October 6, 2011 at 6:07 pm
L' Eomot Inversé (10/6/2011)
L' Eomot Inversé (10/5/2011)
I find it quite frightening that so few people know enough about normalisation to get this one right.I feel much less frightened now: 22% out of 700 is a lot better than 2% out of 200.
I must be an inveterate cynic. I don't find it at all surprising that the percentage of sucess has risen so dramatically after the next day's newsletter included the correct answers. (never was frightened, though)
October 7, 2011 at 2:21 am
john.arnott (10/6/2011)
L' Eomot Inversé (10/6/2011)
L' Eomot Inversé (10/5/2011)
I find it quite frightening that so few people know enough about normalisation to get this one right.I feel much less frightened now: 22% out of 700 is a lot better than 2% out of 200.
I must be an inveterate cynic. I don't find it at all surprising that the percentage of sucess has risen so dramatically after the next day's newsletter included the correct answers. (never was frightened, though)
Now that you mention it, that certainly looks like a possible explanation. Much more likely than a spontaneous improvement.
Tom
October 7, 2011 at 6:20 am
L' Eomot Inversé (10/7/2011)
Now that you mention it, that certainly looks like a possible explanation. Much more likely than a spontaneous improvement.
Yes, that and the fact that Steve added to the question the number of boxes that you have to tick.
John
October 7, 2011 at 8:01 am
L' Eomot Inversé (10/6/2011)
mtassin (10/5/2011)
I think a big part of this is the flavor of Normalization Kool-aid we're using.For instance, this paper that's referenced isn't necessarily Normalization Form mainstream.
I find it amusing to see a claim that a seminal paper by the man who id now director of the UCLA Web Information System Lanboratory and holder of UCLA's N. E. Friedmann Chair in Knowledge Science, an associate editor of VLDB Journal, who has been program chair or general chair of VLDB (more than once), SIGMOD CMD, NACLo, NID, EDBT, and chaired HotSWUp in April this year should be dismissed as not necessarily mainstream.
Look, you've been drinking his Kool-aid. And you agree with him.
But I also find no mention of EKNF on Wikipedia
http://en.wikipedia.org/wiki/Database_normalization
I'm not saying it's an invalid concept, what I'm saying is that expecting people to know a branch of Normalization that isn't part of the maintstream, and then getting worried when only 2% of us know what the heck you're talking about, and you talking about it like it's the best thing since sliced bread is just silly.
October 7, 2011 at 7:25 pm
mtassin (10/7/2011)
L' Eomot Inversé (10/6/2011)
mtassin (10/5/2011)
I think a big part of this is the flavor of Normalization Kool-aid we're using.For instance, this paper that's referenced isn't necessarily Normalization Form mainstream.
I find it amusing to see a claim that a seminal paper by the man who id now director of the UCLA Web Information System Lanboratory and holder of UCLA's N. E. Friedmann Chair in Knowledge Science, an associate editor of VLDB Journal, who has been program chair or general chair of VLDB (more than once), SIGMOD CMD, NACLo, NID, EDBT, and chaired HotSWUp in April this year should be dismissed as not necessarily mainstream.
Look, you've been drinking his Kool-aid. And you agree with him.
But I also find no mention of EKNF on Wikipedia
http://en.wikipedia.org/wiki/Database_normalization
I'm not saying it's an invalid concept, what I'm saying is that expecting people to know a branch of Normalization that isn't part of the maintstream, and then getting worried when only 2% of us know what the heck you're talking about, and you talking about it like it's the best thing since sliced bread is just silly.
Frankly I find these repeated kool-aid cracks repulsive. I usually try to be less bloody offensive than you seem to regard as the norm for serious technical debate - but of course maybe you don't think this topic is worthy of serious debate.
Ok, maybe I understand where you are coming from. Your concept of what may be useful and relevant is what's in Date's textbook and consequently (since most people think Date walks on water) what's in Wikipedia (without regard to the talk pages, which of course ought to be an essential part of the discussion although you haven't looked there).
I can easily change your view the wikipedia view by changing the page you reference (the inclusion of EKNF on the page you reference was suggested some time ago and has been supported by other comments on the talk page , with no adverse comments at all, so such a change would be perfectly reasonable). I can't change Date's text book, and he won't because he regards the representation principle as nonsense, in distinct conflict with, for example, Ron[ald] Fagin - whom you maybe regard as mainstream as he is responsible for the standard definitions of 4NF and 5NF - as made eminently clear by Date's very public attack on Fagin's DKNF suggestion - although I agree that RF got that one wrong, and CJD was right to challenge it, the style of the attack and some of its content (suggesting that attempting to achieve uany useful degree of representation was a waste of time) was appalling. You appear to be relying on an appeal to an "eminent authority" to suggest that anything you have not been taught by that "eminent authority" is irrelevant in our field of work.
If that's the way you think, I have to accept that that's the way you are and give up on changing your mind - I just have to hope that that silliness doesn't influence other people who may read this discussion.
Tom
October 12, 2011 at 5:43 am
I notice that we are now up to approimately 25% correct.
How on earth did that happen?
December 8, 2011 at 2:44 am
I think that volcanic air is getting to you Tom, This question is way over most of our heads - a really hard question. No amount of googling seemed able to turn up any clues. And... well, once you see the answer its no easier to work it out.
I got the not null part. And its probably not the primary key, (otherwise why call it an elementary key). The middle one long complex answer I chose this because it was long and complex - don't pretent to understand it even now, and random choice for the other one.
December 8, 2011 at 5:20 am
Tom Brown (12/8/2011)
I think that volcanic air is getting to you Tom, This question is way over most of our heads - a really hard question. No amount of googling seemed able to turn up any clues. And... well, once you see the answer its no easier to work it out.I got the not null part. And its probably not the primary key, (otherwise why call it an elementary key). The middle one long complex answer I chose this because it was long and complex - don't pretent to understand it even now, and random choice for the other one.
I probably overestimated how much relational theory people would be familiar with, and put the question into a form that was harder than it could have been if presented in a different form. Multiple choice questions where you have to select several answers are inherently more difficult than multiple choice questions where you select only one answer (choosing 3 out of five means choosing one combination out of 10 possible combinations, choosing 1 out of five is easier than choosing 1 out of 10).
I'm surprised that googling came up with nothing, though - it finds the Halpin,Morgan and Morgan book, Joe Celko's book, the Wikipedia normalisation article, Zaniolo's paper, and Scot Becker's rather silly (he parades his ignorance of why EKNF is important) paper on normalisation (Google gives a broken link to that one, but it's there in the cache - at least it is today) all on the first page of results when I try it - but maybe Google knows I'm interested in normalisation.
Tom
December 8, 2011 at 6:06 am
Yes I found the Wiki, and other articles, but the whole topic of normalisation seems couched in academic terms.
I was trying to google for simple definitions of things like elemetary attribute key or prime attribute key - definitions that don't refer to other similar terms, sending my round in circles of cyclic definitions. I've since read your 3 normalisation articles, and heard mention that you may be planning a 4th. Perhaps you could have a bash at some good practical definitions of the terms used in normalisation - as a kind of 'demistification' level zero article. I'm sure there is something simple I'm not getting in normalisation, however here are my 'simple' definitions (which may be wrong)
Attribute: This is any piece of data. In 1NF equates to a column
Attribute Key: a piece of data that on its own, or when combined with other attributes could uniquely identify a row.
Candidate Key: any column, or combination of columns in 1NF that would uniquely identify a Row
Primary Attribute Key: A single column candidate key
Elementary Attribute Key: ~ still unsure about this one.
Viewing 10 posts - 46 through 54 (of 54 total)
You must be logged in to reply to this topic. Login to reply