REFACTORING SCOTT AMBLER

Scott Ambler CCed me an exchange with Ravi Venkataraman who initiated Scott Ambler and His Strawman to criticize an article by Ambler. An email exchange ensued in which I referred him to explanations of why I do not deem engaging him, as he thought I ought to do, productive:

Aside from the fact that Ambler is, of course, free to rebut any criticisms on his many platforms without my participation (as I do with respect to criticisms of me), here are some further arguments in support of that position.

To Ravi’s claim in the above-mentioned group exchange that he does not know data fundamentals, Ambler responded as follows:

Really? Because I've written about both normalization as well as denormalization as well as several times in print (my book AGILE DATABASE TECHNIQUES, won a Jolt productivity award). I've also written extensively about data modeling with the UML (interestingly, when you Google the term "data modeling" my article is the first one to come up, so apparently there's a lot of people out there who think that what I have to say about databases is pretty good).

First, most of what is being published on database management is either regurgitation of tool manuals, jargon generation, or cookbooks of trips and tricks aimed at relieving practitioners from thinking too much. And given the reasoning and critical thinking ability of the average practitioner, books sell precisely because that’s what they are. An almost exclusive focus on tools also means that vendors like, often sponsor, and promote them. Worse, even academia and scientific journals are lowering themselves to the industry level (see, for example, Et Tu Academia? and Chasing Dayflies: Reply to Riggs). Serious books or articles on data fundamentals don’t have a chance.

Second, due to the disregard for foundation knowledge and reason—caused by educational failure—most practitioners are Unskilled and Unaware of It:instead of judging reputation by content, they judge content by reputation. At his limited knowledge level the average practitioner often believes (and, as I have documented, states) that “a published author must be an expert”; there is a general lack of awareness that anybody can and does write these days, whether they are technically or logically accurate or not, be it books e.g. Basic Concepts in UML: A Request for Clarification (Ambler is a proponent of OO and UML), or articles e.g. Joins Explained). So Invoking authorship does not, per se, mean much.

As an aside, when Ravi asked “As for getting a Jolt award, isn't it true that books submitted for

these awards have to pay some money to be considered?”, Ambler responded

In recent years the publisher needs to pay a $200 entry fee. The fee helps to focus the publishers, and software vendors for that matter, to nominate only products which are good enough to have a chance.

It may well focus them, but if you believe that “good enough to have a chance” means technical and pedagogical merit, I’ve got a bridge in Brooklyn to sell you.

· Who exactly are those who judge the books? Are they knowledgeable, competent database professionals? How were they selected and by whom?

· What exactly are the technical judgment criteria and how objectively are they applied?

· Are these details publicly available prior to submissions and judgment?

The fact is that neither publishers, nor award conferrers base decisions on the quality of content, or technical merit. Most of these awards are based at best on some measure of popularity/sales (if anything), which is significantly affected by marketing and promotion efforts, which in turn affects the decision to pay the submission fee (in some cases there are blatant conflicts of interests, as when the awarding organization is, for example, a magazine whose writers are authors of submitted books).

Thus, belaboring under any notion that popularity or sales reflect content quality is either an illusion or delusion. That is why I insist of warning myself constantly that it is when I become popular that I will have to start worrying what I am doing wrong. If the old adage that quality correlates negatively with quantity applies anywhere, it certainly does in the database management field, which makes those who want to take advantage of it quite profitable.

In his emails Ambler also boasted about interest by vendors to implement some of his ideas, which he interprets as an indication that they are “fairly good” because vendors “are motivated to compete effectively”.

Economics 101: To assess how good an idea is and to implement and sell it on merit, four critical things are needed

(a) informed idea providers

(b) informed implementers of the idea

(d) informed media and experts to aid consumers in their assessment

None of these exist in the database field. In a blind-leading-the-blind system the market will not operate efficiently, as per Ambler self-servingly claims. Not only does the industry, with help from its various uncritical trade media and “experts”, operate like the fashion industry, but just like the latter, it actually comes up with “retrospective lines”. The industry keeps reinventing and re-labeling old discarded ideas as “new paradigms” and gets away with it simply because there is no knowledge or appreciations of the history and fundamentals of the field. Ambler, for example, promotes heavily OO jargon, and calls true data professionals (including me) “dinosaurs”. Consider, therefore, the following comment to him in Responsibility-Driven Design versus Functional Decomposition:

Tim: By the way, OO came into being in the 1960s with the creation of Smalltalk and Simula (preceding it, I think), relational technology came into being in the 1970s with E F Codd's seminal paper. Hence relational technology is newer than OO technology. Therefore, anybody who refers to database experts as "data dinosaurs" should probably refer to OO practitioners, who use even older technologies, as those belonging to a Pre-Cambrian era! Let's see how it sounds: Pre-Cambrian OO practitioners. Has a nice erudite ring to it.

Witness the “revolutionary” ODBMS and XML—throwbacks to network and hierarchic DBMSs that were made obsolete decades ago. In fact, Ambler’s own material suggests a throwback to application-based data management which databases and DBMSs were devised to eliminate. Labels and catchwords are not substitutes for foundation knowledge, even if they do attract an audience that cannot, due to lack of proper education, grasp beyond marketing gimmicks.

Ambler also accused me of being a charlatan because I don’t provide URLs to the material I debunk to allow readers to make their own judgments. This is not entirely accurate—this policy holds only for weekly quotes and I do post links to online exchanges—but I did make my reasons public. Popularity is measured on the Internet by traffic, and it would be inconsistent of me to debunk nonsense, yet contribute to its popularity ranking (but I do keep URLs and can provide them upon request). And Ambler’s accusation is rather rich coming from somebody who, invokes that very popularity as evidence of expertise, and who condones locking out of online forums those who challenge him in ways that he cannot address (see

On Straw Man Tactics in Software, forthcoming). But even though I will not publish (as I’m sure he would like) his URLs—readers can find them if they care to—this editorial does provide two links to group discussions in which he participates, so that readers can judge for themselves who is the real charlatan.

It’s the industry that needs refactoring, if databases are to be refactored the way they should, and not the way Ambler promotes.

Revised 5/5/06

Posted 4/7/06