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

In article <UgClf.13909$ea6.6355_at_news-server.bigpond.net.au>, terabitemightbe_at_bigpond.com says...
> The trouble with your definition is that it is nonsensical - more than a
> bit like what Alice encountered!

I don't think you have understood the nature of our discussion.

> The root problem is that the "UNKNOWN" (aka NULL) itself is not and
> never will be a member of that or any other domain. Accordingly others
> have called is a "Special" value.

You misunderstand. In 3VL, there are three truth values, often denoted by the symbols 0, 1 and 2. Neither of those is identical to the mysterious SQL NULL; on that, both vc and I agree (I think).

Perhaps you are thinking about the "obvious" implementation of booleans in SQL, where you would have two possible values---TRUE and FALSE---and the possibility to set the attribute to NULL, like an attribute of any other type in SQL. That is a very different thing, though this difference may actually be hard "to get".

FWIW, I agree that { TRUE, FALSE, NULL }---using "NULL" in its usual SQL sense---is not a domain, because NULL is not a value.

> When the attribute value does become known (lets say it is "MAYBE") then
> it is no longer UNKNOWN, it is MAYBE and it is a member of the domain
> (YES, NO, MAYBE) (sic) :-)

Or are you just suggesting that I should use MAYBE instead of UNKNOWN as a more user-friendly name for the third 3VL value that is neither TRUE nor FALSE? It seems strange (to put it mildly) to say that my "definition" is "nonsensical" for that reason alone.

> Personally I don't understand why NULL is so hard "to get".

It's not that hard to memorise the rules about NULLs in SQL, but they are not particularly intuitive or self-evident. But it makes sense if you squint a little and don't think too hard about it, I suppose. :) Or YMMV with regard to what you mean by "to get".

-- 
Jon