Re: Examples of SQL anomalies?
Date: Wed, 02 Jul 2008 16:57:36 GMT
Ed Prochak wrote:
> No. Consider the case of a credit card bill which does not have a
> repeatable balance month to month. There are months when the account
> balance is zero. But zero is not the correct answer to give your
> accountant when you haven't yet received the bill. The correct answer
> is "haven't got that bill in yet", aka, NULL But zero would not tell
> you the difference.
That is likely just a example of incomplete design. Probable castle in the air too, like as so often happens because of people failing to nail down specs/requirements, then proceeding to build a system that is partial and resorting to porridge phrases like "real-world" in reply to why it produces ambiguous answers.
Even a book-keeper, at least a competent one, would then ask you to provide an accrual instead. Can't count the many systems I've seen that were tied up in knots because or people trying to answer two questions with one fact, which is not always possible. I guess this shouldn't surprise me - for example, Joe C whose devious bumpf is apparently bought by many people, regularly introduces new questions without changing the original scheme, does it here all the time. Beyond me why so many don't seem to recognize his boringly repetitive M.O. which is not even a one-track horse. To be fair, he's not the only, the phenomenon seems rampant and the big consultancies love it for the unnecessary but extra billings they can make. Such systems need perpetual fix-ups. Lazy, incompetent and naive users and managements who allow db developers to make up biz requirements don't help either.
In general, such users are in the majority and it is no surprise that anybody who points this out, such as Fabian P, gets shouted down. It figures - English composition holds many more possibilities for nonsense than sense. Received on Wed Jul 02 2008 - 11:57:36 CDT