ISSN1361-6161ComputerScienceUniversityofManchesterNotesonP-Algebra(4):AlgebraoverProcessStructure,PartI.KoheiHondaDepartmentofComputerScienceUniversityofManchesterTechnicalReportSeriesUMCS-95-12-5NotesonP-Algebra(4):AlgebraoverProcessStructure,PartI. KoheiHondaDepartmentofComputerScienceUniversityofManchesterOxfordRoad,Manchester,UK.khonda@cs.man.ac.ukDecember17,1995Copyrightc 1995.Allrightsreserved.Reproductionofallorpartofthisworkispermittedforeducationalorresearchpurposesonconditionthat(1)thiscopyrightnoticeisincluded,(2)properattributiontotheauthororauthorsismadeand(3)nocommercialgainisinvolved.RecenttechnicalreportsissuedbytheDepartmentofComputerScience,ManchesterUniversity,areavailablebyanonymousftpfromftp.cs.man.ac.ukinthedirectorypub/TR.The lesarestoredasPostScript,incompressedform,withthereportnumberas lename.Theycanalsobeobtainedon Refereedby:Cli JonesAbstractWedeveloptheoryofalgebraoverprocesses,basedonanabstracttreatmentofprocessstructure.Theabstractframeworkofprocessstructuresisdeveloped(PartI),and,onitsbasis,theoryofalgebraisdevelopedwherethebasicresultsincludingBirkho -likeVarietyTheoremareproved(PartII).Wealsodiscusspropertiesofsigni cantconcreteexamplesofalgebras,takenfromextanttheoriesofconcurrency.PartIdevelopstheabstracttheoryofprocessstructureencompassingawiderangeofstructuresforprocesstheoryincludingconcretestructureswehavediscussedintheprecedingnotes.ThetheoryisusedasfoundationsofthealgebraicdevelopmentinPartII,andisinterestinginitsownright.1IntroductionTheaimofthepresentnotesistwo-fold.The rstaimistodevelopanotionofalgebraoverprocessstructurestogetherwithanappropriatesystemofnotations.Theabstractframeworkofthetheoryofalgebrafairlyfollowstheusualcategoricaltreatmentof\alge-brainacategorywiththecategoryofprocessstructureandp-mapsastheunderlyingcategory.Thenewpointshoweverlieintheconcreterepresentationofgeneralnotionslikehomomorphismsinthesettingofprocessstructure,aswellasbasicdi erencesinseveralconstructionssuchasquotientandotherconstructions.Suchdi erencesneces-sitateanewnotationforpresentingalgebraicnotions.Thetheory,whichtreatsstaticaspectofmanipulationofpureprocesses,willbecomethebasisoftreatmentofprocessdynamics,tobedevelopedinthesubsequentnotes.Thesecondaimconcernstheframeworkinwhichthesenotionsandresultsarepre-sented.WedeveloptheoryofalgebrainanabstractframeworkunlikeintheprecedingNotes[7,8,9].Thus,ratherthanrelyingonconcreteconstructionofpermutationgroupsandpartialinjectionsover nitesets,webaseourdevelopmentonapairofcategoriesconformingtocertainconditions,bywhichallessentialresultsandconstructionsoftheprecedingnotesareguaranteed.Forthedevelopmentoftheoryofalgebra,thisresultsinclarityofpresentationofthenotionsandtheproofs.Theapproachisalsoimpor-tantwhenweneedtoextendthenotionofprocessesbeyondwhathasbeenstudiedintheprecedingstudyofconcurrencysofar,cf.Examples2.19.Ourprimarypurposeinthispaperhoweverliesinexploitationoftheabstractframeworkforlucidpresentation,leavingdeeperstudyofsuchpossibilityelsewhere.Thenotesaredividedintotwoparts.PartIgivesabstracttheoryofprocessstructure,whichwillbeusedinPartIIandalsointhesubsequentNotes.Thetreatmentofthetheoryisself-contained.PartIIisdevotedtothedevelopmentofelementarytheoryofalgebra.IntherestofPartI,Section2presentsthefoundationalstructuresoftheab-stracttheorycalledConnectionDomains,whichgivesarelationalalgebraofconnectionsamongprocessesinacategoricalframework.Section3thenstudiesanabstracttheoryofprocessstructurebasedonconnectiondomains.Byvaryingtheunderlyingconnectiondomain,quitedi erentkindsof\processstructuresarise(includingthefamiliarnotionofsetsthemselves),eachofwhichareshowntoinduceaset-liketheoryaswedeveloped1intheprecedingNotes.AllthemainresultsofNotes1and3areestablishedinthisabstractsetting,purelyrelyingontheabstractalgebraofconnections.Surprisingly,thegeneralityresultsinmuchshorterandmorelucidproofsthanthoseoftheprecedingNotes.Acknoweldgemensts.TheauthorthanksCli Jonesforaseriesofenlighteningdis-cussionsandencouragement.HethanksPeterAczelforstimulatingdiscussionswhichdeepenedhisunderstandingonthesubject.TheauthorgratefullyacknowledgesthesupportofanEPSRCVisitingFellowship.2ConnectionDomainInfunctionsorrelationsoveraset,someelementsofasetarerelatedtosomeelementsofaset.Thewaytorelateanelementtoanotherelementisquitesimple:wejusttakeatupleoftwoelements.Inthetheoryofprocesses,thissimpleschemeisnolongervalid.Twoprocesses(correspondingtotwoelements)mayberelatedindiverseways,consideringhowinterfacepointsofprocessesareconnectedtoeachother.Thiswasimplicitinthestudyofprocessalgebra[11,6],embodiedinthenotionoffreenames,orstudyofcompositaanditsgeneralisations[2,3,13],embodiedinthenotionof\variables(of,say,openterms),andismadeexplicitbythestudyofprocessstructure[7],aswellasbysuchworksas[1,5,12].Speci callywhat[7]showedisthatacoherenttheoryofmapsandrel