Path: newssvr20.news.prodigy.com!newsmst01a.news.prodigy.com!prodigy.com!news.glorb.com!newsrout1.ntli.net!news.ntli.net!newsfeed.vmunix.org!fu-berlin.de!uni-berlin.de!66.250.41.69!not-for-mail From: "Alan" Newsgroups: comp.databases.theory Subject: Re: Normalization by Composing, not just Decomposing Date: Mon, 12 Apr 2004 16:03:24 -0400 Lines: 54 Message-ID: References: NNTP-Posting-Host: 66.250.41.69 X-Trace: news.uni-berlin.de 1081800183 949144 I 66.250.41.69 ([114862]) X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2800.1158 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Xref: newssvr20.news.prodigy.com comp.databases.theory:25264 Haven't you ever heard of a joke (voodoo)? "Putting it back together" was not my concept, it was Dawn's, and I was explaining what the correct term is for what she was describing. I made no mention of lossy joins, etc. The final product (a physical implementation) is not just based on the domain being modeled, though, as you point out, it is based on a certain reality that must be considered. The implementation of a normalized/denormalized schema also considers physical performance factors. BTW, the best scientists do more than just follow rules. They show a great deal of creativity, intuition, and insight. A bit like voodoo. _You_ might be surprised... "Eric Kaun" wrote in message news:Zczec.53528$aU6.24525@newssvr16.news.prodigy.com... > "Alan" wrote in message > news:c54a0e$2ohurg$1@ID-114862.news.uni-berlin.de... > > You are assuming that (good) normalization is a science. It is not. It is > > part science and part art- that's where experience (as well as ESP to read > > the user's minds and clairvoiance to predict future needs) comes in to > play. > > Oh, it is also part voodoo. Sometimes waving a dead chicken in a paper bag > > over your head produces the results you need. > > It might not be science, but it's at least a discipline based on logic > (specifically functional dependencies). It's always going to require > interpretation with respect to the domain being modeled, because we're > trying to model part of reality, which is messy, in such a way that we (and > computers) can extract meaningful data, which requires clarity. > > That's all a far cry from voodoo, unless you're defining voodoo as > everything which is not science. And you might be surprised what real > science is like... > > > By the way, the process of > > putting it back together is called denormalization, > > Putting it back together implies that information was lost during > normalization, which isn't the case - in fact, the normalized schema doesn't > risk data loss (e.g. inconsistency) the way a denormalized schema does. > > - erk > >