Update: In response to Monash’s spewing further nonsense and insidious misinterpretations of Date’s and my comments on his vociferous ignorance (see my forthcoming column at www.tdan.com), Date comments as follows:

I blame myself. I should have known better than to stoop to bandying words with a chronophage like Monash. Life is too short. Monash can say or think what he likes about me; I prefer to let my record speak for itself.

</response>

(I hope I’m using those XML tags properly!)

Posted 2/24/06

Ed. Comment: This is Date’s reply to Curt Monash’s comments in the online Computerworld. It is almost impossible to get Chris Date to call a spade a spade, no matter how much I tried—may have something to do with his British roots—he came closest to it in this piece, but it is still much too bland for my taste, and certainly does not put Monash in the place he really deserves. But I don’t have any British compunctions and, in fact, I believe it is irresponsible not to call a spade precisely a spade, so check out my own reply, Monash Balderdash.

PREAMBLE

A couple of recent threads on the Web^* include a number of insulting remarks aimed directly at myself. The remarks in question are by one Curt Monash—or, rather, " Curt A. Monash, Ph.D.," as he refers to himself (and in one of the threads he says his Ph.D. is "in mathematics" and was obtained from "an elite university"). Now, I have little idea as to who Curt Monash is, nor do I really wish to know (or even care). I'm sure we've never met. We've certainly never corresponded. And I don't think I've ever commented on anything of his in any of my previous writings. So I don't know who or what I have to thank for his recent attentions and apparent animosity toward myself, but I certainly do want to respond to some of the more egregious of his remarks.

The next section consists of Monash's remarks, unedited by me except for a couple of explanatory notes (also, I've numbered his remarks for reference purposes). The subsequent section examines the remarks one by one.

MONASH'S REMARKS

1. From The DRIPping folly of the " Information Principle" (title of a posting): A couple of annoying relational theory-purists have been hurling anathemas [at me] ... They are particularly hung up on Date's Relational Information Principle (DRIP) ... It's a nice theory, and Chris Date has surely done a great service educating people to some good ideas in database design. But when his acolytes start trying to interfere in practical discussions of commercial technology use, the DRIPping can get hard to take.

2. Later (after Fabian Pascal pointed out to Monash that it was actually Ted Codd, not me, who came up with The Information Principle): Wrong again ... Date has been quoted saying that while the idea of the Information Principle was Codd's, the phrasing is Date's.

3. Fabian: Can you by any chance justify your and Chris Date's claims to be basing your views on "science"? I could understand why you might claim to base them on math and logic, but your claims to base them on "science" seems about as credible as Tom Cruise's blatherings on psychiatry.

4. If you want to argue that logic and math are sciences ... you can wait until I get around to more carefully debunking [Pascal's] and Date's outrageous intellectual fakery on the subject directly, and respond at that time ... Mathematics = studying abstract systems which follow various sets of assumptions (axioms) ... Science = studying of [sic] the real world ... Engineering = changing the real world ...

5. It seems that whenever I look at an argument by Chris Date or Fabian or [Alf Pedersen], it quickly gets to the point where you [sic] say "I'm not talking about how to use actual products that are shipping today. I'm talking about how to use products that I hope will ship in the future" ... [At] least [Alf is] not repeating the "science" nonsense of Chris Date and Fabian Pascal.

6. With reference to my O'Reilly interview: Chris Date is BADLY confused in that discussion of geometry. HE DOESN'T KNOW WHAT AN "AXIOM" IS!!!!!!!!!!!!!!!!!!!!!!!!!! Or else he DOES know, and is deliberately debasing the discourse for his own personal advantage. Either alternative is sickening. I want to throw up.

7. I STRONGLY reject the argument "relational theory is based on math/logic/science, and hence must be the best guide to practical database management."

SOME RESPONSES

Before getting into details, I'd like to say that Monash's remarks throughout the threads under discussion—not just the remarks I've actually quoted, but most of the others as well—display a quite extraordinary lack of clear and careful thinking. As I've stated elsewhere,^* it's very distressing to find such sloppiness in connection with relational technology of all things, given that one of Codd's objectives in introducing the relational model was precisely to inject some sorely needed precision and clarity of thinking into the database field. Anyway, let me get down to some specifics. Here again is the first quote:

----------

^*In Relational Database: Further Misconceptions Number Three, in C. J. Date and Hugh Darwen, RELATIONAL DATABASE WRITINGS 1989-1991.

----------

1. A couple of annoying relational theory-purists have been hurling anathemas [at me] ... They are particularly hung up on Date's Relational Information Principle (DRIP) ... It's a nice theory, and Chris Date has surely done a great service educating people to some good ideas in database design. But when his acolytes start trying to interfere in practical discussions of commercial technology use, the DRIPping can get hard to take.

As Fabian Pascal quickly pointed out in the original exchange, The Information Principle is not mine but Codd's (in fact, I heard Ted Codd refer to it on more than one occasion as the fundamental principle underlying the relational model). It's obvious that Monash attributes it to me purely in order to be able to get in a few cheap cracks based on his stupid DRIP acronym. Of course, his game-playing in this regard is hardly very important in the larger scheme of things; indeed, I wouldn't even mention it, were it not for the fact that it's so indicative of the level at which Monash appears to think a discussion of this serious subject ought to be conducted.

"A couple of annoying relational theory-purists": I don't know what a "relational theory-purist" is. First of all, I assume that Monash really meant "relational-theory purists," not "relational theory-purists." Of course, I'm well aware that "if you take hyphens seriously, you will surely go mad,"^* but there is a difference in meaning between the two versions of the phrase—and in our field above all, accuracy is surely paramount. Without it, we're lost.

----------

^*From "an old style guide of the Oxford University Press in New York," quoted by Lynne Truss in Eats Shoots & Leaves. New York, NY: Gotham Books (2004).

----------

Be that as it may, I further don't know what a "theory purist" (of any kind) can possibly be. Either you believe a given theory or you don't. If you do, is that what makes you a "purist"? And if you don't, what does that make you? An impurist? In any case (to adopt Monash's own style for a moment), I STRONGLY reject the suggestion that I might be a purist of any kind. We must be pure, not purist.

Is it possible to be a nonannoying "purist"?

By the way: Neither of the "annoying relational theory-purists" was me (I mean, I wasn't a participant in the threads under discussion). So I still don't know what I did to deserve Monash's contumely and abuse.

"They are particularly hung up on [The Information Principle]": Monash doesn't seem to understand the crucial point about The Information Principle, which in essence is that relations are the only data structure we need at the logical level—they're both necessary and sufficient to represent absolutely any data we want to represent (again, at the logical level), and so we can get away with the absolute minimum of complication for the user: more specifically, we can get away with the minimum number of object types (one) and the minimum number of operators.^* Add any extra object types, say XML hierarchies, and you fundamentally need a whole set of extra operators (not to mention the fact that it's not at all clear that an appropriate abstract model has yet even been defined for XML, or the fact that any such model, if it is defined, will necessarily be more complex than the relational model). All of which, incidentally, makes life more complicated not just for the user but also for the DBMS itself—not to mention DBMS implementers, educators, designers, application programmers, DBAs, database standardizers, and on and on.

----------

^*The operators are what we need for expressing queries, updates, constraints, view definitions, etc.

----------

"It's a nice theory": It's not completely clear from the original, but I think "it" here refers to The Information Principle. If so, see my response to Monash's remark number 1. But do I detect a faintly derisive tone here? As I said in the O'Reilly interview mentioned earlier, it's a very unfortunate fact that the term "theory" has two quite different meanings. In common parlance it's almost a pejorative term—"oh, that's just your theory." Indeed, in such contexts it's effectively just a synonym for opinion (and the adverb merely—it's merely your opinion—is often implied, too). But the meaning in scientific circles is quite different, as I'm sure Monash knows very well. The net is, I regard his "it's a nice theory" as a cheap shot, unworthy of a debater who wants to be taken seriously.

"Chris Date has surely done a great service educating people to some good ideas in database design": Well, gee, thanks, I suppose. But "some good ideas in database design"? To me, that phrase means, basically, just the principles of normalization ... and I'd certainly like to think that my contributions have included rather more than just teaching people what third normal form is. Is Monash merely damning me with faint praise? If so, I take it as another cheap shot.

"But when his acolytes start trying to interfere in practical discussions of commercial technology use, the DRIPping can get hard to take": I am SO tired of the all too ubiquitous suggestion that relational theory isn't practical! I could get very eloquent on this issue, but I'll limit myself here to just one observation: The entirety of today's "commercial technology use" that's based on SQL DBMSs—valued, conservatively, at several billions of dollars a year—can be regarded as a practical application (albeit flawed) of one great theoretical idea. What price theory?

Now, I don't want to be misunderstood here. I've never said (and nor has anyone else that really understood relational technology ever said, I'm quite sure) that relational technology is the solution to all of the world's problems. But it is the basis for the solution to far more of the world's problems than people like Monash even begin to comprehend, at least if their public pronouncements are anything to go by. What's more, I do think it's incumbent on anyone who proposes some replacement technology for attacking some specific problem to show first that relational technology can't solve the problem in question. I also think that, given that relational technology is founded on set theory and logic (elements of which go back over 2,000 years), to show that something is better in some way is ... well, not impossible, perhaps, but certainly a tall order.

I didn't know I had "acolytes." More to the point, I reject the implication that we're talking about religion here ("acolyte" being a religious term, at least originally). Relational technology is a matter of science (yes, it is, Curt!—see later), not religion, and we'd all be better off if discussions of it were more scientific in tone.

2. Wrong again ... Date has been quoted saying that while the idea of the Information Principle was Codd's, the phrasing is Date's.

Now this is just silly. You might as well say Hamlet's soliloquy is mine, not Shakespeare's, because (having a tin ear) I give it in the form "To exist or not to exist: that's the conundrum."

We do claim that our "views" are based on math and logic. We also claim that math and logic are science! First, Bertrand Russell is on record as asserting that math and logic are the same thing ("it has become wholly impossible to draw a line between the two; in fact, the two are one," quoted by William Dunham in The Mathematical Universe, John Wiley & Sons Inc., 1994). Now, I believe Russell and Whitehead's attempt to derive the whole of mathematics from logic did ultimately fail, but to a first approximation and for present purposes (where in particular the only sets we ever have to deal with are finite) we can take Russell's dictum as valid. Second, I lost count of the number of respected thinkers I checked (Gauss was one) who simply stated, in effect, that of course mathematics is a science. I'll content myself here with just two observations:

· CHAMBERS TWENTIETH CENTURY DICTIONARY defines mathematics as "the science of magnitude and number, and of all their relations" (!).

· A recent "popular science" book—one I liked a lot, incidentally—is titled MATHEMATICS: THE SCIENCE OF PATTERNS (by Keith Devlin, Scientific American Library, 1994). Here's a quote: "As the science of abstract patterns, there is scarcely any aspect of our lives that is not affected, to a greater or lesser extent, by mathematics; for abstract patterns are the very essence of thought, of communication, of computation, of society, and of life itself."

Actually I don't really care whether we regard math and logic as science (even though I do so myself). Let's argue the substance, not the labels.

4. If you want to argue that logic and math are sciences ... you can wait until I get around to more carefully debunking [Pascal's] and Date's outrageous intellectual fakery on the subject directly, and respond at that time ... Mathematics = studying abstract systems which follow various sets of assumptions (axioms) ... Science = studying of [sic] the real world ... Engineering = changing the real world ...

On the question of whether math and logic are sciences, see my response to Monash's remark number 3. It does seem to me, however, that a debater who accuses his opponents of "outrageous intellectual fakery" needs to be prepared to back up that accusation with solid evidence—which Monash hasn't done, so far as I know—or else be accused, deservedly, of the very same thing himself. It's pretty strong language! Speaking for myself, I've always tried in my writings to explain why I hold the opinions I do. I've also had occasion to change my mind from time to time, but again I've always tried to explain why I've done so. In other words, I've tried to be open, clear, and scientific in those writings. By contrast, I find Monash's remarks to be none of these things.

"Mathematics = studying abstract systems which follow various sets of assumptions (axioms) ... Science = studying of [sic] the real world ... Engineering = changing the real world ...": These glib "definitions" add nothing to the debate. However, I'd like to say this: Mathematics does indeed involve the study of "abstract systems," but the abstractions in question are often derived from empirical observations (euclidean geometry provides an obvious example). And when they aren't, they often turn out later to have useful concrete realizations anyway (group theory is a good example here). Also, of course, we have the classical division of mathematics into "pure" and "applied" branches; both of my examples fall into the "pure" side of the house, which serves to illustrate the point that even "pure" mathematics has practical applications. To get back to database management, I regard relational theory as an application of "pure" set theory and logic.

Regarding "various sets of assumptions (axioms)," see my response to Monash's remark number 6.

5. It seems that whenever I look at an argument by Chris Date or Fabian or [Alf Pedersen], it quickly gets to the point where you [sic] say "I'm not talking about how to use actual products that are shipping today. I'm talking about how to use products that I hope will ship in the future" ... [At] least [Alf is] not repeating the "science" nonsense of Chris Date and Fabian Pascal.

"[They're] not talking about how to use actual products that are shipping today": Well, this simply isn't true, is it. Much of what we (or at least I) write is indeed concerned with concrete advice on how to use today's products without digging yourself into some pit or other. A very recent example is a detailed response I wrote last week to a question regarding the representation of missing information (that response is due to appear soon on www.oreilly.com). I would also argue that my books AN INTRODUCTION TO DATABASE SYSTEMS, 8th edition and DATABASE IN DEPTH: RELATIONAL THEORY FOR PRACTITIONERS are both very much concerned with how to use today's products properly—not to mention the regular column I used to write for Database Programming & Design.

"[They're] talking about how to use products that [they] hope will ship in the future": Well, this one isn't true either. We aren't so much concerned with trying to tell people how to use some hypothetical future product. Rather, we're trying to lay down a foundation for some such future products ... just as Ted Codd did, in fact, when he first described the relational model in his papers of 1969-71. That's a legitimate exercise! Or do you want products without a solid foundation? (That's a rhetorical question, perhaps. From what Monash says in the Computerworld threads, it looks like he does want such products. Certainly many of the things he discusses, apparently approvingly, are things for which we don't have products with any solid foundation.)

"[At] least [Alf is] not repeating the 'science' nonsense of Chris Date and Fabian Pascal": See my response to Monash's remark numbers 3 and 4.

6. Chris Date is BADLY confused in that discussion of geometry. HE DOESN'T KNOW WHAT AN "AXIOM" IS!!!!!!!!!!!!!!!!!!!!!!!!!! Or else he DOES know, and is deliberately debasing the discourse for his own personal advantage. Either alternative is sickening. I want to throw up.

I don't much care for ad hominem arguments at the best of times, but this one has to be one of the worst such that I personally have ever been subjected to. I categorically deny that I was "badly confused" in my interview in that "discussion of geometry." Of course I know what an axiom is (yes, I too have a degree in mathematics—not that such a qualification is or should be necessary in order to understand such a simple concept). I find it deeply offensive that anyone who claims to be knowledgeable in our field should suggest that I don't.

The second part of Monash's remark here—"or else he DOES know, and is deliberately debasing the discourse for his own personal advantage"—manages to be even more offensive. Who does Monash think he is? And why on earth would he think I might "debase the discourse for my own personal advantage," when anyone who (unlike me, according to him) does know what an axiom is would be able to find me out in a moment? And how on earth could "debasing the discourse" in such a manner act to "my own personal advantage" when it would (according to his allegations) serve only to show me up as an ignoramus? Monash, you owe me some very public apologies.

"Either alternative is sickening. I want to throw up." Go ahead, be my guest. (See, I can descend to your level of discourse if I really want to.)

7. I STRONGLY reject the argument "relational theory is based on math/logic/science, and hence must be the best guide to practical database management."

I don't think relational advocates have ever made exactly this argument; at least, I'm quite sure I haven't. But I do think that a "guide to practical database management" that's "based on math/logic/science" is likely to be better than one that isn't. I'm a believer in the importance of careful reasoning; indeed, I agree with Edward Abbey, who wrote (in Abbey's Road, in an essay titled "Science with a Human Face") that "Sweet Reason [is] the newest and rarest thing in human life, the most delicate child of human history." The relational model is based on careful reasoning; and while I don't deny that (for example) an "XML model" might also be based on such reasoning, I think it's obvious that what would result would not be as suitable as the relational model as a basis for "practical database management." To paraphrase something I said in the O'Reilly book Database in Depth: Suppose it's your job to implement some front-end application, say a decision support or data mining application. Which would you prefer as a target?—a relational DBMS, or some other kind, say an XML or object-oriented DBMS? I think the answer is obvious. Of course I could be wrong, but Monash signally fails to provide any evidence that I am.

Posted 9/30/05