : applications" sl
ž: - meet. certain re: fi

MARTIN [ieaornke

BRA

"Nak
8m s

d Aa
is:

s Ha
: — pa a ma h 1 h." da
- z — . - š č s j z ] z 4 ui 1 b i
z zom z m II —-- o ej VA Medi O oi Na kh Pi nea on a a oi A EM VE 8 im ino bi mna, Iz —- čr Mai L TI NŠNI TRTI eno mami. —-
| h a j i k .
-a Gi HI L 8 s iz mh | B s ] 1 i . si pi ; s d
X ui x 4 z FI ' j . u laj a k ' m k P br:

a. Pr k" - s z — s ki m.

d 1 ta E' sla di NE j ; ni t : HI x

i : kj ih zd j ič k i 8

po Če
"

The above
examples might be SPATALJJ Hi TA eo —— |
found in a medical reference guide. nasi Ni (UMIM The concepts in this article are de- .. |
iticles. iin j.: sej pepela JIM jo REČ EPA Se SV aa Di PAL AA ARA k Šum a midi scribed tn a new volume, Teri Man |
1) To minimize: zi Er nalchesi sei ni PA sodi? ogeksleko si ua dg ia pobija sei ča odlič; nada sili. DUŠI z BILE :Jom Avakčan agement, of The James Martin ke-
'|<mosttext-retrieval software.offers' aj it, fujeb pfusaLineji kate BEE SR OTU NU Sai e nei "ova Gg Guid si port Series. For more information, .../
' |rtechnigue known:as proximity search: »"Meyer? End the ability to browse: the ix: sein o atrolica ji erabri Uččej ETJOJI ' on this volume, call (617) 639-1958.

ing, which lets the user specify the"; ' index.itself, letting the user gain famil-:;;:' terms into symbolic. concepts and dne » For information on seminars, con-':;:

number. of words, sentences. or para-. - V darity with the nords that Por pise. the. čethese concepts to each other via a : tact. (in the United States and Cana- .

'|-graphs that can separale; individual, zslext. Collectionjši pi ske aii — gesraphical representation that can be: di da) Technology Transfer Institute, :.:
search terms.i' xi). gi ge | ' To minimize USer rfn istration with - browsed and assigned a label. These rep-: 441 10th St., Santa Monica, Calif.....

us |vTo support, proximity ATAI a: [ 'auery results; many. vendors. offer. tet si zl resentations, called '"topics,' are used tos . 90402 (21 3) 3994-8305. In Europe, '

'—..| postings file is created in addition to" ' 'saurus capabilities to improve recall and" search criteria and can be combined us-:; eponlori Savant, 2 New St., Movia
—| the dictionary file.:This file maintains; rati v precision. in the retrieval process, jz gi ing Boolean operations. : nes., LA5 9BX. Ma d Kingdom.

detailed. information ! ertainir to začni | Geč PA. thesauruš can val ii y premo. si ile is get another text: manage... sl (0524) 734 205:

aa m li vj oDBMS: Tecinolog j Is Rolling Along.
lise df na pijači Ugobami Wač klifi

dis ki v, sa ti a | Ko SAT, ge pi sjan sa ksi, tut R |
nIŽ a kto steči telic sj PNaP EGA? Masji . Ki N- AUKN Pika
pi sit Be na mij ae zi: pi 4 m švit s: oa nei zi ja bie: BI 4 še, Biroren ni preds i' SA ti; tlači '" k (min jeni b Mi 5 JAH Poadateket. TEL pa Es

| Continued. from, Page 68. ša zi die "Once. yow ve: designed" vyining: to. 7 | ti inieč, ifrom Page, 63. jetra a a n: hi .. were:done i in a relational kite čali

—ilj ' capture data, you can! let'the.end user: — -- Todd Bashuk, technologist for Southwest:

—and menu: 'ehoices: to: ET Oifferent iztretrjeve data and: do. the analysis. them-č]|t NU 'Ontos database nee "with: a; new: -ern Bell Technology Resources Inc., a divi-
| fourth generation language (4GL) tool... ; sion of the telephone company in St. Louis,

omen of: the data;-regardless' of where! Pis selves,:without; needing help from the'sys«:|
in'what format it resides on the net-:::tems department: every time their needs | :: ODBMS technology, which is urrenti vh ;;Ontol also in Burlington,  Mass., is.
EI ; work; he said. € ag 4 mia ii" 4914 change, U according to' Bob :Holmes,: com: ENIHSE from only a handful of players,"'  tossing paišekk entry into the ODBMS
JE 5. -% Coca-Cola plans to tap DDANencE to '' puter:technology: research analyst for the | .stores complex data types directly, unlike . ring. Its Ontos 2.0 upgrade runs on Sun-3,
%.. build.a nationwide retail. survey system! ' Southern California, Gas. Co. jin Los; An: :|-more traditional relational databases that Sun-4, SPARCstations, Apollo and
ME at snana multipla. LANG in order to help: geles,:: GORAN ENE Kao lek SA a) ij "organize data in tables of rows and col-:  DECstation uorolatona as well as, PCs

H users: perform break-dov "analyses of: ..-'This kind. of: fexibility will boost; the '|umns. This approach, which is well-suited .. running OS/2."- J
:for. creating large, multimedia databases '..:Ontos'is also showing off Ontos Short: m:

-of bit-mapped graphics, geometric forms hand a new 4GL option for its database.

" market data.i':/". MjAEB NB ae Wi Efnciency: of users, alleviating the need to :
zand:compound objects, can:also serve to: -'The 4GL features an interface called Ontos

VES (%2/With this. system;'a user at.one retail;' consult a database programmer whenever-
zj |e sš V chain; for. example,:may. choose to view: new:development: needs arise, Byers. sndčji
|| :boost, performance, according to officials |. ; Studio, which provides access to blocks of
of both firma. |: preprogrammed code for buttons, menus

" sales of'a particular.Coca-Cola product in '; 'other InfoAlliance users said, ,

GT$: "one orstwo regions; InfoAlliance will auto-'. |. InfoAlliance, which is the resulto£SPCs |. oo
č ti H Object Designs ObjectStore 1 O comprises "and text boxes. Both options will Ha avail-
'three major components: a dat: ase mod-.. "able in December." prav

si / /matically.cull the data from. odg dis- - "reported: $10 million, five-year 'develop-"
ju, the Objecištore Data Manipilation Wi Pricing foč OLJečStore. vrhici IE deteč,
age and a mix of development; tools, -: mined by the number of users per site,

nili nee - similar databases. AE NE eva /" ment effort, will be available for both DOS.

IH ubi " SLater, that same user could review sales "and: "Windows. by the; second gvarter. of"
include SchemaDesigner, a: graphical da-. anges from $2,000 to $9,000. ses ma
labase-des ) tool and code generator, said ; 'Ontologic's Ontos upgrade is free to reg-

7 of that item in other regions without hav'.". next; year. ic 47 ruge eea%
"Ang to irespecify: to. the InfoAlliance soft //- V  Pricing opi InfoAlliance 'ranges' from/
|" ware where the data is located, said Byers, -1 $8,500.for a single-server, 10-worksta
iš > 24" People: developing .lviews] won't have" .configuration to $99,500 for a nine-server, y officials in Burlington, Mass. MN ated Use otithe Ata keše Sla licenses.
s AE leže uje VI need an object.orjen ted ''are availabl starting at, $1,500 per user.
Sil kožča Ee sd part, of the sadi ia. SPC, of si View,.C 4 mi e. |, database is that Im modeling pretty com: lez Object can.be ge ned at (GI7),
SLI ; | pa Red Kupi Soiga PRE oce at (li gosooee,a O op iglečajigjan 'plex sbece: and it a aige 4 it z KeTOSTOT,  Onilogiiset (611) 272" z1l0. E

eo o Beja ELE oj Mor pa as A tujki

"im

s
meči f

pi

d.m 4
LA HI 4 4

toda - KE,

ua bož
i ma 5, ji
ce -

ši ti udi

s Zi
i, S li Pdd.E 1x3

hi
i
ke
f

s. v, 4 s: te
Ki i rta, s

m mu |
LE.

|

ae tace bad Zamik rebus
po

sa ža

|| ze to think o£ everything while. they; are " iz 200-workstation version. | gtikaš

4% i

i i .
zu HM mad. UH Ai A A
ah ak 1: A ; ue. sta LI H ui 1 da Ro H j H LI sj [ k » m |
Re 7 dN A H ? na" Li k m! bt" l ee mol za K a; " h a ( note tl
da m, ka ia ae 4. P fei oz ki ' . ki t -
sd 4 4 gi na ae im 'a —- 1

o
me
si
] a
a RE BA) PRA A EM SA, ME a si
bina ma "a Si mi — iii ! biu s ši sy ke $IU Er va nd voda na Hm HE N, he sie mia. y g B $ že a" ii "a d A.
kd ; - de ej ra a.» % fee: vsa pa x soji NI piši ii ge $ a a i kg siti, MJ Mara MIZ da Za.» H Že jata na - i tat. me pa ij i 4 de: J si
J LI k k ] a nA že P di... in. A ME i ji 1 i ie s sei na ; - a a im. 1 id 1 z ; : sa i
ne miš " Ami, ublgv pAč nd ja Ula č a de — ki. 4 i ke ga A — ue V Tap pi ze nA? LAM Zda, E— EEA " MA od ML, LI, Y
h Ž ip a ii A, 1 | 4 l | - H -a NE U za s ; lai, La
ia 9 BE. pooiaj yi AI Ed me je k S IVE] ad Ed ral Me oa ei Su ATE Ur EEA x te Me s Mi je, ma |
| 5 Ee ej k uri ii, Le be v ZOE i .Ww« : . v i ( ps] pa ie Ke, Tess « ubjs: sar, A. m. Če, pd silk k ua z a L: tar i m ja. "a F z" sa za Se" | i.-a i f ; zi Va a - " ga pi H k na ki... t. zad : h SA k j J h i kve -.
iko, ee bej m Be PRE k f nm raru die, | . et AA A — | Ma,:
dai Ne A, Pa Na e" LEE Zi 4, TAT ai" s bie SEME pe na ko, JE MENE, h P h ic,
H ie Le sjaki sa! pouka Ča ei ma dia me ba m si ia sti ko La ae NE; "m sh ee s —iu - iii 4 : - - " AR h s?
sma! i | 4, na a s; sa — IME ga Ri ie soja : —e- poi A gali Se ie kl h paj ee Eme! h ua
es zi age Pod Pa ie ni: oi ka ui... učija po zjabi odli Je Z MALI gl A RE SE rk uta RE biu črča Ne PP a
ie ireksi Np k ez z Mo no pa Ee AE UE pe ogleni Mae Seas tai, SU ere
ae, E ua a šč A [ Gen k POLI dr dai Pa dil rel ši ih a Ena s sila 5] s: či