If you subscribe to the Kimball Design Tips, you’re likely already well aware of Ralph Kimball’s contributions to the field of data warehousing and business intelligence. However, many of you may not know about Ralph’s contributions and accomplishments prior to turning his attention to our industry. Ralph’s not someone who seeks the limelight or frequently toots his own horn; I need to do that for him sometimes. I recently shared this historical perspective with a client who insisted it be disseminated more broadly.
Ralph’s official title is Dr. Kimball. He received his PhD in electrical engineering from Stanford University in 1973 after building a computer-assisted tutor that taught calculus concepts to students, but at the same time learned superior problem solving techniques from the students themselves. Shortly after earning his doctorate, Ralph went to work at the Xerox Palo Alto Research Center (PARC). PARC was a research think tank that attracted the best and the brightest engineers and computer scientists in the early 1970s. While PARC was never a household word, its pioneering innovations still impact us daily – client-server computing, the Ethernet, laser printing, bitmapped graphical user interfaces (GUI) including windows and icons, the mouse, object-oriented programming, and the list goes on. The PARC alumni roster reads like a veritable who’s who of early Silicon Valley leaders.
While at PARC, Ralph spent five years doing research on user interface design, and then joined a group dedicated to building products based on the research prototypes demonstrated at PARC. Contrary to most product development practices at the time, a guiding principle for Ralph and his colleagues was to focus first on the user experience, then back into the design of the underlying hardware and software. They tried to follow the advice of their PARC colleague, Alan Kay: “the simple things should be simple; complex things should be possible.”
With the early adoption of some PARC innovations in Xerox’s experimental Alto workstation, Ralph’s team developed the first commercial product that incorporated a bitmapped display for a graphical user interface including icons and windows, Ethernet networking, file and printer servers, and a mouse. Their user interface mimicked the office paradigm with a desktop and icons depicting documents, folders, and email so it was intuitive to users. Xerox introduced the Star workstation in 1981, the same year that IBM introduced PC-DOS. Here’s a link to a snippet from an early Xerox promotional video for the Star workstation: http://www.youtube.com/watch?v=zVw86emu-K0. Ralph was the product manager of the Star workstation; he made his acting debut about 7 minutes from the start of the video.
Steve Jobs and several other Apple engineers saw a demonstration of Xerox’s Alto computer in late 1979, and Ralph gave Steve Jobs a demo of the Star workstation during its launch announcement in 1981. Needless to say, Steve grasped the commercial potential of PARC’s mouse-driven GUI; the concepts were subsequently incorporated in the Apple Lisa which was released in 1983 followed by the Macintosh in 1984.
In 1982, Ralph and several other Xerox colleagues envisioned leveraging the Star workstation beyond its traditional document creation market to tackle the world of decision support. But Xerox senior management decided not to pursue that opportunity. So in the fall of 1982, Ralph and several research colleagues and business leaders left Xerox to launch a company called Metaphor Computer Systems to focus on the data access and analysis needs of business professionals. During the product research phase, Ralph conducted interviews with Fortune 500 analysts to better understand their requirements. When speaking with folks in the consumer package goods industry, Ralph was introduced to the syndicated data from A. C. Nielsen, IRI and SAMI that was used for competitive analysis.
Ralph spent time with the experts at A. C. Nielsen who described their data in terms of dimensions and metrics. While Ralph didn’t invent dimensional modeling, he was quick to envision the broader applicability of the concepts for decision support across a variety of industries and functional application areas.
Metaphor released an integrated hardware and software product in 1984 with file servers, database servers, and workstations (including unattached keyboards and mice), linked together by the Ethernet. The workstations delivered a graphical user interface with query, spreadsheet, and plotting tools, along with a Capsule tool that Ralph invented. The Capsule, originally called “graphical pipes,” was a programming tool which connected the separate icons on the desktop with arrows that directed the flow of data from icon to icon. The Capsule could export the results of a query into a spreadsheet, then into a graph, and finally to an email outbox or a printer. We take these capabilities for granted today, but it was rocket science nearly 30 years ago.
True confessions… I saw a Metaphor demonstration in 1984 just prior to the product’s release and was so blown away that I immediately joined Metaphor’s initial class of field consultants. Ralph taught star schema design to my fellow new hires and me. The mantra from his formative years at PARC continued to resonate: focus on the business users’ experience by making the data model as understandable as possible, then work backwards through the design and development. Fortunately, old habits sometimes die hard.