MHonArc v2.5.2 -->
ubl-ndrsc message
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [Elist Home]
Subject: Re: [ubl-ndrsc] Functional Dependency and Normalization
[I hope it's appropriate that I kept all the original recipients...]
Bill, thanks for putting this together, and thanks to the other experts
for providing us with helpful material!
I've tried to get the "gestalt view" of the notions of normalization and
dependencies by reading/skimming the Celko chapter. It would take me a
really long time to get comfortable reading SQL (I'm new at it), but
I've started to get a vague sense of the dependency requirements. One
comment below on this.
Burcham, Bill wrote:
...
> Another example is that the arguments around "data anomalies" (insert,
> update, delete anomalies) don't seem pertinent to our situation where no
> one "inserts" into a business document. Rather, we form the business
> document all at once. Same issue for updates -- we don't update a
> business document: we form it and we send it. I believe that we can
> motivate the applicability of the theory for our situation, but I
> believe that to do so we will have to rely on arguments not about data
> anomalies but about maintainability of processing code (e.g.
> stylesheets) and about specialization of the vocabulary
> (specializations, ability to apply context methodology
> <http://www.oasis-open.org/committees/ubl/cmsc/>)
...
I suspect that we need not resort to examining processing logic to
eliminate data anomalies (though maybe it's a good idea to do that too).
Rather, I bet that we can just use the "insert, update, delete" tests
mentally to figure out where we have inappropriate dependencies. This
is something I usually do intuitively with the process of modeling
directly in XML, and the experts apparently have extensive ways to
formalize this mental exercise. (Some of this seems to be shown in
Ullman/Widam's course lecture notes; I'm looking at Chapter 3's notes now.)
Finally, one request: While the Celko chapter naturally uses relational
examples throughout, it would be great if Tim's writeup could
essentially walk us through an *ABIE* version of those examples, and
show us how it changes from inappropriate to appropriate with each
anomaly removed and each dependency justified.
Eve
--
Eve Maler +1 781 442 3190
Sun Microsystems cell +1 781 883 5917
XML Web Services / Industry Initiatives eve.maler @ sun.com
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [Elist Home]
Powered by eList eXpress LLC