Re: So what's null then if it's not nothing?

Jon Heggland wrote:
> In article <1134354197.583597.21280_at_o13g2000cwo.googlegroups.com>,
> boston103_at_hotmail.com says...
> >
> > Hugo Kornelis wrote:
> >
> > > There is a reson to object to the use of "the null value" to represent
> > > the truth value Unknown. Elsewhere in the standard, the Null value is
> > > defined as "A special value that is used to indicate the absence of any
> > > data value". Of course, if Unknown is considered to be a data value for
> > > a boolean data type (and the rest of the standard clearly indicates it
> > > is), it should not be represented by the same symbol that is also used
> > > to represent the absence of a data value, since a data value can not be
> > > absent and present at the same time.
> >
> > Right.
> >
> > > Further in the document, in the part where <boolean value expression> is
> > > detailed, the boolean truth tables are given. According to these tables,
> > > False AND Unknown should evaluate to False and True OR Unknown should
> > > evaluate to True. Makes sense. But since Unknown and NULL are considered
> > > equal in the context of a Boolean data type, this also means that False
> > > AND NULL evaluates to False and True OR NULL evaluates to True. The
> > > basic rule of NULL propagation is out of the window!!
> >
> > Apparently, whoever created this part of the standard had no clue
> > about what any kind of logic is.
>
> Or perhaps they thought it was not a problem at all to use the same word
> "NULL" for different concepts, behaving differently. After all, it is
> obvious from the context when NULL means 'missing' and when it means
> 'unknown', no? :)

There is indeed no problem with using the same word, but why use *two* ? Even the unnecessary synonymy would be OK unless there were more serious problems with their 3V 'logic' treatment I mentioned in my response to Hugo.

> --
> Jon
Received on Mon Dec 12 2005 - 15:41:32 CET