Re: Proposal: 6NF

From: David Cressey <dcressey_at_verizon.net>
Date: Sun, 01 Oct 2006 13:47:43 GMT
Message-ID: <3GPTg.1980$6S2.607_at_trndny02>


"Cimode" <cimode_at_hotmail.com> wrote in message news:1159707552.227184.276470_at_k70g2000cwa.googlegroups.com...
> Imaginary numbers are not values belonging to domain of reals.
> Nevertherless they are values.
>
> NULLS are NOT values.
>
> NULLS is SQL's poor way of handling missing data by making arbirtrary
> assumptions about what *should be there*. There is a number of
> possible values greater than one for each NULL value stored in a
> database and therefore predicate can not be validated (therefore you
> have no facts in your database). In a way, one could conceive NULL's
> as a occurrence of some random function in a limited set of
> possibilities but certainy not as a value. To make your database state
> facts about reallity under the form of predicate you use values that
> belong to well defined domain of values that are derived from one
> another.
>
> NULL is the carpet under which you hide the dirt of missing data
> instead of throwing it to the trash. It is practical YES. But it has
> the drawback of making your count results false if you, your team and
> all people working on the db during the system life cycle, do not put
> additional IS NULL/IS NOT NULL conditions in all procedures selecting
> values from tables allowing NULLS. Adding such conditions kills
> performance at overall server level. So. Result FALSE/ Performance
> (Response Time/Concurrency) Down.. Do you still believe that NULLS are
> a good idea?
>

What's your proposal for a systematic way of dealing with missing data? Received on Sun Oct 01 2006 - 15:47:43 CEST

Original text of this message