"--CELKO--" <jcelko212@earthlink.net> wrote in message
news:51589058-a7a2-4a1e-a9c2-46ba44f49d4c@59g2000hsb.googlegroups.com...
>>> Once upon a time there was a questionnaire which has the following
>>> question:
> Do you have sex regularly? <<
>
> LOL! That is an old Woody Allen joke about a man and woman going to a
> therapist and being asked that question:
> He: "Almost never, 3 times a week!"
> She: "Constantly, 3 times a week!"
>
>

http://en.wikipedia.org/wiki/Coolidge_effect

--
Greg Moore
SQL Server DBA Consulting Remote and Onsite available!
Email: sql (at) greenms.com http://www.greenms.com/sqlserver.html

> right now. I am getting a physical on 2008-04-01 and just had to fill
> a four-page basic intake questionnaire I got in the mail.

This is a VERY appropriate date for you Joe.

>> And do you think the fact that John likes lemonade should be described by other facts in the database <<

How about a list of beverages, with the value Lemonade in it? That
would be nominal scale and not a flag. When we decide to research
other beverages, we extend the scale. Unlike a flag, I can ask how
much of a given beverage he drinks.

I extrapolate that if John drinks s certain number of Cokes and is of
a certain age, then I stand an 85% chance of selling him Pepsi (i.e.
kids the sweeter Pepsi to Coke). You cannot get that kind of
information from flags.

> How about a list of beverages, with the value Lemonade in it?

As I have asked several times already, what if they are only interested in
lemonade? Why should they bother with a list?

Anyway, you are still clearly either missing the point or intentionally
disregarding it. This is not about lemonade; this is about the fact that
some things are just yes/no indicators on their own, without "" from
other facts. If you don't get it, you don't get it, and I'm afraid I can't
to educate you any further. :-(

> Anyway, you are still clearly either missing the point or intentionally
> disregarding it. This is not about lemonade; this is about the fact that
> some things are just yes/no indicators on their own, without "" from
> other facts. If you don't get it, you don't get it, and I'm afraid I can't
> to educate you any further. :-(

To my (mostly) unbiased eye, it looks like both sides are missing the
point (either intentionally or otherwise)... Joe has already said
that he accepts that there are rare cases where a 2 value domain is
valid, but that in *most* cases modelling a flag is not "correct".
But we all know that Joe's only interested in "correct" from a
standards and theoretical perspective though, so it is pointless to
argue the application of flags from a practical perspective.

From a purely theoretical perspective I can see his point ... from a
practical perspective I have no problem disregarding his point if it
makes sense in the context of the problem I'm facing
For example (and in keeping with the theme of this thread), I've heard
that in market research it is fairly normal to build a database purely
to record the results of a single survey. This is essentially a throw-
away piece of work that exists only to analyse data from that single
survey, the reason being that it's more cost-effective than building a
"proper" database to store more generic survey results. In this
scenario it is a waste of time to design flexibility into the database
as the analysis is usually predefined by whatever research model the
company is using. The results of the analysis however, would likely
be stored in a well designed database to be compared against past/
future surveys...

J

>>
> Anyway, you are still clearly either missing the point or intentionally
> disregarding it. This is not about lemonade; this is about the fact that
> some things are just yes/no indicators on their own, without "" from
> other facts. If you don't get it, you don't get it, and I'm afraid I can't
> to educate you any further. :-(

To my (mostly) unbiased eye, it looks like both sides are missing the
point (either intentionally or otherwise)... Joe has already said
that he accepts that there are rare cases where a 2 value domain is
valid, but that in *most* cases modelling a flag is not "correct".
>>

Well, to be fair, I did not say that *most* columns should be a two-value
flag, either. But in my experience they are more common than Celko's
"advice" would lead one to believe.

> Well, to be fair, I did not say that *most* columns should be a two-value
> flag, either. But in my experience they are more common than Celko's
> "advice" would lead one to believe.

Perhaps what I should have said is "in *most* cases where a flag has
been modelled, it is not "correct" to have done so"

>> As I have asked several times already, what if they are only interested in lemonade? Why should they bother with a list? <<

If the only concern is about lemonade preferences and consumption,
wouldn't everything in that table deal with lemonade consumption? If
so, why would they bother with a flag? It would be like a Personnel
table with a flag that asks "Are you an employee?" when the answer
would have to be "yes" to get into the table.

Let me recover a bit from my physical exam, x-rays and booster shots
and see if I can get a short article about scales, measurements and
data values versus question/answer and other types of flags.

> If the only concern is about lemonade preferences and consumption,
> wouldn't everything in that table deal with lemonade consumption?

Celko, you are too much.