ON INCOMPETENCE

From: Ted Hills

Date: 09 Feb 2006

Well, I certainly did not know whether to laugh or to cry when I discovered that the only required (i.e., not nullable) column in the central fact table of a major corporation's data warehouse is the row

id. I got quite a chuckle when I began to imagine what a row in this table means when it contains nothing but the row id. On the serious side, you could imagine that the table predicate is, "The row exists!" On the humorous side, perhaps the row means, "We know that somewhere out there is information that just maybe should be in this row—or perhaps another row just like it, except for the row id." You can't make this stuff up!

The NULL is the database equivalent of the unrestricted GOTO in programming. It lets you do any undisciplined thing you like.

From: Fabian Pascal

Why are you surprised? Without any proper education, why should knowledge and ability to reason materialize? Incompetence is fast becoming an american hallmark, and IT is hardly the most serious of the consequences.

From: Tim Armstrong

Date: Fri, 10 Feb 2006

Another example of the current state of affairs in education (this from the Instructor):

It is pretty much accepted that normalization beyond 3NF is pretty much an academic exercise and not all that beneficial to DB's in the real world. So we, like most will only be normalizing to 3 NF.

The above was in response to a suggestion that a relation that one of the students was using to demonstrate modification anomalies could actually be in DKNF if there were certain business rules in place. BTW, I’m one of the students and I want to learn how to fix lousy logical designs and more importantly, how to create correct conceptual models. I guess I’ll end up learning it on my own.

If it’s an academic exercise, why aren’t we learning it?

I took your advice and bought your book, PRACTICAL ISSUES IN DATABASE MANAGEMENT. A lot cheaper than the DB class I’m taking and far more informative.

From: Fabian Pascal

The educators today are the product of the same failed educational system or, worse, are uneducated practitioners who are hired by academic institutions because of their “practical experience”, to do essentially product training.

From: Alf Pedersen

Date: 20 Mar 2006

Just got this remark from one of my pages:

I was asked to look at NULLs in a Database by someone for whom I'd done some performance improvements to their production database, which included making a number of fields nullable. The basis of the argument in the article - that the value of a NULL field is unknown - is completely wrong. We know precisely what the value is. A NULL value in a field means that it "has no value". This is not the same as unknown, unsure, don't know, forgotten, haven't been told, etc... as all of those imply that is does have some value. It doesn't. Space precludes me from giving a full explanation of this concept, but on request I will happily do that plus give examples. But your example of count(available) from accounts belies your lack of understanding of the concept. With three rows in the table of which one is NULL, the count MUST return two, as only two rows have a value. The rules of mathematics state this.

Regarding your final points, introducing NULLs does not mean n-valued logic - it means 3-valued logic (three state logic). Neither does it introduce uncertainty. In fact, the reverse is true. I have a recent example where a number field was made not nullable, and held zero when it should have held null. The problem was they could not differentiate between records that really had a zero value and those that had just defaulted to zero (and should have been NULL). I have been designing relational databases for twenty years now, and it worries me that such articles get published where a lack of understanding of basic mathematics is shown. People will get the wrong impression, and write bad solutions. The only good thing is that as a freelance, I may get plenty of work.

--Nigel N.

Well, well.

From: Fabian Pascal

I think he said the same nonsense in some database group, but can't recall in what thread.

From: Alf Pedersen

And I, who actually thought my article about nulls was pretty good, showing what false conclusions could be drawn from using nulls!! :-)) How could I be so wrong... :-)

To laugh or cry, yes... Unbelievable.

From: Fabian Pascal

But if it's possible to have a career without knowledge, what do you expect other than Unskilled and Unaware of It.

Anyway, here's the world's only superpower: More Than Half of Americans Reject Evolution, Back Bible.

Database issues are nothing in the scheme of things, and the real issues are clearly nothing to laugh about: the west is committing suicide with the help of the US christo-fascists, with the islamo-fascist barbarians at the gate. If those two groups want so much to be close to their god in heaven, why don't they die for their gods and leave rational people in peace, free of nonsense?

Posted 3/31/06