In the course of his polemic, Dan is kind enough to say that I am a friend - I hope that is still the case when he is done reading this. He adds that I have provided useful background on dynamic data warehousing in a related publication.2 Thus, my esteem for his good judgment is high, his rant is an excellent one, and it deserves a considered reply.
Dan's immediate complaint is that Claudia Imhoff, Bill Inmon and himself have received insufficient recognition for their work on dynamic data warehousing. It is a matter of public record that Claudia led a multi-city road tour with IBM on dynamic data warehousing, so presumably she was able to clarify to the market in person her fundamental role in validating this concept as used by IBM.3 Bill's startup enterprise, Inmon Data Systems, has been extending the technology of ETL, of which he was the inventor, into the realm of unstructured data.4 his was presented in a public forum in Chicago last year at which both IBM (as the sponsor) and end-user enterprises were in attendance. So if Bill will pardon the expression, he was "grandfathered in" as IBM was in the process of getting out the word.
If the rant was therapeutic for Dan, then it is well and good. However, inaccurate statements still require correction. At least half of Dan's statement is accurate - but which half? He says "I applaud IBM for beginning the effort to meet the needs of DDW, but I don't think they should be claiming credit for being the first to create a solution."
The latter part is simply not accurate. No such claim was made nor will any be found in the brochure in question in which IBM makes many claims such as: "Dynamic warehousing is IBM's new strategy to help organizations squeeze the greatest possible business value out of their information. ... IBM believes that dynamic warehousing is the next logical step in the evolution of data warehousing. ... IBM is the only company with end-to-end capabilities - that are proven to work seamlessly together - to implement dynamic warehousing."5
No doubt each of these statements, especially the last one, can be discussed. I will not do so here. Rather let me make a bold statement of the obvious. With a high concept such as "the grid," "dynamic data warehousing, "predictive analytics," it is hard to say who first thought of it. The first reference to "dynamic" is the philosopher Aristotle in his Physics. The first reference to the use of a database was when Tiberius Caesar conducted a census of his empire including a couple named Joseph and Mary. The first occurrence of the term often means something specific in a research context. Dan provides an example of DDW from the year 2001 in the context of XML. Other examples refer to time series databases, since "dynamic" invokes a sense of changing over time. Since this is itself a dynamic - that is, changing as opposed to static - context in which innovations are extending what is possible with data warehousing system, the goal is to stipulate a definition that 1) captures existing capabilities of products, tools and solutions; 2) indicates the method or approach to implementing the data warehouse; 3) allows for future innovations that are useful and incremental, and protect the client's investment in existing infrastructure. At this point, a kind of semantic contagion takes over in the (relatively) positive sense of viral marketing. The buzz in the market increases thanks also to rants such as Dan's incorporating as it does firm facts, dangerous half truths and total nonsense.
Dan and the colleagues at the Colorado Business Intelligence Community Connection (COBICC) are doing excellent work on dynamic data warehousing. So are the colleagues in the IBM laboratory system. Both are contributing, and we don't have to choose between them. What I would say to Dan is: do not let perfection be the enemy of the good. DDW is indeed for real. Is it perfect? Is it the silver bullet? Wouldn't that be nice?
What is incomplete gets improved incrementally in a step-by-step process until the capabilities become a best practice and, in turn, are driven into the DNA of the enterprise as confirmed, inheritable abilities. Insofar as Dan's rant is a sustained attack on an IBM "marketing brochure," this is a soft target. To correct a marketing brochure by means of a research agenda from 2001 or 2007 is - how shall I put it - to mistake innovation for a message about innovation. Marketing is about messaging. Research is about redefining the limits of what is possible by means of incremental and disruptive technological advances. In my opinion, Dan's polemic is itself a well-developed piece of marketing disguised as a rant. The brand? Dan Linstedt. It's not easy being dynamic. Of course, IBM is accountable for working to popularize the concept of dynamic data warehousing, including promoting a public debate, and to bask - not unlike Dan himself - in the halo effect thrown off by the viral, virtual brand.
Another inaccurate statement is "dynamic data warehousing is not yet here." Of course, if you make the definition too tight, no one will ever implement it. If you make it too loose, then every Access database will be "dynamic." If the reader considers the four principle components of DDW called out in IBM's definition as necessary and sufficient conditions, which is how I read them, then leading-edge enterprises are indeed operating DDW in production. The four components are 1) real-time access to information in context update - this does not refer specifically to accessing the most recent data, but just as much to accessing data as part of real-time business processes (this is the part that most mirrors "active data warehousing;" 2) analytics as a part of the business process, which means going beyond just delivering information, the crux of "active" data warehousing, to generating business insights in a timely manner so that it can make a difference in the day-to-day operations, including at a tactical level; 3) integration of unstructured data into the process of analytics and decision-making; 4) extended infrastructure tightly integrated, which I understand to include the entire decision support ecosystem with high performance inbound processing, message brokering, information integration, master data management and efficient, timely information delivery (reporting and access), not just the atomic data warehouse
Be the first to comment on this post using the section below.