ebook img

The architecture of symbolic computers PDF

380 Pages·23.584 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview The architecture of symbolic computers

THE ARCHITECTUREO F SYMBOLICC OMPUTERS McGraw-HSielrlii enSs u percompauntdiP nagr alPlreolc essing ConsultEidnigt oKra:i H wang,U niversoifSt oyu theCrna lzfornia Hwanga ndB riggCso;m puteArr chitecatnudrP ea ralPlreolc essing Hwanga ndD eGrooPta;r alPlreolc essfionrSg u percompuatnedrA sr tificial Intelligence KoggeT:h eA rchitecotfuS ryem bolCiocm puters PeteMr. Kogge LakshmivaraahnadDn h alAln: alysisD easnidg onfP aralAlleglo rithAmrsi:t hmetic andM atriPxr oblems IBM FederSaelc toDri vision QuinnD:e signiEnfgf iciAelngto ritfhomrPs a rallCeolmp uters SiegeIln:t erconneNcettiwoonr kfso rL arge-ScPaalrea lPlreolc essing McGraw-HIinlcl., New yo rkS t.L ouisS anF rancisAcuoc klanBdo gota CaracaHsa mburgL isboLno ndonM adriMde xicoM ilanM ontreal New DelhPia riSsa nJ uanS aoP auloS ingapoSryed neyT okyoT oronto This bwooakss eti nT imeRso manb yt heC olleCgoem positUinoint inc ooperatwiiotnWh a ldmaGnr aphiIcnsc,. Thee ditowresr e DaMv.i Sdh apiarnod I raC .R oberts; thep roductsiuopne rviwsaosrL e roy YAo.u ng. ABOUT THE AUTHOR Thec over wdaess ignbeydC arlBaa uer. R. RD.o nnell&e Syo nsC ompanwya sp rintaenrdb inder. PETERM . KOGGE isa S eniToerc hniSctaalmff e mberin I BM'Fse derSaelc ­ torD ivisiIonnt .h icsa pacihteys erveosn s everal technical advisory boards responsifbolroe v erseeaidnvga ncecodm putitnegc hnologyb uasnidn edsis­ THE ARCHITECTURES YOMFB OLICC OMPUTERS .. rectiofonrs a pplyitnhge m. pHriess eanrte aosf e xpertairseeh ighlcyo ncur­ rencto mputenrosn,- voNne umanna rchitecatnudrp erso gramming languages, Copyrig©h 1t9 9b1y M cGra-wHilIln,c A.l lr ighrtess ervPerdi.n tientd h eU nited I real-tliomgeip cr ogrammianngde xpesryts temuss,eo fV LSIi nc omputiemr­ StateosfA mericEax.c ept pearsm ittuendde trh eU niteSdt ateCso pyriAgchtto f r plementatainodn s,t he theodreys igaonnfm d a tchipnagr alalleglo rithms. 1976n,o p arotf t hipsu blicamtaiyob ne r eproduocredd i stribiunat nedyfo rm orb y • anym eanso,r s torienda datbaa seo rr etriesvyaslt ewmi,t hotuhtep riowrr itten L Dr.K oggei sa lstoh ea uthoorft he internabteisotnsaelTl hleeAr r chitecture permissoifot nh ep ublisher. f ofP ipelinCeodm puterhsa,sp ublishfreedq uentilny ttheec hnilciatle rature, andh asl ectureexdt ensiavtem layj ourn iversaintdir eess earccehn teHres . has 2 3 4 5 6 DO7C 8D O9C 90 5 4 I0 3 2 participiantt heede ditionfmg a nyc omputpeurb licatiinocnlsu,d tihneIg E EE TransactoinoC nosm putearnsd t heI nternatiJoonuarln oaflP aralalnedlD is­ ISBN 0-07-035596-7 tributed Compauntdi nwgaG,se neraTle chnical Chaoiftr hmea1 n9 8I9n ter­ nationCaoln ferenocneP aralPlreolc essiDnrg.K. o gger eceivheidsP h.Di.n LibraorfyC ongreCsast aloging-in-PDuabtlai cation electreincgailn eerfrionmg Stanford Univerissai Fteyl.l ooHwfe t heI EEE, and mae mbero ft he IBAMc ademoyf T echnology. Kogge, PMe.t,e( rd ate). Thea rchitecotfsu yrmeb olic comIpP uetteeMrr.s K ogge. p. cm.--{McGraw-Hilli nss uepreirecso mpuatnidpn agr alplreolc essing.) Includiensd ex. ISBN0 -0-7035596-7 { I.C omputearr chitectu2r.eE .l ectrodniigci ctoamlp uters-Deasnidg n I constructioIn.T. i tle.I IS.e ries. QA76.9.A73K64 1991 004.2'2-dc20 90-42806 t v L CONTENTS Preface xv ParIt Fundamentals Chapte1r FundamentoaflC so mputation 1 1.1 Probleamnsd A lgorithms 2 1.2 LanguageSsy:n taaxn dS emantics 5 1.3 SemantMiocd els 7 1.4T hev onN eumannM odelo fC omputation 10 1.5O bjects 13 1.6 Sets 13 1.7 Relations 16 1.8 Functions 19 1.9P roblems 25 Chapte2r ExpressiaonndTs h eiNro tation 27 2.1 Expressions 27 2.2 FunctiDoenfi nitions 33 I 2.3 FormaSly ntaNxo tation 36 2.4 Problems 40 I Chapte3r SymbolEixcp ressiaonndAs b straPcrto grams 41 I 3.1 SymbolEixcp ressions 42 I 3.2 AbstraPcrto grams 52 3.3 Recursion 55 1/ 3.4 AbstraScytn tax 62 3.5P roblems 65 i ix u __ - x CONTENTS CONTENTSx i 7.4T heB asiIcn strucSteito n 153 ParItI Function-BCaosmepdu ting 7.5C ompiliSnigm plSe- expressions 163 7.6S amplCeo mpilation 168 Chapte4r LambdaC alculus 67 7.7T he FunParrogb lem 171 4.1 SyntaoxfL ambdaC alculus 68 7.8O ptimizations 173 4.2 GenerMaold elo fC omputation 69 7.9P roblems 177 4.3 StandaSridm plifications 71 Chapte8r MemoryM anagemenfotr S-expressions 179 4.4 Identifiers 72 4.5 SubstituRtuiloens 74 8.1A llocation 180 4.6 ConversRiuoln esR,e ductiaonnd,N ormaOlr der 77 8.2 Mark-SweCeopl lection 183 4.7 TheC hurch-RoTshseeorr em 80 8.3 ReferenCcoeu nts 192 4.8 Ordeorf E valuation 81 8.4G arbagCeo mpaction 195 4.9 MultipAlreg umenCtusr,r yianngd,N aminFgu nctions 83 8.5A lternaLtiisvRtee presentations 204 4.10B asiAcr ithmeitniL ca mbdaC alculus 86 8.6P roblems 211 4.11B ooleaOnp eratiionnL sa mbdaC alculus 88 4.12R ecursiionLn a mbdaC alculus 90 Chapte9r Demand-DrEivvae luation 213 4.13P roblems 92 9.1E xpliDceilta yasn dF orces 215 9.2P rograEmx amples 217 Chapte5r A FormaBla sifosr AbstraPcrto gramming 94 9.3 RecipePsr,o misFesu,t ures 219 9.4L azyE valuation 224 5.1 A BasiScy ntax 95 9.5S treams 229 5.2 Constants 96 9.6C ontinuations 233 5.3 FunctiAopnp lications 97 9.7P roblems 241 5.4 ConditiEoxnparle ssions 98 5.5 LetE xpressionsD-Leoficnailt ions 99 Chapte1r0 LISPV:a riatiaonndsI m plementations 243 5.6 RecursiDveefi nitions 103 5.7 GlobaDle finitions 106 10.1T heO rigiLnIaSlP 244 5.8 Higher-OFrudnecrt ions 106 10.2S cheme-A Pure LISP 257 5.9 An Example-SymbDoilffiecr entiation 108 10.3C ommonL ISP-AM odemS tandard 265 5.10P roblems III 10.4M ultiLISPP-aAr alLlIeSlP 270 10.5T heC ADR Machine 273 10.6C ADR Derivatives 278 Chapte6r Self-Inteartpiroent 113 10.7O theLrI SPM achine.s 285 6.1A bstraIcntt erpreters II5 10.8B enchmarking 291 6.2L ambdaE xpressiaosSn -sE xpressions II9 10.9P roblems 295 6.3A n ExpandeIdn terpreter 124 6.4A ssociaLtiisotns 127 Chaptre 11 CombinataonrsdG raphR eduction 296 6.5C losures 133 11.1B asiCco mbinators 297 6.6R ecursiCvleo sures 135 11.2B rackAebts traction 300 6.7C lositnhgeL oop-Read-Eval-Print 137 11.3A dditioCnoamlb inators 303 6.8 Problems 138 11.4O ptimizations 309 11.5G rapRhe duction 312 Chapte7r TheS ECDA bstraMcatc hine 140 11.6S upercombinators 322 7.1L isMte mory 142 11.7C ombinatMoarc hines 326 7.2B asiDca taS tructures 143 11.8T heG -Machine 332 7.3T heS ECDM achinSet ruc�ures 150 11.9P roblems 336 xii CONTENTS CONTENTSx iii Chapte1r2 OtheFru nction-BCaosmepdu tiSnygs tems 338 16.4T heP ROLOGI nfereEnncgei ne 463 12.lL ogo 340 16.5S peciFaela turoefPs R OLOG 471 12.2F P 343 16.6S omeE xamples 480 12.3H ope 351 16.7P roblems 483 12.4T reeM achines 357 12.5A lice 359 Chapte1r7 TheW arreAnb straMcatc hine 486 ""'' 12.6R editlow 363 17.1P rograSmt ructure 488 12.7F unctiCoanc hing 365 17.2M ajoDra taS tructuarnedSs t atRee gisters 491 12.8P roblems 368 17.3M emoryW ordF ormat 494 17.4S implifCiheodi cPeo int 496 ParItI IL ogic-BaCsoemdp uting 17.5S implifWi AeMd I nstrucSteito n 497 17.6P ROLOG-to-WACMo mpilOevre rview 512 , I 17.7S upportBiunigl t-ins 516 I, I Chapte1r3 A LogiOcv erview 369 17.8D etailEexda mple 520 13l. InformaOlv erview 370 17.9P roblems 525 13.2F ormaLlo giScy stems 376 13.3P ropositLioogniacl 384 Chapte1r8 OptimizatainodnE sx tensions 526 13.4A SimplIen fereEnncgei ne 387 18l. An OptimizWe AdM Summary 527 13.5A SimplPer oblem 390 18.2O ptimizSiinngg Cllea uses 527 13.6P roblems 393 18.3O ptimizMiunlgt ipCllea uses 539 18.4O theOrp timizations 544 Chapte1r4 PredicLaotgei acn dt heF irsItn fereEnncgei ne 394 18.5A nnotatiaonndBs a cktraCcokn trol 550 14.lB asiScy ntax 396 18.6P roblems 562 14.2B asiIcn terpretations 398 14.3S tandard Expression Forms 400 Chapte1r9 PROL OG Implementations 564 14.4H ornC lauses 407 19.1M easuring PROPLeOrGf ormance 565 14.5D ecidabiIlsistuye s-The HerRbersaunldt 410 19.2C ompilitnoAg ConventiCoonmaplu ter 570 14.6T heH erbraInndf ereEnncgei ne 415 19.3T heO riginala nTdiW cakr ren Machine 578 14.7P roblems 417 19.4F ifth-GenerPaRtOiLoOnG Machines 580 19.5T heP LM andi tDse rivatives 585 Chapte1r5 FundamentoaflP sr actiIcnafle reEnncgei nes 419 19.6T heL ow RISCA pproach 589 15.lP attern MatacnhdUi nnigfi cation 420 19.7U sinAgs sociaMteimvoe ry 596 15.2C ommonI nfereRnuclee s 426 19.8P roblems 605 15.3R esolution-IBnafesreedn Rcuel es 428 15.4R efutatCioomnp leteness 431 Chapte2r0 All-SoluItnifoenrse Enncgei ne 607 15.5D ecisiPorno cedures 432 20.l DatabaOsvee rview 608 15.6S trategieOsv-eArnv iew 438 20.2 TheR elatioMnoadle l 614 15.7E xample 443 20.3 ComputatioEnxaplelnys iOvpee ratoJrosi:n s 619 15.8D eductiTorne es 444 20.4 RelatioAnlagle bra 623 15.9P roblems 446 20.5 Implementations 624 20.6 ProductionS-yRsutleem s 633 Chapte1r6 TheP ROLOG InfereEnncgei ne 449 20.7 TheR eteA lgorithm 639 16l. Structoufra e P ROLOG Program 450 20.8 An AbstraMcatc hine 645 16.2P roceduIrnatle rpretation 455 20.9 Machines 651 16.3A PROLOG-LikPer opositIinofnearle Enncgei ne 457 20.10P roblems 655 I :: xiv CONTENTS Chapte2r1 ParalIlnefle reEnncgei nes 657 21.lT axonomoyf L ogiPca rallelism 658 PREFACE 21.2V ariatioofOn Rs Parallelism 665 21.3V ariatioofAn NsD Parallelism 676 21.4A ND/ORP roceMsosd el 685 21.5C ommitted-CLhaonigcuea ges 694 21.6I mplementations 702 21.7P roblems 706 References 707 Index 721 Theries a revolutliooonm inignc omputer scnioetjn ucseit,n t hes peeadn d memoryc apaciotfty h ec omputing equipumseenb,tu tiw nte h ev erwy ayw e thinakb out, spaencdip feyr,fo rmc omputatiEosnsse.n tiatlhleby a,s ivco n Neumannm odeolf c omputiwnhgi,c hha sr uleudn challenfogre tdh el as4t0 yeariss,e ncountenroitno nge bsuetv erpaolt entihaelalvyy weicghhatl lengers. Thist exrte preseannta st temtpotp rovisdoem ei nsigihntt toh ec onceptual seedosf t her evolutainodnt ,h ed irectiinow nhsi cthh eayr eg rowinAgs.s uch, itw asw rittweint thw oo veraglola lisn miFnidr.s wte, wishd etvoe lionpt he readear fi rmu nderstanodfit nhgem athematircoaolto sft wot runkosft hese newc omputatimoondaell sS.ec ondw,e want dteom onstrianta e v eryc on­ cretfaes hiohno ws ucmho delcsa na nd habveee np ut ipnrtaoc taiscr ee aplr o­ gramminlga nguages rounnr neianplgr ocessoofrts,ew ni trhe volutiodnea­ry sigcnh aracteriAscthiicesv.i ng thegsoea lstshw oou lgdi vet her eadetrh e abiltiotfoy l lowo,rb ettyeertp articiipnac,to en strucotfti ohnel anguaagnesd computisnygs temtsh at wbieclolm teh ee verydcaoym putatitoonoalolsf t o­ morrow. I, VIEWPOINT ii Thev iewpoitnatk einn t hitse xits t haotf a practiccoimnpgu tearr chitect, wherea rchiteisct taukerinen i tbsr oademseta ninags " thea rto rs cienocfe buildi.n.g.e, s ,.p h abitasbtlreu ctu(reWse"b stTehri'Irsnd t ernaDtiicotniaol­ nary)I.no urc asteh es tructuarreecs o mputisnygs temasn,dt heiirn habitants aret hes ophisticlaatregden,lo yn numerpirco grams fotuondda iyn p rototype I ' ' form inh ighilnyt eractive environmeinnttesl,l iagrietnnitcfieecl,il aidlga etnat­ I !1 baseas,n ds imilar adavpapnlciecda tbiuotmn osv,i ng raipnitdtolo ym orrow's Ii' maisntraem. Whenb uildsiuncgh s tructuart ersu,ae r chitmeucsttc onsidtehrei nterplay 'I betweenb otht heh ardwa(rmea chionreg anizaatnidot nh)es oftwa(cr oem pil­ ' rl: ers,i ntperretearnsd,r untime routinetso)s unseteatdiheneid n habitaInnt s. 1I,: 1 i!' xv xvi PREFACE PREFACEx vii thitse xtth ee mphasoinst hesset ructausrpaelc itscs o ntinuobuasl,a ncaendd, mappinfrgosm ones eto fs ymboltsoa nothewri,t shu bstitusteirovnia nsgt he interlewaivtedhdi scussoifot nhsefo rmativbaec kgrouonfdt sh ea ppropriate conceptiumapll ementavteihoinc Tlhei.ms o deo fc omputihnagss omes imilar­ programmilnagng uagFerso.mp ersoneaxlp erienictie si, m possitbold oeo th­ ititeost hes tructoufrc el assilcaanlg uagients h at odnees cribgeisv,e n some erwisAe .c levhearr dware organiidzeiaast isoinm ply gwaatseitsfet dh es up­ speciifincp uttsh,ep rocetsosfo llowto b uiludp t hed esirreeds ulIttd. i ffers porsto ftwardeo esn ote mploiyte fficientalnyd,a high-perforcmoamnpciel er from classiccoamlp utiinntg h atth eriesn od ireecqtu ivalteon,ot r n eedf or, techniiqsuo efo nlayc ademiinct eriefts hte u nderlying hahradsn woah roeo ks sucht hinagss" variablceosr"r espontdois npge cimfeimco ryl ocatio"nass,­ withw hictho s ustaiitnF. u rthenre,i thhears vailftu hee appliocrap trioo­n signmesntta tement"sp,r"o gracmo unters",s"e quentsitaalt emeenxte cu­ gramminlga nguaugsee dd oesn otp ermiotr e ncouraegxep ressitohnasut s e tiont'o'c hangteh ec ontenotfsv ariabilnea s c ertasienq uenocfes teposr, them. "sideeff ectsw"h iccha np otenticahlalnyg seo me objaetac gtr eadti stance Thef loowf p resentarteifloenc tthsiv si ewpoiwnitt,th h ei nitfoicauls o nt he from thec urresnitto efc omputatTihoenr.ae r en o" commandst"oc hangteh e underlymiantgh ematifrcaamelw orktsh adtr ivteh ev ariocuosm putimnogd els, valueass signteond a mesI.n steaodn,es impl"yd eclartehsa"st o mes ymbol followebdy d etaildeids cussioofhn osw somek eyl anguagbeass eodn such "is"t hes ameo bjecastt hadte scribbeysd o me expresasnidbo uni,l dnse w frameworkasff ectth ea rchitecotftu hrece o mpilaenrdm achinhea rdwaursee d objecotnst opo ft hat. tos uppotrhte mI.n g eneratlh,et ransibteitowne emna thematimcoadle lasn d Interesteinnogulgyh , htahsin so effecotn thec omputatipoonwaelr o f reailm plementaitsit ohnrso utghhe u seo fa bstrmaaccth ineis.,e c.o,m putaerr­ function-pbraosgerda mmlianngg uages. Fwuhretndh oenrep, r operiltyy i,e lds chitecttuhraemtsa yn otn ormalbleyi mplemenatse d rehaalr dwabruet t hat programmlianngg uaignew sh icfuhn ctioanrset rul"yfi rst-ccliatsisz enis."e;. , aree speciaelalsyyo rg oodt argefotrs c ompilfoerr ssu clha nguagOensc.et he anythitnhgac ta nb ed onet oo bjecstusc ha si ntegienrc so nventiloannaglu ages abstraarccth itecatnudmr aet chicnogm pilteerc hnolhoagsby e edni scussreeda,l canb ed onet ofu nctionTsh.e yc anb e passeads a rgumenetxsa,m inebdy languagmeasc,h ineasn,di mplementations becomdei rreeclta teixvteelnys ions. othefurn ctionasn,dr eturnaesdr esulTthsi.ms e anst haat p iecoef c odec an Sincteh ee mphasiisos n a rchitectthuear pep,r oatcahk etno wartdh em ath­ dynamicatlelsymt o,d ifayn,d g eneranteew p ieceosfc ode, whmiacyhe ither ematiicsso fi ntroductihnekg e yi deaasn dn otatitohnad tr ivteh el anguage bes toreodre xecutiemdm ediately. definitioannsdn, o to nd etailed parnodao rfgsu menStism.i lartlhyer, e ader Logic-bacsoemdp utionng ,th eo thehra ndi,n volvfiensd insgu bstitutions shoulndo te xpecttot akaew ayt hea bility teoip trhoegrr aimns pecific lan­ thamta kec ertaeixnp ressioorns se,to sfe xpressihoanvses, o med esirperdo p­ guageosrt od upliciandt eet ail particuldaers imgoanrcsp h rionger ammsiynsg­ ertiIetsms.a thematibcaasiles t haotfr elations-niadmeenltyi,f wyhiennvg a r­ temsR.a thehre,o rs hes hould extpore eccte iavfo er mali ntroducttosi oent s iousse tosf o bjechtasv es omep roperPtryo.g rammiinsag g aidne claraitniv e ofa rchitectural desitghnah tap vrpeir nocviepsdlu eccse ssifnau clt usayls tems. thaat t ypicsatla temesnitm pl"yd eclartehsa"it fc ertaoibnj echtasv e some I properttiheesn, soomteh eorb jechtasv eo theprr opertCioemsp.u tatoifotne n i cosisotfsm ethodosfg enerating psoutbesnttiitaulta inodtn hse,in d entifying I THE NATUREO F THE BEASTS � I ' whicohn esa rea pproprioaftteeb,ny s ophistipcaattteedmr ant chibnegt ween Allt hem ajocro ncepdtiss cushseerdre e volavreo untdh ei detah amtu cho ft he propertoifet sh ep roposseudb stituatnidop nrso pertoifet sh ep robleTmh.e computatoifto hnefu ture wpirlolc essysm borlas thtehra jnussti mplneu mbers, underlypirnogc essing iesrn egsipnoen sfoirb filned ing an apptrroapcroeif a te I''I I I ora ggregaotfen su mberTso. be specifitch,i mse ansc reating, converting, �uchs ubstituwthieonnp sr esenwtietdah s peciifincp uTth.i tsr acyei elds such I evaluatainodni ,n terpretingo fss ytmrbionlgss, ubsyu tahlels yub stituotri on mterestcianpga bilaistt iheeas b ilitota ys k" why"o r" how"t hea nswewra s matchionfgo nes eto fs ymbolisn toor w itahn othseerti n s omefa shionW.h ile deriveodrt; o r unt hes amep rogram wtawyos e,i thienrt hec lassifcoarlmo f thimsa ys eemt ob et heo pposiotfew hato ned oesi nt odayp'rso gramming "giveann i npuqtu estifionnd,a no utpuatn swero,r",m oreu nusualilnty h,e languagaensy,t hitnhga cta nb ed onei na conventiloannaglu acgaena lsboe form," giveanna nswewrh,a tw ast heq uestioAng?a"i n, tahreecs aep abilities doneu sintgh esnee wc onceputssu,a lmloyr ee asilmyo,r ec oncisealnydm, o re thaatr en otfo undi nc onventicoonmaplu tisnygs tems. naturaFlulryt.h etrh,ee liminatimoann yoo fft heo ldc rutchoefst eonp enusp opportunifotri tersu e paralalnedlt ihsuvmse, r yhi gphe rformanscoem,e thing ORGANIZATION thaits s implnyo tp ossibwliet tho dayc'osm puting models. Thet wop ossibdlier ectifoonrss u chp rocessdiinsgc usisnet dh ibso oka re Thec hapetrsi nt hitse xatr ed ividiendt toh remea inp artTsh.e fi rspta rte,n ­ function-bcaosmepdu tianngdl ogic-bcaosemdp utiBnogt.h a reo ftecna lldeed­ compassiCnhga pte1rt sh rou3g,hr evietwhsem athematifcuanld amentaanlds claratilvaen guagaesso pposetdoc onventiiomnpaelr atliavneg uages. �ommonn otatiuosne tdh roughourte stothf te h et exTt.w ot hinagrse o fm ost Function-bcaosmepdu tiinngv olv"easp plyinsgu"b stituotfio opnesr ands importnaceh ereF.i rsitsa ni ntroducttoit ohnec oncepotfs ubstituatsi on inteox pressiaosin tsms a jocro mputatioopnearla tiTohnen. a mec omesfr om some.thign useftuold escribcionmgp utatiSoenc.o nids a bstrapcrto gramming them athemiactatlh eoroyff unctiownhsi,c ihn volvtehsefo rmald efinitoifo ns notaotni andt hec oncepotf a bstrascytn taTxo. gethetrh,e yw ilble used xviiiP REAFC E PREFACEx ix REQUIRBEADC KGROUND throughtohuett e xtto p ermivte rsyh orbtu,t e xtremerleya dabdlees,c riptions ofh owt oi ntrepprreotg rawmrsi ttiennt hev ariomuosd els. Thist exwta sd esignteodb es uitafobrl pee ople wai btrho arda ngoef b ack­ Theo thetrw om ajopra rtesa chi nclundien ceh aptearnsd,c ovefurn ction­ groundIsn.m ostc asetsh eb asimca teriiascl o mpatiwbilteha n u pper-level baseadn dl ogic-bacsoempdu tirnegs,p ectiEvaeclhyp .a rsth areas c ommon computesrc iencoer c omputeern gineeruinndge rgradbuaactkeg rouInnd . organizatFiiornsi.tsa seto ft hreceh aptedresv otetdot he commmoant he­ termosf a recentplryo posuendd ergradcuoamtpeu tsecri enccuer ricu(lcufm. maticfauln damentoaftl hsep articcuolmapru tatimoondaellT .h ee mphasiiss AlfsB erzti"sAs ,M athematicFaolcluys eCdu rriculfourm C omputeSrc i­ oni ntroductihneb ga siicd eanse edetdo e xprecsosm putatiiontn hse m odel, ence,C"o mmunicoaftt hAieCo Mnv,so l3.0 ,n o.5 ,M ay1 987p,p .3 56--365), ando nt hoskee ym athematical wrheiscduhel fitnse rtahneg oer l imiotfss uch a minimbaalc kgrouwnodu lidn clu"dLea nguagaensdP rogram(sC"S 5)",A d­ computations. vancePdr ogrammin(gC"S 7)",D iscreMtaet hemat(iIc)s, a"n d" Computer Nexti sa seto ft hreceh apterss ttahrawttis t ah c leafno rmo ft hep rogram­ Architect(uCrSeI"5 0T)o.g ethetrh,e sceo ursweisl gli ve rtehaed etrh en ec­ mingl anguamgoes tw idely recoagsnb ieziendrg e presentoaftt ihveme o del essarfaym iliarwiittyh sevceornavle ntipornoaglr ammlianngg uagseosm,ea d­ discussiendt hatp artT.h isi sL ISPf orf unction-bcaosmepdu tianngd vancepdr ogrammcionngc epstusc ha sr ecursitohnec, o nceopfta ni nstruction PROL OG forl ogic-bcaosmepdu tiFnogl,l owingi stt hhedi ess cripotfia os ni m­ seta rchitechtouwrm ea,c hines mbiebg uhittl ote xecutthee mt,h ec onceopft � plea bstramcatc hinfeo rw hiciht i se asyt og eneraetffiec iencto defr om po­ a compileanrd,a ni ntroducttoit ohnen otioonffo rmald efinitioofna s l an­ gramwsr ittientn h el anguaTghei.is n t urpne rmisthso rbtu tc ompledtees cnp­ guages'esm antics. tionosfh owt oe xtentdh essei mplceo mpilearnsda bstramcatc hinteosr eal Althoungohtr equirtehde,ra er ec oursaetst he mastleerv'ewslh icghr eatly II I, compilefrosrr eamla chines. enhancae s tudenetd'usc atioenxpaelr ienIcnte e.r ms ao pfr oposmeads ter's 1,, Finalelayc,h p arcto ncludweist ah s eto fc haptetrhsad te scriobteh eprr o­ curricu(lcufKm.. I .M ageelt a l.",R ecommendatfiooMrna ss terL'esv ePlr o­ gramminlga nguaganedsm achinfeosrt hactl asosfc omputatwiiotnhe, m pha­ gramisn C omputeSrc ienceC,o" mmunicoaftt hiAeoC nMvs,o l2.4 ,n o.3 , '11 siosn t hosteh ahta ves omeu niquceh aracter(issutciahcs cs o mbinatroerlsa,­ March1 981p,p .1 15-12a3 )p,r iocro ursseu cha s" CompilCeorn struction" tionadla tabaseosr, p roduction-sryusltee mso)r exhibifeta tureosr (CS19o)r( BerztCiSs2s2,1 w)o uledn ablae t ypicraela detroa ctuaaltltye mpt performapnocses ibil(istuicaehss e xtenspiavrea llelisamr)en otthf aotu nidn usintgh em aterifraolm thitse xats a guidteo t hec onstrucotfiw oonr king theo thers. translafotro rresa lla nguages. Likewisoet,h ecro ursseusc ha s" HighL eveLla nguagCeo mputeArr chi- AUDIENCE (t ectur(eMsa"g eClS,2 6a)n d" LargCeo mputeArr chitectu(rMeasg"eC lS,2 6) wouldp ermiat s tudetnotr ecognimzoer ef ulltyh osper ocessdoers igcno n­ Whilteh itse xwti lble o fv aluteo a nyonien teresitnue ndc onventicoonma­l cepttsh aatr en ovetlot hesnee wm odesof c omputiSnigm.i laarc loyu,r sseu ch putinigt,is n tendaeudd ienacreet hosset udenatnsdp racticpirnogf essionals as" FormaMle thodisnP rogrammLianngg uage(sM"a geClS,2 0w)o uladl low interesitnte hdei nnarodfsc omputation. Spetchiiifisnc calluldyhe,as r dware for a deepa ppreciaotfis ounc ht opiacssp arallelliaszemyv, a luatbiaocnk,­ architercetssp onsifbolrne e w computeorr ganizatainodnt sh,e isro ftware trackianngd,n ondeterminFiisnma.l al yc,o ursien ' 'Artificial Intelligence'' counterpwahrotc so nstrmuacttc hilnagn guage-proacnedss syisntge m-support (MageClS,1 2o rB erztiCsSs1,6 0o rB erztiCsSs2,7 2) woguilvtdeh es tudean t programs. High-perwfoorrkmsatnactpeir oonj ecctosm et om inda sa prime backgrouwnidt hw hictho j ustibfyye xperiewnhcyet hec oncepdtess cribed examploeft his. ares oi mportaanntd,fo r whatt ypickailn dosf a pplicattihoenaysr ep articu­ Nexta,p plicatpiroongsr ammefarcesd w itht het asokf c hoosinlge,a rning, larlsyu ited. andu sinag n on-voNne umann langwuialfglie n tdh et exuts efiunld efinitnhge origiannsd m eaninogfst hose language tchoaantrc eef potrse itgonc onven­ ·, ,' I ,, tioncaolm ptuinag,n d uinnd erstantdhiekn egy i mplementaitsisounte hsa atf ­ CURRICULUUSME :1 'I··" fecat p rogramp'esr formaonnca e r eamla chine. ��: As implied atbhoevm ea,t eriiantl h itse xcta n( anwda sd esignteod)s upport •i' Additionparlolgyr,a mmlianngg uadgees igncearnsu set het exats a starting 1•. : poinfto ri dentifnyeiwnc go ncepitnse xpresscionmgp utatiwointe,hm phasis advancuendd ergradaunadtg er aduate-cloemvpeult er archictouercstieunsr e ·;�I I ·1:'. 11,�o n theimra thematibcaaclk grouannddso n implementatteicohnn iqauneds severwaaly sF.i rsatn,d m osti mportainnt th ea uthorv'ise wi,si tdsi rect use ';�- ina courseen titsloemde thilnigk" eA lternaCtoimvpeu tinwgh,i" ch brings to­ trade-offs. ·:;-" �;·I. I Finalrleys,e archineu rnsc onventicoonmaplu ticnagnu set hem aterihaelr e gethwehra ti st odasye verdailv ergtirnagc iknsm ostc omputsecri enccuer ric­ �;I I"" I asa basifso ru nderstantdhiefn ogu ndatisotnrse,n gtahnsdl, i mitatoifot nhse ulan,a melpyr,o gramming languages, computeran dpo rrogcaensis­or design most up-tom-oddaetlesa ,n dc anu set haats a mapt oi dentpioftye ntsiuacl­ zatioann,d s ystesmo ftwaSruec.h a courscea nb efo rmatteadtl easttw o .'I< waysA. single-semveesrtseicroo nu lcdo veorn ltyh ec orceh aptefrrosm the 1! cessors. XX PREFACE PREFACE xxi twom odelwsi, tah fe w lessoantst hee ndo ns electteodp itcosb ec hosebny herep utts heq uantifiveadr iabalfet etrh eq uantifiaenrds eparatbeyad ''I'' thei nstrucTthoirws.o ulidn trodtuhceeb asipcr inciipnlveosl vientd h em od­ from thee xpressbieoinn cgo ntrolel.egd.,3, x l( re(ax/\)l ( 1< x)/\ (x<2)). elss electaenddg ive, thtrhoeur gehl evaanbts tramcatc hineasn,o vervioefw Whilteh imsa tchenso p articusltaarn dairtdi ,sc onsistaenndti ,tm inimizes theiirm plmeentatitoenc hniquTehsis.w oulbde m ostc ompatiwbilteah na d­ confusiowint thh eo thenro tations. vanceudn dergradsutautdee bnotd y. Finalilnyt ,h ep rintoifnt gh itse xtt,h efi rst tai tmeer mi si ntroduictie sd A two-semesvteerrs icoonu lcdo vebro thm odelosfc omputiinnmg o red e­ highligihnta ef do nltit kheiT hsen. a meso fo bjevcartisa,bs leetrs,se ,l ations, taiWli.t hc aret,h es emestecrosu lbde s tructutrobe edr elatiivnedleyp endent an�f uncatrieao ss nhsow nw,h ilkee yworfrdosm programmifrnagg mentasr e ofe acho thepre,r mittsitnugd enttots a kteh es econwdi thotuhtefi rsto,rv ice i?a sepaa_rtefo nt,a si ni f-theSynn-taeclttesircem d.se finebdy B NF produc­ versOan. ew ayt od ot hiwso ulbde t oc ovefru nctional coimntp huetfi irnsgt tionasr em brackeatss( ,e xpression). semestearn,dl ogic-bcaosmepdu tiinntg h es econAdg.a int,h eriess ufficient Fors implicaintdey a soef r eadabitlhietfo yn,t su sedi nfi gureasn dt ables materifoarl a ni nstructtopo irc akn dc hoosteo piocfss peciianlt ereTshti.s area subseotft hese. ,'. woulbde m ostc ompatiwbilteah g raduastteu deanutd ience. ,:,> Thea uthohra ss uccessfuusleldby o thfo rmatisn c oursteasu ghotv ert he CONCLUSION . i las5ty earast t heS tatUen iversoiftN ye w Yorka tB ingham.t on � Alternatitvheitlsey x,ct o ulbde a v aluabalueg mentattioso env ereaxli sting Thise :'th asb eena laboorfl ovfoer thea uthoIrn.m oret ha2n 0y earass a courseCsl.a ssiensL ISPo rP ROLOG,fo re xamplceo,u luds et hea ppropriate practiccmogm putaerrc hitaencd1t 3 y earass a t eachIeh ra,v ec omet or ealize chapteforrs a morei n-depltoho akt t hel anguagseesm'a ntiacnsdt ypiciaml­ thef undamenitmaplo rtanocfte h ei ntertwionfil nagn guafguen damen�ta ls plementattihoancnsa nb efo und mions tp rogrammmianngu alAsd.v anceadr ­ compil�ant,ai ndm achinaer chitecWtiutrhe .s uch undercsotmaenasdd ieneg ii chitectculraes sceosu luds et het exats a sourcoefi nformatifoorn s ymbolic sensoef r ightnaensdsc ompletiaonndt, h ea biltiotsy e eb ehinad s nippoeft processjourssat,s o thetre xtasr eu seda sr eferenfocre psa rallel, piopr elined, codet ot heg uidipnrgi nciapnldei sm plementamteicohna nistmhsa stu ppoir.tt distribpurtoecde ssIonrts h.ic sa saen a dvantaogfte h itse xitst hec losmei xing Thiiss w hatI w antt op astso t her eader. 'i ofc omputoerrg anization taencdht nhiecq oumepsi lteerc hnolongeieedse tdo �Tot hosree adewrhso h avea ctivpealryt icipiantt heedd e velopmoefnm ta ­ I, takea dvantaogfet h emS.i milaraldyv,a ncperdo grammilnagn guacgoeu rses chneso rl ang�ueast hasth oulpde rhabpesr eferencheedr beu t areI a npootl,­ couluds et hitse xats a ni ntegrarteefde retnoca e w ide varoifle atnyg uages. ogizTeh.e fi elids s ol argaen,dg rowisnogfa stt,h aittw oulbde a monumental Finaltlhye,t exhta sb een desiegxnpeldi ctiotb lely a rgely self-icno ntained tastko d oa dequajtues titcoae l all ternatMiyv heosp.e i st haIth avep icked concepbtustw itehx tensrievfeer encteots h eo rigisnoaulr cepse,r mitteixn­g onest hahta ves omel ongeviatnydI, w ouldb eo verjoyteohd e arfr om re­ periencperdo fessitoonu asleis t fo r eithreerfe renocres elf-study. searchwehrosc ouledd ucamteeo no thetrhsa sth oulbdec onsideforrefu dt ure editsi.o n !� Fianllyal,t houIg hha vee njoywerdi tienvge rsye ntenocfet hibso oki,t i s _ NOTATION notm me �loneM.y visioofnw hati tm eanst ob ea na rchitect frcoomm meys �� Ina nyb ookt haatt tempttosc ovears b roaad r angaes t hiosn ed oesi,tb e­ .D.a dvisaotrS tanforHda,r olSd. S toneA. l argmee asuroefc reaticvrei t­ comesi mportatnopt i cnko tatitohnai tsb otihn terncaolnlsyi staenndrt e flects Ic�sm ands uggested improvement bmeyl sotnugdse nwatitst S hU NY atl easatp proximawthealty rseyaslt emuss eW.h ilseo mewhaatr bitraanrdy , Bmghamtownh,o s uffertehdr ouegahr lvye rsions of thea ndmw aittethrh iea l, guarantteoec do nfliwcitt sho mep arotf e achr eader's backtghreno outnad­, fa�ultayt S UNY andm y coworkeartIs B M,w ithout wchoomspel estuep port ' ,, tioIn h avec hoseins a na ttemaptts tandardizFaotrie oxna.m plfoer, l isIt s thiwso rkw ouldn oth aveb een possTihbele ed.i torsitaalaff tM cGraw-Hlil l havea doptetdh ed otn otatiaonndp arenthneostiast ioofLn I SPT.h idso esn ot especiaD�lalvyiS dh apiarnod I rRao bertasl,sh oa da majoirm pacotn t hefi na agrewei tmha nys tandaPrRdO LOG notatiobnusth, a sp rovevde ryr eadable producIt .o uldal sloi ek tom entiotnh efo llowirnegv iewewrhso p rovided _ throutghhe y earOsn. ei mpacotfi tuss agoen o thenro tatihoonw,e veirs,i n valuabclnetq Iueso ne arlier odfrt ahfemt asn uscriApnta:n dA garwaMla,s sa­ _ Backus Nfoarumr ( BNFs)y ntadxe scriptTioom nisn.i miczoen fusitohnefo, r ­ chusetItnss titoufTt eec hnoloJgoyh;nP .H ayesU,n iversoifMt iyc higaJno;h n matc hosehne rdee liberadtoeelsny o tu s"e( ) "a sa b asiBcN F symboTlh.e ir PetersUonni,v ersoifAt ry izonaan;dK eshaPvi ngaCloirn,e lUln iversPietry.­ placiest akebny " {)." haps?.h oweveerv,e nm orec redsihto ulgdo tmoy familyp,a rticulmayr sloyn As anotehxearm plnee,a rlayl lt heb asimca thematniecesd eidn volves M1caelfo r keyboardtihneig n dexa,n de speciamlyl wyi fMea ryE llefno,r quanteixfiperde astss oimoep noisn wth,e rea quantisfiyemrb oolf s omes ort puttmugp w itlho sMto ndayn ighatnsd extepnedreido odfsh and-keyboaatr­d (sucahs V ,3 ,X. , )c ontroslosm ea specotfa n identiufiseargi'ens s omee x­ tahcment. .•• pressiEoanc.ha uthosre emtso a dophti so rh ero wnn otatifoorn s uche xpres­ PeteMr. K ogge sionwsi,t uhs eo f".,""( ) ", andt hel ikbee incgo mmonT.h en otaticohno sen ' .

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.