Analysis Paralysis: when to stop?

An a ly s i s P a r al y s i s : Wh e n to sto p ? Er .A ks h ay B h a rdwa j , U ni v er s i ty I ns t i tu te of I n fo r ma t i o n T e c h n olo g y , H i ma ch al Pr a d es h Un i v e rs ity ,S hi m la - 5 ( In d i a ) A b stra ct A n al ysis of a s ys t e m cons t i tu t e s t he mo s t i mp or tan t asp e ct of th e sys t e ms d e v el op me n t l i f e c ycl e .B u t i t i s a l s o t he most confu s i n g an d ti m e cons u ming of all th e s tag e s .T he cr i t i cal q uesti o n al w ays re ma i n s: How mu ch an d ti l l w h e n to an a l ys e ? E d Y ourd on h as cal l e d th i s p h e n om e no n as Anal y s i s P aral ys i s. I n th i s p ap e r, I su gge st a mod el w h i c h can actuall y h e l p in arr i v i n g at a s at i s factor y an s w e r to th i s p rob l e m. Tab l e of C on t e n t s 1) Int roduct ion 2) Princ ip le s 3) Imp le m ent a t ion 4) Dia g r am m at ic Rep resent at i on of t he AA P mode l. 5) Concl us ion 6) Refe renc es 1 I n trod ucti on T here are s eve n m aj or st age s in t he S y st em s Deve lop me nt Li fe Cy cl e.T he y are 1)Re cog nition of ne ed 2)Re qui r em ents Anal y s is 3) S ys t em s A nal y sis 4)S ys t em s D esig n 5)S y st em T es t i ng 6)S y s t em Imp le me nt at ion 7)Pos t Im p le m ent a t ion.Out of al l t hese st e p s t houg h al l a r e e qua ll y im p ort ant t he m os t im p or t a n t am ong t hese i s S y st e ms Anal y s is. T his s t ag e by defi nit ion st a t es: S y s t em s Anal y s is is t he di ss ec t ion of a s y s t em int o it s com p onent p ie ce s t o st ud y how i t s com p onen t p i ec es int e rac t and work.[1]. Broadl y sys t em s a nal y sis p ha se e nc omp ass es t he fol lowin g m ain ac t ivi t ie s or s ub p hases: 1)S urvey 2)s t ud y a nd 3)defi ne.[2].T hese sub p hases ac t ual l y enc om p ass t he fol l owing ac t ivi t ie s .In 1)A n effort is undert a ken t o s t ud y t he e x ist ing s y st em where the cha ng e t o be brou g ht i s desire d.In 2) t he p hase i s s t udie d and t he t e chni ques t o bring about t he de sire d ch an g e ar e e x a m ine d.In 3) t he ne w s y s t em is defi ne d kee p i ng in v ie w t he i nform at i on t hat has bee n a c cum ul at ed from t he p revi ous p ha s es. How eve r i n al l t he a bove m ent i oned st ag es t hree t hin g s t hat a re c ruci a l a nd ne ed t o be de fi ned p rop e rly are : a )D e s c rip t ion of i np u t s t o t he s y s t em b)descri p t i on of t he p roc ess es t hat are be ing used and t hos e t ha t w il l be app li ed t o t he s y s t em a nd c) de s cri p t ion of t he out p uts t ha t are desire d from t he sy s t em aft e r p roce ss ing .T hes e p roce s s es t houg h very cl ea r and se em in gly succ inc t t o defi ne a ll hav e one c h ara c t eri st ic , t hey al l a re s o m u ch a bsorbin g t hat s ome t im es it bec om es di ff ic ul t for t he a na l y s t as wel l a s t he p roje ct t ea m t hat how far should t he a nal y sis cont i nue.In ot he r w ords i t e nt ai ls aski ng t he que st ion: A re we rea dy t o de s i g n?.I t ry a nd p resent a m ode l whic h w il l t ry t o solve or at l ea s t g ive t he a nal y st s s ome y a rds t i ck b y w hic h a solut ion t o t his p roble m ca n be ac hi eve d.I c al l t his mode l t he A A P ( A ks h ay’s ana l ys i s p aral y s is) mode l. T he roa dma p of t he p a p er i s as foll ows . I mot iva t e t he nee d for ana l y s is in sy s t em s anal y sis and de s ig n. I foll ow by p resent ing an a l g ori t hm whic h coul d e ff ec t ive l y count er t he “ a nal y sis p ara ly s is” p roble m .I t hen p resen t t he A A P m odel . U lt im at el y , I conc l ud e. 2 Pr i n c i p l es T he m aj or p ri nci p le underl y ing a nal y s is in sy s t em s ana ly s is and de sig n is t hat t he a na ly s is p hase serve s as a st ar t e r or a bl uep rint for ca rry ing out a ll t he m aj or ac t ivi t ie s t hat ac t ual ly le ad t t he de s ig n and t hen t he i mp le m ent at ion of a s y s t em t hat is bei ng deve lop e d. T hus t his p ha s e i s t he m os t cruc ia l in terms of t he e ffec t t hat it has over t he e nt ire s y s t em . If t h e p roble m i s not underst ood or ana lysed p ro p erl y t hen i t would l e ad t o a p oor desig n and l at er on m i g ht le a d t o an entire ly w rong im p l em entat ion. How eve r a n over or an unde rest im ation of t he p roble m a t hand would re s ul t in a ca t ast ro p he.S o, t he m aj or p rinc ip le unde rl y in g a na ly s is effe ct ive ly me a ns t he c orre c t dia g nos i s a nd a p la n for s ol vin g a p art ic ula r p ro bl em . 3 I mp l e me n tat i on T he que s t ion t ha t most of t he s y s t em a dmi nist rat ors/a nal y sts are fac ed wi t h is w hen t o cross over from t he a nal y s is t o t he de sig n p ha s e? W hi ch e ffe ct ive l y me a ns when t o s t o p p la nning a nd s t art desig n in g ? A t yp i ca l s y s t em s ana ly s t is resp ons ibl e for, a nd ac t s wit h t he fol lowing set of ent it ie s . a) P E O PLE , w hi ch i nc lude s m an a g ers, users and ot her de vel op e rs . b) D A TA i n cl udi n g ca p t ure , val ida t ion, or g a ni z at ion, st orag e and usa g e. c) P R O C E S S E S bot h au t oma t ed a nd m anua l. d) I N TER F A C E S bot h t o o t her s y s t em s and a pp li cations, as w el l as t o t he a c t ual users e) G E O GRA PH Y t o effe ct i ve ly dist ri but e t he da t a a nd p roce s s es t o t he p eo p l e. [3].K e ep ing t he a bove m ent ione d det ai l s i n mi nd we re al i z e t hat a sy s t em s anal y s t has t o be i nvolve d a t al l p hases of t he S D LC. But it is in t he fi rs t t hree p hases t hat his/he r role i s of t he ut mos t sig nifi c anc e . In my p ro p o s ed m odel t he fol lowin g are t he t erm s /p ara me t ers t hat ne ed t o be defi ned: 3.1. Pe op l e I nd e x (PI ) : T his ac t ual l y is a m e asure that w il l he lp t he a nal y s t t o as s ess t he a m ount of cont ribut ion t hat t he p eo p le invol ved i n t he a na ly s is s t ag e ha v e done . . T he re a r e basic al l y t w o t yp es of p eo p l e who are i nvolve d i n t he c oll e ct ion of i nforma t ion: a) I n formati on Gathe re rs : T heses are the set of p eo p l e who are resp ons ibl e for c oll e ct ing t he i nforma t ion t hat i s neede d by t he m duri ng t he i nform at ion ga t heri n g p hase. T hey a ct ua ll y c ons is t of al l t he ana ly s t s , p rog ram me rs, int ervi e wers et c. who col le c t ive ly const it u t e t he i nform ation g at he ri n g t ea m. b ) I nformati on S ourc es : T hese a re t he set of p eo p l e who ac t ual l y p rovi de t he i nformation t hat is nee ded by t he i nforma t ion ga t heri n g t e am . T hese m ay be or m ay not be a p a rt of t he s y s t em t hat is under i nvest ig at ion. T houg h m os t ly t hey consis t of t he e nd us e rs i .e. the set of p eop le for w hom t he s ys t em i s int ende d t his mi g ht not hol d t rue i n al l t he c ases. P I h e l p s u s i n as s ess i n g th e amou n t of contri b u ti on t h at h as com e from b oth th e s e p ar ti ci p ati n g e n t i t i es . T o hel p s e t t he va lue s of PI for a p art ic ula r p roje c t t he fol lowing quest ions ca n p rovide a useful est im at e: How ma ny typ es of fac t findi ng t ec hniqu es [4],[5] w ere foll ow e d ? How ma ny t im es was t he s y s t em unde r st ud y visit ed? Wha t w a s t he c om p os i t i on of t he fa ct fi nding t eam ? Wha t w a s t he e x p eri enc e le ve l of t he t ea m ? T o w hat ex t ent w as t he l iterature of t he e x ist ing s y s t e m re vie wed? Did t he va ri ous que s t ionna i res and fe edb ac k forms hav e some dist in g uishin g e nt rie s or was it s t a ndard run of t he m i ll s t uff? How muc h use of aut oma t ed t ool s a nd t ec hnolog i es ma de v is a vi s ma nu al st rat eg i es? How muc h of t he i nform at i on gat he red was ma int ai n ed i n p rop er doc ume nt ed form ? T here are al so a nu mbe r of m at hem at ic a l form ul ae l ike t he m e t hod of L e ast S quares [6] t ha t coul d p rove effe ct ive ,a mon g ot hers. A s t he va lue of P I ha s be en ke p t be twee n 0 and 1 i t me a ns t hat if t he a s s ess m ent a s p er t he p rec e ding qu est ions fal ls le ss t han ha lf i . e. 0.5 t hen t he t ea m ha s not done it s w ork p ro p erl y and ha s no t be en a ble t o gat her t he re qu isite i nforma t ion. W hi ch a ct ual ly im p li es t hat t he c omp os it i on of t he i nform a t ion ga t heri n g t ea m ne e ds re t hinki n g . H e nce t his is a ve ry cruc i a l fa ct or in de ci din g whet her t he a na ly s i s can be ca rri ed out p ro p erl y or no t . 3.2 Data gath e r e d (D G ): T his ac t ual ly hel p s t he a na ly s t in de ci di ng how mu ch da t a ha s ac t ual ly com e i nt o hi s / her p os s ess ion. T he da t a c an be furt her subdivi ded i nt o t he fol lowin g t w o ca t eg or ie s : a) Dat a that is im m edi at el y us eful (U ): T his is t he am ount of dat a t ha t is g oi n g t o c ome i n h andy im me di at el y . T his ess ent ia ll y me a ns t hat wh at eve r dat a ha s c ome t o t he Anal y s t he/ she c an, ba s ed up on t hat da t a t ake im me d i a t e dec isions as t o t he desig n of t he p rop o s ed ne w s y s t em . Cons ide r e . g . t hat a sys t e m a nal y s t g ets t he i nforma t ion t hat a p art ic ula r p roce ss is no t w o rking due t o a ce rt ai n fa u lt w hic h, w it h a l it t l e a mount of modi fic at ion c a n be ea s il y handl ed T hen t he a nal y s t ca n cha n g e or mod ify t he e rring p roce s s w i t hout muc h of a n effort . b) D a t a for fut ure use (F): T his im p li es t he da t a t hat i s no t im me dia t el y rel eva nt or us eful but ca n c o me i n handy at a l a t er st ag e. Conside r e. g . t hat an a nal y s t w ho is inve s t i g at ing t he Fina nc e Dep art me nt of a c orp orat e house rec ei v es inform a t ion t hat 4 H uma n Resource p ersonnel ha ve le ft t he c om p any in t he p as t one month due t o l ow sal ary . T houg h t his inform ation seem s t o be not us eful i mm edi at el y but m i g ht com e i n us e fu l l a t er. M an y t ec hnique s have be en de ve lop ed for Dat a Gat heri ng w hi ch m a y be de duc ed m a t hem at i ca l ly .[7],[8]. As t he va lue of U and F has bee n kep t be twee n 0 and 1 i t me a ns t hat if t he a s s ess me nt as p er t he p rec e di ng di s c us sion fa ll s le s s t han ha lf i . e. 0.5 t hen t he ac t ions as p er t he A A P mode l a l g ori t hm ne e d t o be t ake n. 3.3 P roc e s s I n d e x (PRI ) : T hi s wil l g iv e t he a nal y s t an i dea of how muc h he /she know s of t he p roce s s es in t he e x ist in g s y s t em . A lso based up on t his inde x a s t rat e g y of ide nt ify ing t he c ore p roble m p roce ss es and t hei r re ct i fic at ion c an be formu l ated [9] ,[10]. T he p roce ss es ca n be di vided int o t he fol lowing s e gm ent s: 3.3.1 Core Proc ess e s : T hese a re t he p roce s s es t ha t are absolut el y ess ent i al t o t he working of t he curre nt sys t em . If t hese p roce s ses are rem oved or de ve lop a fa ult t hen t he ent i re s y s t em would c ol la p s e. 3.3.2 S up p or t ing P roce ss es : T hese are t hos e p roce ss es, w hic h t houg h p la y im p or t ant role s in t he functioni ng of t he sy st em , but t he y only p l a y a su pp ort ing role t o t he core p roce s s es. T hu s t he y i ndire ct ly affe ct t he funct i oning of t he sy s t em . T he da t a g at here d c an a c t ual ly be c omp are d and ana ly s ed by u s ing c ert ai n em p iri ca l a p p roac hes w hi ch i n ass oci at ion wit h p roce s s ide nt ifi c at i on coul d p roduce insi g ht ful re sult s [11] K ee p ing t he va lue of P RI bet w ee n 0 and 1 me ans t hat if t his inde x fa ll s t o a va lue le ss t han 0.5 t he n t he a nal y s i s t ea m h as not underst ood t he func t ioni ng of t he p roce s s es w hi ch ac t ual ly me ans t hat t he a nal y s i s ha s t o be done aga in sinc e if p roce s ses are not underst ood t hen t he ne w s y s t em would not p rove t o be be t t er t han t he e x ist i ng one . On t he ot her ha nd if t he va l u e of P RI c ome s out t o be g re at er t ha n 0.5 and m oves t ow ards t he va lue of 1 it mea ns t ha t t he p roce s ses have bee n w el l underst ood and ba sed up on t hese ha ve be e n i dent i fie d as cor e or supp ort ing and a s s uch c a n be p ut t o us e i n t he ne w s ys t em or be t o t a ll y done away wi t h. 3.4 I n te r face U ti l i ty (I U ) : T houg h not s t ric t ly a p art of t he a nal y s i s p hase t his p ara meter al s o has t o be t ake n i nt o conside rat i on bec ause f ina l ly t he e ffe ct i vene s s of t he s y s t em w il l be g au ge d by t he users and a p oor i nt erfa ce cou ld de s t roy a wel l desig n ed sy s t em . S p ea kin g in term s of c onvent i ona l sy s t em s deve l op me nt t he de vel op me nt of us e r int e rfac e fa l ls int o t he desig n p hase. How eve r in t he a na ly s is p hase t he ana ly s t ca n l ook at t he ki nds of outp u t s t hat a user i nt erfa c e i s e x p ec t ed t o p rovide a nd kee p ing t his in m ind c a n fr am e a few set of quest ions whic h coul d be used t o p rovide an i np ut t o t he de sig n p hase. S om e que s t i ons c ould be l ike : 3.4.1 Wha t are t he out p u t s t hat are t o be e x p ec t ed of t he Us er In t erfa c e? 3.4.2 Wha t are t he p roc ess e s t ha t coul d be refl e ct e d in t he U ser Int erfa c e? 3.4.3 Who a re t he t yp es of us ers t hat t he i nt erfa ce w ould c at er t o? 3.4.4 Woul d a de skt op based i nt erfa ce s uffic e or would it be a w eb ba sed i nt erfa ce ? T he l i s t of quest ions, t houg h not ex haust ive in na t ure he lp t o ide nt if y in g ene r al t he c har ac t eri st ic s t hat t he i nt erfa c e would p os sess and t ha t w ould serve a s a n inp ut t o t he de s i g n p hase. A l so s ome hel p ma y be de rive d by foll owin g al rea dy est a bli s hed t ec hnique s s uch a s an It era t ive Int erfa ce Desi g n[12] .If t he val u e of UI t ent a t ive ly fal ls s hort i.e . l es t han 0.5 t hen t he a nal y s is t ea m nee ds t o rethink a nd have a m ee t in g w it h t he i nt ende d desig n ers of t he ne w sys t em . 3.5 G e ograp h i cal Quotie n t (G Q ) : Ag ai n t houg h t he i s s ue of di st ribut ion c rop s u p in t he i mp le me nt at ion p ha s e but eve n in a nal y s is if t hi s fa ct or i s t ake n i nt o ac count i t has a sig nifi ca nt bea ri n g on t he out com e of t he p roje c t . Con si d e r a sc e nari o : L et us ass ume t hat a t ea m of a nal y s t s ha ve succ ess full y im p le m e nt ed a n E- g overn ance p roje ct in T am i l Nadu(T N ). N o w t he s a me t ea m i s ca ll e d up on t o im p le me nt a sim i la r p roje ct in Him a cha l Prad esh(H P ). T he t ea m based up on it s p as t ex p eri enc e t rie s a s i m il ar st rat eg y but s ome how here t hing s do not s ee m t o be g o in g t hei r w ay . T hey find t hat t he g eo g rap hic a l di ss im i la ri t ie s bet w ee n t he t w o p la ce s have give n rise t o a numb er of fa ct ors: 3.5.1 T he p refe rred m ode of UI int e rac t ion was Engl ish in T N but has t o be Hindi i n HP 3.5.2 T he p eo p le are resist ant t o ide a s i n H P ,w hic h were ea s il y ac cep t ed i n T N 3.5.3 T here is a l ot of red t a p ism i n HP w hic h w as mi s s ing in T N 3.5.4 T he hi l ly nat ure of t he t erra in i n H P is ma ki ng inform at ion g at heri n g d iffi c u lt vis a vi s T N . T here coul d be m any m ore hi ndranc es t hat t he t ea m c oul d hav e i dentifi ed. T hus t he g eo g rap hic al d iss i mi l ari t ie s i nt roduce a wide var ie t y of ot h er i s s ue whic h coul d range from soci al t o cul t ural t o ec onom ic a l t o any t hing el s e. T he re are a wide v ari et y of t echni ques t hat ha ve be en m a t hem at i c al ly t rie d and used for g at he rin g s p at ia l da t a a nd for ac cura t el y ide nt ify ing t hem .[13],[14],[15].Us in g t hese t ec hnique s some s ort of dat a c a n be a cqui red whi ch c ould g o some w a y in solvi ng g e o g r ap hic al p roble m s for t he a na ly s t .H oweve r it i s c le ar t hat if GQ fall s t o l ess t han 0.5 i t me ans t hat t he solut ion we a re st rivi ng for ca nnot be m u ch g ene r ic i n n at ure a nd ca nnot be re us ed. A l l th e ab ov e me nti o n e d te rms ar e ass u me d to hav e a r an ge of v al u e s b e tw e e n 0 a n d 1 w he re 0 d e n ote s th e b as e v al u e an d 1 d e not e s th e p e ak v al u e . S o havi n g d efi ned t he t erm s t hat t he m ode l would use l et s now de fine t he st e ps t ha t t he m o de l wi ll have whi ch c a n b e defi ned usin g t he fol l owin g al g ori t hm( A A P mod e l ): 1) Che ck a l l t he da t a t ha t has bee n g a t here d so far. B ased up on t he t w o c riteri a i .e. i mm e d ia t e useful ness a nd fut ure use di vide it int o t w o rel e vant p or t ions. 2) For ea ch of t he tw o differe nt p il es of dat a s o c rea t ed m ake a l i st ing of t he cont ri but ion of t he p eo p l e i nvolve d (PI). 3) a) If PI > 0.5 but ei t her U<0.5 or F <0.5 t hen t he da t a g at here d ha s no re l eva n ce . D isca rd and be g i n ana ly s i s a fresh. b) If PI< 0.5 but U > 0.5 t hen da t a i s ha n d y . Chec k for fut ure useful ness . c) If PI< 0.5 but U >0.5 and F > 0.5 t hen y ou have dat a t hat is bo t h us eful and ha ndy for fu t ure. How eve r t he t ea m nee ds reworkin g . d) If P I>0.5 but U > 0.5 and F>0.5 t hen y ou have da t a t ha t is bot h u s eful and ha ndy for fut ure. M ove t o t he ne x t s t ag e. 4) Out of t he da t a g at here d so far a g ai n divi d e t he d at a i nt o t w o p il es. A gai n che c k for t he im me di at e useful ness and fut ure us e. H ow e ver t his t im e t hese p ara m et ers w ould be d efi ne d rel at ive t o only one fa ct or i.e . P roce ss Inde x (PRI). 5) If PRI< 0.5 t he n ana ly s is nee ds t o be done a fresh. Bac k t o s t ep 1. 6) If PRI> 0.5 but <1 t hen a nal y s is nee ds t o be cont i nued but onl y a g rea t er unde rs t andi n g of t he p roce ss es nee ds t o be ha d and shouldn’t s t a rt from scra t ch. 7) If PRI= 1 t he n w e ca n move t o t he ne x t st e p . 8) Che ck for IU. If IU <0.5 t hen re t hink. Che ck a lso if PI<0.5.If bot h condi t ions hold t hen i nvolve m ore p eop le . E l s e move t o t he nex t s t ep . 9) Che ck i f G Q <0.5.A lso che c k if e i t her PI< 0.5 or P RI<0.5.I f y es t hen m ay be G Q is affe ct ing bot h PI and PRI t o s ome e x t ent . T r y and find a lt erna t ive s . 10) If G Q> 0.5 t hen we a re rea dy t o move int o t he de sig n p hase. As is cl e ar from t he a bove al g ori t hm we c an a c t ual ly t r y and a s sess our p rog ress unt il we ha ve re ac hed a desi red le ve l. 4 D i agrammati c R e p r e s e n tat i on of th e A A P M o d e l start Input all collected data, PI ,PRI, U,F, DG , IU , GQ Divide data into two piles based upon U and F For both piles compute PI Is PI>0. 5 and U<0. 5 and F<0.5 NO yes Discard Data Is PI<0. 5 but U>0.5 A B C A Is PI<0. 5 and F>0. 5 and U>0.5 Is PI>0. 5 and U>0. 5 and F>0.5 Based upon U and F relative to PRI divide into two piles. Is PRI<0.5 B Is 0.5< PRI <1 C Is PRI=1 Is IU<0. 5 and PI <0.5 Is GQ<0. 5 and Pi <0. 5 and PRI <0.5 Is GQ>0.5 Move into Design phase Stop Rethink.Involve more People Find Alternatives F i gu re 1 5 C oncl u si o n All t he a bove m ent ione d t erm s ar e a s s ume d t o have a r an g e of va lue s bet w ee n 0 and 1 where 0 d enot es t he base va lue and 1 de not es t he p ea k va l ue. T he quest ion t hat cro p s u p in a l l t he above ca ses is how t o dec ide t he va l u es for ea ch of t he p ara me t ers? T his is p art i al ly based up on t he di s cre t ion of t he s y s t em s ana ly s t or w hoeve r i s i n cha r g e of t he a na ly s is t eam . It is hi m/ her /t hey w ho w il l de ci de what ran g e of va lue s t hey g ive for e ac h of t he fa ct ors bec ause t he cr i t eri a for t hese would cha n g e from p roje ct t o p roje ct . How eve r t he m arg i n for di cre t ion would be very s nma l l as nume rous ma t hem a t ic al and st at is t ic a l t ools ex ist t o a id the ana ly s t as p oint ed ou t b y t he p rec e di ng d iscuss ion.The data col le ct ed by t he num erous t ec hni ques refe ren ce d t o above c ou ld be norm al i z ed by ma t hem at ic a l t ools t o fa ll bet w ee n 0 and 1. T he AA P mode l and a l g ori t hm serves as a gene ri c g ui de li ne and p rovide s a y ardst i ck t o t he a nal y s t t o g aug e and a ss ess t he answer of when and where t o s t o p ana l yz ing . In t he p rec edi n g di s c us sion, howeve r we hav e a l most al w ay s adva nce d t o t he ne x t st e p if we g et val ue s > 0.5.T hi s is s o bec a us e in try ing t o rea ch t he va lue of 1, w e a ct ual ly fal l i nt o t he t ra p of “ Anal ysis Para ly s is” 6 R e fe r e n c e s [ 1] Whit t en, B en tl e y. Sy s t ems A nal ysis And Design Me t hods .4 th Ed.120. [ 2] Whit t en, B en tl e y. Sy ste ms Anal ysis And Design Me thods.4 th Ed.120-164. [ 3] Whit t en, B en tl e y. Sy ste ms Anal ysis And Design Me thods.4 th Ed.10. [ 4] Kenne th E.Ke ndal l,J ul ie E . K endal l , S yste m s Anal y sis And Design, Page ( S):83-203 [ 5] John R . Ehrm ann And B arbar a L . Sti ns on, Joi n t F ac t-F i nd ing A nd The U se Of Tec hn ic al Ex pe rts, H t tp:/ /We b.M it .Edu /Dus p/E pp/Music /P df /J f f- H ow% 20to- Chapt 09 Y iu -Wing Leung , Le a s t-Squ ar e-E r r or E s timat e Of I n di vi d u al Con tr i bu t i on I n G r ouppr oj ec t , Educ at ion, Ie e e Tr ans act i ons O n,Vol ume 41, Is s ue 4, Nov 1998, Page (S):282 - 285 M.Be r m an,T he Rout ine Fi t ti ng O f Ki n et i c Data To Mod el s A Mat h emati c a l F orm ali s m For D igi ta l Compute r s, B iophy s ic al Journal, V olume 2, Iss ue 3, Pag es 275 -287 [ 8] Bradle y P. Car li n And Thom as A. L ouis, Bay es And Empiric al Bay es Me thods For D ata A nal ysis, Stat ist ic s And Computi ng, Spr i nger Net herlands, Vo lum e 7, Num be r 2 / June , 1997 , Page (S):153 -54. [ 9] Ronal d Mai er * , Defi n ing Pr oc e s s -O r i ente d Kn owl edge M an agem ent St r ate gi e s , Kn owl edge An d M an age me n t, Wil ey In te r s ci e n c e V ol u me 9 I s s u e 2, Page s 103 - 118 [ 10] Paul S. Ad le r , Av i Mande lbaum, V iê n Nguy en A nd E li zabet h S chwe r er, From Proje c t To Proce s s Management : A n Empiric al ly -B as e d Fram ework F or A naly z ing P r oduct Dev el opment Time, Jstor:Management Sc ie nc e,V ol ume 41,No.3(March 1995),Page (S):458-84 [ 11] Vi c tor Basil i, Barr y P erione ,So ft war e E rr or s And Comple xi ty :An Empiric al Inv esti ga ti on, Co mm u n i cat i on s O f T h e Acm Vo lume 27 , Iss ue 1 (J anuary 1984), Page (S) 42 - 52 Y ear O f P ubl ic at ion: 1984 [ 12] J.Ne i lson,Ite r ati v e Us er Inte r fac e Design,Compute r ,Iee e Tr ans ac ti ons O n,Vol um e 26,Iss ue 11,Nov 1993,P age(S):32-41 [ 13] Robe r t P . Hain ing,Spat ia l Data A nal ysis In Soci al And E nv ironm ent al Sc ie nc es, [ 6] [ 7] [ 14] Sudipt o Ban erje e, Bradle y P. Carli n, H ie rar chi cal Mode li ng And Anal ysis For Spat ial D ata,Cr c Press . [ 15] Nei l Wrig le y, Robe rt J ohn Be nne tt ,Quati tat iv e Geography,R ou tl edg e.