Databases Reference
In-Depth Information
To repeat, this topic represents my most recent thinking. But I make no claim that what it
has to say represents the last word on the subject. There are still some loose ends—indeed, there
are one or two points on which the ideas could quite reasonably be challenged (and in the interest
of full disclosure, I'll call out such points explicitly, when we get to them). But I'm an optimist;
I believe those loose ends can and will be tied up in good time, and I don't believe there are any
showstoppers lurking among them. Indeed, part of my reason for wanting to publish this topic at
this time and in its present form is to solicit constructive suggestions. Certainly I'm open to
discussion in those areas where such suggestions might make sense. Though perhaps I should
stress that qualifier
constructive
... Perhaps I'm being a little defensive here, but the fact is that
I've seen quite enough negative criticism in this area in the past, and I would prefer to keep the
dialog positive.
One final point to close this chapter: The topic of view updating can unfortunately be quite
confusing, even when it's not particularly controversial. Part of the difficulty lies in the fact that
there's unavoidably a lot of detail to wade through, and it's easy to lose one's way in debates and
discussions. In particular, it's all too easy to forget, especially in examples, which relvars are
base ones and which ones are views. It's important to keep a clear head! Again I have to say:
Caveat lector
.
