Re: WWW/Internet 2009: 2nd CFP until 21 September x
Date: Sat, 15 Aug 2009 19:05:49 -0300
paul c wrote:
> Walter Mitty wrote:
>> Where a lot of database designers go awry is to assign "meaning" to >> missing data. For example, if spouse's first name is missing, it >> means the person has no spouse. People who set up information systems >> with this kind of conventional interpretation are entering the >> quicksand. I guess, to come back to your comment of a few replies >> back, this means that I share your qualms about the CWA. If there is >> no passenger named Yussuf Islam on the passenger list, it could mean >> that there is no such passenger on the plane, or it could mean that >> the passenger list is incomplete or incorrect. (I may be >> misrepresenting your position on the CWA, but you get the idea.)
> You are misrepresenting it. I have more than qualms about the OWA, not
> the CWA.
> 1) db meaning is expressed by predicates (and to be precise, meaning
> isn't assigned to data, rather data (values) are substituted in the
> predicate's place-holders, giving a statement that by the CWA convention
> is true. If the predicate is "spouse's first name is X" and there is
> no value to substitute for X, the meaning is that there are no spouses's
> first names. That predicate doesn't mention anything at all about there
> being no spouses, the sentence "there are no spouses" is not even a
> predicate, it is a proposition. This is very simple, simple enough for
> a child to learn but adults like Scott are so easily tripped up, like
> trying to substitute four values into six placeholders don't ask me why,
> maybe it's because they confuse a big vocabulary with basic understanding.
> 2) it might be quite reasonable in some particular circumstance to
> conclude that somebody isn't on the plane because his name isn't on the
> passenger list, but if the db designer has told the users that the
> predicate is "X is on the passenger list", that is the only predicate
> expressed by the db. Concluding X is not on the plane has nothing to do
> with the db. This is the essential agreement required by the CWA.
> Whereas if there are no values for X, the OWA allows any user to
> conclude that somebody is on the passenger list, regardless of what's
> recorded in the db. Actually, people who subscribe to the OWA, don't
> really need a db, lucky them.
> 3) Whether db information is incorrect is nothing to do with RT. Whether
> it's incomplete is very much to do with RT, that's why we have
> constraints. Don't remember if I mentioned it before, but I think
> whenever talk of incomplete data comes up, the best way to be clear
> about whether a given constraint handles it is to phrase it in the
> algebra and then ask whether the predicate's place-holders are satisfied.
Walter wants a clue. If every table with null represented a materialized outer join, as he claims he thinks it does, then presumably something is being joined. Unnamed perhaps. Implicit for sure. But some ephemeral tables or relations are joined (outer joined) to materialize the table.
Each of those unnamed, implicit, ephemeral tables represents a predicate. And the CWA tells us that any row not in the table is false according to that predicate. Thus, a NULL spouse's name, obviously and necessarily, means someone has no spouse. If we accept Walter's nonsense that tables with null represent materialized outer joins, that is. Received on Sat Aug 15 2009 - 17:05:49 CDT