0% found this document useful (0 votes)
140 views

WorldWideWeb Proposal For A HyperText Project

This document proposes the WorldWideWeb project, which aims to create a hypertext system linking information across CERN. It would allow users to easily navigate between different types of information (reports, notes, databases, documentation, etc.) stored on various servers, through hypertext links. The project has two phases - first implementing basic browsers and analyzing information needs, then expanding to allow user-added content. It requires 4 software engineers, workstations, and funding for software/consulting. The goal is a universal interface providing access to CERN's stores of machine-readable information through a common protocol.
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
140 views

WorldWideWeb Proposal For A HyperText Project

This document proposes the WorldWideWeb project, which aims to create a hypertext system linking information across CERN. It would allow users to easily navigate between different types of information (reports, notes, databases, documentation, etc.) stored on various servers, through hypertext links. The project has two phases - first implementing basic browsers and analyzing information needs, then expanding to allow user-added content. It requires 4 software engineers, workstations, and funding for software/consulting. The goal is a universal interface providing access to CERN's stores of machine-readable information through a common protocol.
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

WorldWideWeb:ProposalforaHyperText

Project
To:
P.G.Innocenti/ECP,G.Kellner/ECP,D.O.Williams/CN
Cc:
R.Brun/CN,K.Gieselmann/ECP,R.Jones/ECP,T.Osborne/CN,P.Palazzi/ECP,N.
Pellow/CN,B.Pollermann/CN,E.M.Rimmer/ECP
From:
T.BernersLee/CN,R.Cailliau/ECP
Date:
12November1990
TheattacheddocumentdescribesinmoredetailaHypertextproject.
HyperTextisawaytolinkandaccessinformationofvariouskindsasawebofnodesinwhichtheuser
canbrowseatwill.Itprovidesasingleuserinterfacetolargeclassesofinformation(reports,notes,
databases,computerdocumentationandonlinehelp).Weproposeasimpleschemeincorporating
serversalreadyavailableatCERN.
Theprojecthastwophases:firstlywemakeuseofexistingsoftwareandhardwareaswellas
implementingsimplebrowsersfortheuser'sworkstations,basedonananalysisoftherequirementsfor
informationaccessneedsbyexperiments.Secondly,weextendtheapplicationareabyalsoallowing
theuserstoaddnewmaterial.
Phaseoneshouldtake3monthswiththefullmanpowercomplement,phasetwoafurther3months,but
thisphaseismoreopenended,andareviewofneedsandwisheswillbeincorporatedintoit.
Themanpowerrequiredis4softwareengineersandaprogrammer,(oneofwhichcouldbeaFellow).
Eachpersonworksonaspecificpart(eg.specificplatformsupport).
Eachpersonwillrequireastateoftheartworkstation,buttheremustbeoneofeachofthesupported
types.Thesewillcostfrom10to20keach,totalling50k.Inaddition,wewouldliketouse
commerciallyavailablesoftwareasmuchaspossible,andforeseeanexpenseof30kduring
developmentforoneuserlicences,visitstoexistinginstallationsandconsultancy.
Wewillassumethattheprojectcanrelyonsomecomputingsupportatnocost:developmentfilespace
onexistingdevelopmentsystems,installationandsystemmanagersupportfordaemonsoftware.
T.BernersLeeR.Cailliau

WorldWideWeb:
ProposalforaHyperTextProject
T.BernersLee/CN,R.Cailliau/ECP

Abstract:
HyperTextisawaytolinkandaccessinformationofvariouskindsasawebofnodesinwhichtheuser
canbrowseatwill.Potentially,HyperTextprovidesasingleuserinterfacetomanylargeclassesof
storedinformationsuchasreports,notes,databases,computerdocumentationandonlinesystems
help.Weproposetheimplementationofasimpleschemetoincorporateseveraldifferentserversof
machinestoredinformationalreadyavailableatCERN,includingananalysisoftherequirementsfor
informationaccessneedsbyexperiments.

Introduction
Thecurrentincompatibilitiesoftheplatformsandtoolsmakeitimpossibletoaccessexisting
informationthroughacommoninterface,leadingtowasteoftime,frustrationandobsoleteanswersto
simpledatalookup.Thereisapotentiallargebenefitfromtheintegrationofavarietyofsystemsina
waywhichallowsausertofollowlinkspointingfromonepieceofinformationtoanotherone.This
formingofawebofinformationnodesratherthanahierarchicaltreeoranorderedlististhebasic
conceptbehindHyperText.
AtCERN,avarietyofdataisalreadyavailable:reports,experimentdata,personneldata,electronic
mailaddresslists,computerdocumentation,experimentdocumentation,andmanyothersetsofdataare
spinningaroundoncomputerdiscscontinuously.Itishoweverimpossibleto"jump"fromonesetto
anotherinanautomaticway:onceyoufoundoutthatthenameofJoeBloggsislistedinanincomplete
descriptionofsomeonlinesoftware,itisnotstraightforwardtofindhiscurrentelectronicmail
address.Usually,youwillhavetouseadifferentlookupmethodonadifferentcomputerwitha
differentuserinterface.Onceyouhavelocatedinformation,itishardtokeepalinktoitortomakea
privatenoteaboutitthatyouwilllaterbeabletofindquickly.

Hypertextconcepts
Theprinciplesofhypertext,andtheirapplicabilitytotheCERNenvironment,arediscussedmorefully
in[1],aglossaryoftechnicaltermsisgivenin[2].Herewegiveashortpresentationofhypertext.
Aprogramwhichprovidesaccesstothehypertextworldwecallabrowser.Whenstartingahypertext
browseronyourworkstation,youwillfirstbepresentedwithahypertextpagewhichispersonaltoyou
:yourpersonalnotes,ifyoulike.Ahypertextpagehaspiecesoftextwhichrefertoothertexts.Such
referencesarehighlightedandcanbeselectedwithamouse(ondumbterminals,theywouldappearin
anumberedlistandselectionwouldbedonebyenteringanumber).Whenyouselectareference,the
browserpresentsyouwiththetextwhichisreferenced:youhavemadethebrowserfollowahypertext
link:
(seeFig.1:hypertextlinks).
Thattextitselfhaslinkstoothertextsandsoon.Infig.1,clickingontheGHIwouldtakeyoutothe
minutesofthatmeeting.ThereyouwouldgetinterestedinthediscussionoftheUPS,andclickonthe
highlightedwordUPStofindoutmoreaboutit.
Thetextsarelinkedtogetherinawaythatonecangofromoneconcepttoanothertofindthe
informationonewants.Thenetworkoflinksiscalledaweb.Thewebneednotbehierarchical,and
thereforeitisnotnecessaryto"climbupatree"allthewayagainbeforeyoucangodowntoadifferent
butrelatedsubject.Thewebisalsonotcomplete,sinceitishardtoimaginethatallthepossiblelinks
wouldbeputinbyauthors.Yetasmallnumberoflinksisusuallysufficientforgettingfromanywhere
toanywhereelseinasmallnumberofhops.
Thetextsareknownasnodes.Theprocessofproceedingfromnodetonodeiscallednavigation.
Nodesdonotneedtobeonthesamemachine:linksmaypointacrossmachineboundaries.Havinga
worldwidewebimpliessomesolutionsmustbefoundforproblemssuchasdifferentaccessprotocols
anddifferentnodecontentformats.Theseissuesareaddressedbyourproposal.
Nodescaninprinciplealsocontainnontextinformationsuchasdiagrams,pictures,sound,animation
etc.Thetermhypermediaissimplytheexpansionofthehypertextideatotheseothermedia.Where
facilitiesalreadyexist,weaimtoallowgraphicsinterchange,butinthisproject,weconcentrateonthe
universalreadershipfortext,ratherthanongraphics.

Applications

Theapplicationofauniversalhypertextsystem,onceinplace,willcovermanyareassuchasdocument
registration,onlinehelp,projectdocumentation,newsschemesandsoon.Itwouldbeinappropriate
forus(ratherthanthoseresponsible)tosuggestspecificareas,butexperimentonlinehelp,accelerator
onlinehelp,assistanceforcomputercenteroperators,andthedisseminationofinformationbycentral
servicessuchastheuserofficeandCNandECPdivisionsareobviouscandidates.WorldWideWeb(or
W3)intendstocaterfortheseservicesacrosstheHEPcommunity.

Scope:ObjectivesandnonObjectives
Theprojectwilloperateinacertainwelldefinedsubsetofthesubjectareaoftenassociatedwiththe
"Hypertext"tag.Itwillaim:
toprovideacommon(simple)protocolforrequestinghumanreadableinformationstoredata
remotesystem,usingnetworks
toprovideaprotocolwithinwhichinformationcanautomaticallybeexchangedinaformat
commontothesupplierandtheconsumer
toprovidesomemethodofreadingatleasttext(ifnotgraphics)usingalargeproportionofthe
computerscreensinuseatCERN
toprovideandmaintainatleastonecollectionofdocuments,intowhichusersmay(butarenot
boundto)puttheirdocuments.Thiscollectionwillincludemuchexistingdata.(Thisispartlyto
giveusfirsthandexperienceofuseofthesystem,andpartlybecausemembersoftheprojectwill
alreadyhavedocumentationforwhichtheyareresponsible)
toprovideakeywordsearchoption,inadditiontonavigationbyfollowingreferences,usingany
neworexistingindexes(suchastheCERNVMFINDindexes).Theresultofakeywordsearchis
simplyahypertextdocumentconsistingofalistofreferencestonodeswhichmatchthe
keywords.toallowprivateindividuallymanagedcollectionsofdocumentstobelinkedtothose
inothercollections.tousepublicdomainsoftwarewhereverpossible,orinterfacetoproprietary
systemswhichalreadyexist.
toprovidethesoftwarefortheabovefreeofchargetoanyone.

Theprojectwillnotaim
toprovideconversionswheretheydonotexistbetweenthemanydocumentstorageformatsat
CERN,althoughprovidingaframeworkintowhichsuchconversionutilitiescanfit
toforceuserstouseanyparticularwordprocessor,ormarkupformat
todoresearchintofancymultimediafacilitiessuchassoundandvideo
tousesophisticatednetworkauthorisationsystems.datawillbeeitherreadablebytheworld
(literally),orwillbereadableonlyononefilesystem,inwhichcasethefilesystem'sprotection
systemwillbeusedforprivacy.Allnetworktrafficwillbepublic.

RequirementsAnalysis
Inordertoensureresponsetorealneeds,arequirementsanalysisfortheinformationaccessneedsofa
largeCERNexperimentwillbeconductedattheverystart,inparallelwiththefirstprojectphase.
ThisanalysiswillatfirstbelimitedtotheactivitiesofthemembersoftheAlephexperiment,andlater
beextendedtoatleastoneotherexperiment.Anoverviewwillbemadeoftheinformationgeneration,
storageandretrieval,independentoftheform(machine,paper)andindependentofthefinality
(experiment,administration).
Theresultshouldbe:
listsofsources,depotsandsinksofinformation,
listsofformats,
diagramsofflow,
statisticsontraffic,

estimatedlevelsofimportanceofflows,
listsofclientdesiresand/orsuggestedimprovements,
estimatedlevelsofsatisfactionwithplatforms,
estimatedurgencyforimprovements.
Thisanalysiswillitselfnotproposesolutionsorimprovements,butitsresultswillguidetheproject.

Architecture
Thearchitectureofthehypertextworldisoneofdatastoredonservermachines,andclientprocesses
onthesameorothermachines.Themachinesarelinkedbysomenetwork(fig.2).Fig.2:proposed
modelforthehypertextworldAworkstationiseitheranindependentmachineinyourofficeora
terminalconnectedtoaclosebycomputer,andconnectedtothesamenetwork.Theserversareactive
processesthatreplytorequests.Thehypertextdataisexplicitlyaccessibletothem.Serverscanbe
manyonthesamecomputersystem,buttheneachcaterstoaspecifichypertextbase.Clientsare
browserprocesses,usuallybutnotnecessarilyonadifferentcomputersystem.Informationpassedisof
twokinds:nodesandlinks.

Buildingblocks
Browsersandserversarethetwobuildingblockstobeprovided.

Abrowser
isanativeapplicationprogramrunningontheclientmachine:
itperformsthedisplayofahypertextnodeusingtheclienthardware&softwareenvironment.
Forexample,aMacintoshbrowserwillusetheMacintoshinterfacelookandfeel.
itperformsthetraversaloflinks.Forexample,whenusingaMacintoshtobrowseonCERNVM
FINDitwillbetheMacintoshbrowserwhichrememberswhichlinksweretraversed,howtogo
backetc.,whereastheCERNVMserverjustrespondsbyhandingthebrowsernodes,andhasno
ideaofwhichnodestheuserhasvisited.
itperformsthenegotiationofformatsindialogwiththeserver.Forexample,abrowserfora
VT100typedisplaywillalwaysnegotiateASCIItextonly,whereasaMacintoshbrowsermight
beconstructedtoacceptPostScriptorSGML.

Aserver
isanativeapplicationprogramrunningontheservermachine:
itmanagesawebofnodesonthatmachine.
itnegotiatesthepresentationformatwiththebrowser,performingonthefly(orcached)
conversionsfromitsowninternalformat,ifany..

Operation
AlinkisspecifiedasanASCIIstringfromwhichthebrowsercandeduceasuitablemethodof
contactinganappropriateserver.Whenalinkisfollowed,thebrowseraddressestherequestforthe
nodetotheserver.Theserverthereforehasnothingtoknowaboutotherserversorotherwebsandcan
bekeptsimple.
Oncetheserverhaslocatedtherequestednode,itwillknowfromthenodecontentswhatthenode's
formatis(eg.pureASCII,markedup,wordprocessorstorageandwhichwordprocessoretc.).The
serverthenbeginsanegotiationwiththebrowser,inwhichtheydecidebetweenthemwhatformatis
acceptablefordisplayontheuser'sscreen.Thisnegotiationwillbebasedonlyonexistingconversion

programsandformats:itisnotinthescopeofW3towritenewconverters.Thelastresortinthe
negotiationisthebinarytransferofthenodecontentstoafileintheuser'sfilespace.Negotiatingthe
formatforpresentationisparticulartoW3.

Projectphases
Providedwithresourcesmentionedbelow,weforeseethefirsttwophasesoftheprojectasachieving
thefollowinggoals:

Phase1Target:3monthsfromstart
BrowsersondumbterminaltoopenreadershiptoanyonewithacomputerorPC.(?)
Browsersonvt220terminalstogivecursororientedreadershiptoaverylargeproportionof
readersAbrowserontheMacintoshintheMacintoshstyleAbrowserontheNeXTusingthe
NeXTSteptools,asafastprototypeforideasinhumaninterfacedesignandnavigation
techniques.
AserverprovidingaccesstotheworldofUsenet/Internetnewsarticles.*
AserverprovidingaccesstoalltheinformationcurrentlystoredonCERNVMandmentionedin
theFINDindex.ThisshouldincludeCERNprogramlibrarynotes,IBMandCERNCMShelp
screens,CERN/CNwriteups,ComputerNewsletterarticles,etc.
Aserverwhichmaybeinstalledonanymachinetoallowfilesonthatmachinetobeaccessedas
hypertext.
Theabilityforuserstowrite,usingmarkuptags,theirownhypertextforhelpfiles.Noother
hypertexteditingcapabilitywillnecessarilybeimplementedinthisphase.
AgatewayprocesstoallowaccessbetweentheInternetandDECnetprotocolworlds.
Asetofguidelinesonhowtomanageahypertextserver.
Arequirementsanalysisoftheinformationaccessneedsforalargeexperiment.
Atthisstage,readershipisuniversal,butthecreationofnewmaterialreliesonexistingsystems.For
example,theintroductionofnewmaterialfortheFINDindex,orthepostingofnewsarticleswilluse
thesameproceduresasatpresent.wegainusefulexperienceintherepresentationofexistingdatain
hypertextform,andinthetypesofnavigationalandotheraidsappreciatedbyusersinhighenergy
physics.

Phase2Target:6monthsfromstart
Inthisimportantphase,weaimtoallow
Thecreationofnewlinksandnewmaterialbyreaders.Atthisstage,authorshipbecomes
universal.
AfullscreenbrowseronVM/XAforthoseusingCERNVM,andotherHEPVMsites
AnXwindowbrowser/editor,givingthesophisticatedfacilitiesoriginallyprototypedunder
NeXTSteptothewideXbasedcommunity.(WeimagineusingOSF/Motifsubjectto
availability)
Theautomaticnotificationofareaderwhennewmaterialofinteresttohim/herhasbecome
available.Thisisessentialfornewsarticles,butisveryusefulforanyothermaterial.
Theabilityofreaderstocreatelinksallowsannotationbyusersofexistingdata,allowstoadd
themselvesandtheirdocumentstolists(mailinglists,indexes,etc).Itshouldbepossibleforusersto
linkpublicdocumentsto(forexample)bugreports,bugfixes,andotherdocumentswhichtheauthors
themselvesmightneverhaverealisedexisted.Thisphaseallowscollaborativeauthorship.Itprovidesa
placetoputanypieceofinformationsuchthatitcanlaterbefound.Makingiteasytochangetheweb
isthusthekeytoavoidingobsoleteinformation.Oneshouldbeabletotracethesourceofinformation,
tocircumventandthentorepairflawsintheweb.

Resourcesrequired
1.People
Thefollowingfunctionsareidentifiable.Theydonotnecessarilycorrespondtoindividualsonaoneto
onebasis.Theinitialsinbracketsindicatepeoplewhohavealreadyexpressedaninterestintheproject
andwhohavethenecessaryskillsbutdonotindicateanycommitmentasyetonthierpartorthepartof
theirmanagers.Weareofcourseveryopentoinvolvementfromothers.
Systemarchitect.Coordinatedevelopment,protocoldefinition,etcensuresintegrityofdesign.
(50%TBL?)Marketresearchandproductplanner.Discusstheprojectanditsfeatureswith
potentialandactualusersinalldivisions.Preparecriteriaforfeatureselectionanddevelopment
priority.(50%RC?)
HyperLibrarian.Overseesthewebofavailabledata,ensuringitscoherency.Interfacewithusers,
trainusers.Managesindexesandkeywordsystems.Managesdataprovidedbytheprojectitself.
(100%KG?)
Softwareengineer:NeXTStep.Providebrowser/editorinterfaceundertheNeXTStephuman
interfacetools.Experimentwithnavigationalaids.KeeparunningknowledgeoftheNeXTStep
world.(50%TBL?)
Softwareengineer:Xwindowsandhumaninterface.Providebrowser/editorhumaninterface
underOSF/Motif.Respondtousersuggestionforeaseofuseimprovementsandoptions.Create
anaesthetic,practicalhumaninterface.KeeparunningknowledgeoftheXworld.(75%RJ?)
Softwareengineer:IBMmainframe.ProvidebrowserserviceonCERNVMandotherHEPVM
sites.MaintaintheFINDserversoftware.KeepuparunningknowledgeoftheCMS,Rexx
world.(75%BP?)
Softwareengineer:Macintosh.Providebrowser/editorforthemac,usingwhatevertoolsare
appropriate(ThnkC,HyperCard,etc?).(50%RC?)
Softwareengineer:C.Helpwritecodefordumbterminalorvt100browsers,andportable
browsercodetobesharedbetweenbrowers.Thiscouldincludeatechnicalstudentproject.
(100%NP?+A.N.Other?)
Weforeseethatademandmayariseforbrowsersonspecificsystems,forspecificcustomizations,and
forserverstomakespecificexistingdataavailableonlineashypertext.Weintendtoenthusiastically
supportsuchwideningoftheweb.Ofcourse,wemayhavetodrawonmoremanpowerandspecific
expertiseinthesecases.

2.Otherresources
Wewillrequirethefollowingsupportinthewayofequipmentandservices.
Wefeelitisimportantforthoseinvolvedintheprojecttobeabletoworkclosetoeachotherand
exchangeideasandproblemsastheywork.Anofficeareaorclosegroupofofficesistherefore
required.
Eachpersonworkingontheprojectwillrequireastateoftheartworkstation.Experienceshows
thataworkstationhastobeupgradedinsomewayeverytwoyearsorsoassoftwarebecomes
morecumbersome,andmemory/speedrequirementsincrease.This,andthecostofsoftware
upgrades,weforeseeasreasonableexpenses.Weimagineusingavarietyoftypesofworkstation
asweprovidesoftwareonavarietyofmachines,butotherwiseNeXTs.ForVMSmachines,we
wouldlikethesupportofanexistingVAXclustertominimizeourownsystemmanagement
overheads.
Wewouldliketobeabletopurchaselicensesforcommercialhypertextsoftwarewherewefeel
thiscouldbeincorporatedintotheproject,andsavedevelopmentandmaintenancetime,or
wherewefeelwecouldgainusefulexperiencefromitsuse.(Approximateexamplesare:Guide
license:CHF750KMSfullauthorlicenseCHF1500,evaluationkitCHF100.FrameMaker:
CHF2000)

Wewillrequirecomputingsupport.Inparticular,wewillrequireareliablebackedupNFS(or
equivalent)fileserversupportforourdevelopmentenvironment.Wewillalsoneedtorun
daemonsoftwareonmachineswithInternet,DECnetandBITNETconnectivity,whichwill
requireacertainamountofsupportfromoperatorsandsystemmanagers.

Futurepaths
Thetwophasesabovewillprovideanextremelyusefulsetoftools.Thoughtheresultsseem
ambitious,theindividualstepsnecessaryarewellwithinourabilitieswithavailabletechnology.
Futuredevelopmentswhichwouldfurtherenhancetheprojectcouldinclude:
Daemonprogramswhichrunovernightandbuildindexesofavailableinformation.
Aserverautomaticallyprovidingahypertextviewofa(forexampleOracle)database,froma
descriptionofthedatabaseandadescription(forexampleinSQL)oftheviewrequired.
Workonefficientnetworkingoverwideareas,negotiationwithothersitestoprovidecompatible
onlineinformation.
Aseriousstudyoftheuseandabuseofthesystem,thesociologyofitsuseatCERN.

References
[1]
T.BernersLee/CN,HyperTextandCERN.Anexplanationofhypertext,andwhyitisimportant
forCERN.Abackgrounddocumentexplainingtheideasbehindthisproject.
[2]
T.BernersLee/CN,HypertextDesignIssues.Adetailedlookathypertextmodelsandfacilities,
withadiscussionofchoicestobemadeinchoosingorimplementingasystem.
[3]
Otherdocumentationontheprojectisstoredinhypertextformandwhichleadstofurther
references.

You might also like