Talk:Domain-key normal form

Highest?
Is this a higher normal form truly? The example simply shows a combination of two attributes in one field, something that wouldn't even pass 1NF if you simply rewrote it as an array stored in a field, which it basically is. Gigs (talk) 05:31, 16 December 2007 (UTC)


 * I agree; based on the First normal form article, it's not 1NF, since evilness and eccentricity are two attributes being stuffed into one column. I wonder, what about a table with an “upper bound” column and a “lower bound” column, with the constraint that the upper bound is greater? That seems to have the same problem as the given example (two columns with an interdependent constraint), but would be much simpler (assuming I'm correct that such a two-column constraint violates DKNF). Also the solution is less obvious in this case. (My guess would be that you'd store the total range — upper bound minus lower bound — instead of the upper bound, and say the range must be positive; that'd pass DKNF now, right?) Luke Maurer (talk) 22:00, 21 May 2008 (UTC)

In my opinion, the example given is flawed. You are both converting to DKNF and changing the data as well. A better example would be that you get two tables: one for millionaires and one for billionaires. That satisfies the DKNF requirement since the keys are unchanged and the domains are now constrained in the tables. Remember: the purpose of DKNF is to split generalized constraints into (testable) Domain constraints and Key constraints. I will add this reason to the page. If no comment arrives in a week or so, I'll change the example to clear it up. RonaldKunenborg (talk) 10:51, 8 July 2009 (UTC)

Incredulous
1 The entire article is an insult to knowledge and reason. "DKNF" is the result of persons untrained and inexperienced with the theory and practice (the two are inseparable) of Normalisation, purporting to "normalise" a database.
 * 1.a It is an ad hoc form of "do-the-best-you-can" organisation, but it has no relation to the terms "database" or "normalisation", and thus those terms should not be used to describe it.


 * 1.b Using such terms to describe it demeans those terms, which were established 30 years prior, and remain of value to the educated database professional.


 * 1.c It attempts to obtain some credibility to itself by identifying itself as relating to established standard terms.
 * 1.d It fails to identify any specific relation to those standard terms, and therefore it has no relation to those terms.

2 The Normal Forms are progressive: 3NF contains 2NF, 2NF contains 1NF; therefore since "DKNF" is neither a formal progression of 4NF or 5NF (which are currently defined, understood and implemented), its claim to being the "highest" Normal Form is entirely false.

3 Since the process identified does not have 1NF, it is not a Normal Form of an kind; from the description, it is merely a status of a non-standard organisation of data in a database storage facility, with no Integrity save for the "domain constraint" and "key constraint".


 * 3.a Even those are supplied as lay descriptions, not technical definitions, and are thus of no value to anyone seeking to understand them.


 * 3.b It avoids confirmation/denial in relation to established and defined terms.

4 The proposition that "The Boyce-Codd normal form, Third normal form, Fourth normal form and Fifth normal form are special cases of the domain/key normal form" is preposterous for several reasons:


 * 4.a In order for those to be "special cases" of it, then "DKNF" would be a general case of BCNF, 3NF, 4NF, and 5NF.  There is no evidence of that.


 * 4.b It would therefore be a lower normal form to them


 * 4.c It therefores contradicts itself as being the "highest" normal form.


 * 4.d In any case, since the relation is made to defined terms, it would be a simple matter to define it in technical terms, relating to BCNF, 4NF, and 5NF.  Such definitions are absent; only a summary lay description is provided.

5 The examples (both the initial and the suggested "compliant" examples) are so poor and confusing the provide neither an example of the undefined proces of normalising to "DKNF", nor the specific definitive difference between the "non-compliant" and "compliant" versions. That is understandable, since the "dknf" is undefined.

6 The example given does not comply with 1NF; this fact further confirms my assertions re [2], [3] and [4]. It is not a "normal" "form" of any kind.

7 The statement "DKNF is frequently difficult to achieve in practice" has no basis. In fact, given the non-specific and non-deterministic lay descriptions given here, and the entire absence of relation to established technical terms (those supplied being false, as identified above), any uneducated person can achieve it; anyone can call any "organisation" of data DKNY. Further, whether it is achieved or not can be confirmed or denied by anyone. It is indeterminate concept.

The rest of the article is marketing hype, and not worthy of specific attention.

At best, "DKNF" is as identified in [1.a]. In fact, it is discussed in educated circles as Don't Know Normal Form.

If Wiki is to be taken seriously as an encyclopedia, we must keep such abject nonsense out of it. DerekAsirvadem (talk) 06:50, 31 August 2009 (UTC)

To the above person; instead of whinging in your high-pitched, academic whine - why did you not improve the article? You have put so much effort into criticising it point by point and none into actually sharing knowledge on the page itself. — Preceding unsigned comment added by 143.238.123.251 (talk) 17:02, 6 November 2011 (UTC)
 * The article seems unfounded, even the last example "normalization" is wrong, since it does explicitly not prevent the mentioned "complex" constraints (for the ranged domain of net wealth). Possibly, this article should be deleted.--&#42;thing goes (talk) 01:30, 26 July 2015 (UTC)
 * Domain-key normal form is a genuine thing, so this article should not be deleted. I don't know enough about DKNF to say whether the current article gives an accurate description of it though. —Ruud 11:58, 28 July 2015 (UTC)
 * True, but at the moment, the article is not even near to encyclopaedic standards. Even less so, if compared to the articles on 3NF, BCNF, 4NF, 5NF.--&#42;thing goes (talk) 17:22, 29 July 2015 (UTC)
 * Feel free to remove any incorrect or unencyclopedic information from the article. Having a short article that points to readers to Fagin's original paper seems at least more useful than having no article at all. —Ruud 19:39, 29 July 2015 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified 2 one external links on Domain-key normal form. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20110606025027/http://dev.mysql.com/tech-resources/articles/intro-to-normalization.html to http://dev.mysql.com/tech-resources/articles/intro-to-normalization.html
 * Added archive https://web.archive.org/web/20071104115206/http://www.utexas.edu:80/its-archive/windows/database/datamodeling/rm/rm7.html to http://www.utexas.edu/its-archive/windows/database/datamodeling/rm/rm7.html

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at ).

Cheers.— InternetArchiveBot  (Report bug) 21:51, 14 December 2016 (UTC)

Book example is nonsensical
The current example claims that "There is a constraint linking Wealthy Person Type ["Type" below] to Net Worth in Dollars ["NetWorth" below], even though we cannot deduce one from the other."

This is clearly nonsense: as the next sentence explains, "The constraint dictates that a Millionaire will have a net worth of 1,000,000 to 999,999,999 inclusive, whilst a Billionaire will have a net worth of 1,000,000,000 or higher." There is therefore a functional dependency NetWorth -> Type. This means the example table isn't even in 3NF, which could be fixed by splitting into Person-NetWorth and NetWorth-Type relations. DKNF adds no further insights here. Charnel Mouse (talk) 14:56, 17 February 2024 (UTC)