Charles W. Bachman interview: September 25-26, 2004; Tucson, Arizona

Charles W. Bachman reviews his career. Born during 1924 inKansas, Bachman attended high school in East Lansing, Michiganbefore joining the Army Anti Aircraft Artillery Corp, with which hespent two years in the Southwest Pacific Theater, during World WarII. After his discharge from the military, Bachman earned a B.Sc.in Mechanical Engineering in 1948, followed immediately by an M.Sc.in the same discipline, from the University of Pennsylvania. Ongraduation, he went to work for Dow Chemical in Midland, Michigan.Bachman discusses his experiences as an engineer there, focusing onengineering economics issues, including his work on the developmentof methods for return-on-investment calculations, using punchedcard machines. In 1957, he was appointed the first head of DowsCentral Data Processing department, where he was responsible forpreparing the way for the arrival of its first big digitalcomputer. Bachman organized a feasibility study to select a newmachine and hired a staff of programmers and analysts. Through theIBM SHARE user group and its 9PAC project, he became involved withHarry Tellier, and Russ McGee in the design of programs to simplythe file maintenance and report generation process. This was basedon the earlier GE Hanford 702 File Maintenance and Report GeneratorSystem, which Bachman discusses in some detail. He also worked onthe theory of information with Bill Orchard-Hays and Tom Steel.However, Dow canceled its IBM 709 order before the machine arrived,ending Bachmans association with SHARE and, shortly afterwards,with Dow Chemical. In 1961, Bachman went to work for General Electric in New YorkCity. He worked for a corporate staff group providing internalconsulting services, and was assigned to an integrated systemsproject to produce a generic manufacturing control system for GEdepartments, using the firms new GE 225 computer. Here he workedwith Stan Williams and Bill Helgeson. This system, MIACS, an earlytransaction processing system, was built around Bachmans newcreation: the IDS (Integrated Data Store). IDS is called the firstdatabase management system, and Bachman explains its origin,development, capabilities, design and relationship to MIACS indetail. IDS used a virtual memory database, with a DDL and DML, andintroduced the network data model. Bachman created the first datastructure diagrams to illustrate the complex data structuresrequired to picture the adaptive manufacturing control logic ofMIACS. In 1964, Bachman transferred to GEs Computer Department inArizona, where he worked on a number of other database relatedprojects including the GE 400 IDS, GE 600 IDS, DataBASIC, apersonal data storage system, and WEYCOS 1 and 2, both built forWeyerhaeuser Lumber. The latter two were an attempt to build acomplex online management information system around a database.Bachman believes WEYCOS 2 produced the first database managementsystem able to support multiple application programs executingsimultaneously against the same database. Bachman also describeshis work with Warren Simmons, Bill Olle and others on the CODASYLData Base Task Group during the late 1960s. The database standardsproduced by this group were heavily influenced by IDS and byBachmans ideas. Bachman also discusses the links between IDS andCullinanes IDMS, a successful commercial system, built to run onthe IBM 360 series of computers, which was based on IDS. When Honeywell acquired GEs computer business in 1970, Bachmanworked for the merged operation in Honeywells advanced researchgroup, in Boston. He continued to work on database topics,investigating what he called the role data model and serving on anANSI-SPARC committee intended to standardize database systems.While failing in this objective, this committee made a highlyinfluential distinction between external, conceptual, and internal(database) schemata. In 1973, Bachman received the ACM Turingaward. He recalls his reactions to this, and his subsequent serviceon committee to select future winners. Bachman was elected aDistinguished Fellow of the British Computer Society in 1977.Bachman also discusses his thoughts on the relational data modeland its creator Ted Codd, including a well known 1974 meeting inwhich he debated Codd on the merits of the two approaches. Bachmanalso chaired the ISO committee working on Open SystemsInterconnection (OSI), well known for its elaboration of a sevenlayer model for computer communication. In 1981, Bachman went to work for Cullinane Database Systems, amajor software company built around IDMS. While at Cullinane, hesupported marketing, and continued the ISO/OSI committee work,while working on tools to model high level schemata for enterprisedatabase design and translate them into data definitions. He alsoworked on the partnership set data model, which was granted a USpatent, and used as the basis for a enhanced network data model andconceptual schema modeling. In 1983, he left to pursue this work athis own firm, Bachman Information Systems. He describes the riseand fall of this company, which obtained venture capital funding,grew rapidly and made a successful IPO. Despite enjoying successwith its Data Analyst product with its reverse and forwardengineering capabilities, the firm struggled financial andmanagerially, including an initially successful, but ill-fated,spell as an IBM Business Partner. It eventually merged with CadreSystems to create Cayenne Software, which was acquired by SterlingSoftware in 1998, which was subsequently acquired by ComputerAssociates. Since then Bachman has worked as a database consultantto Constellar Systems and in 2006 was still working with CbrSystems.