Re: normalisation vs denormalisation

George Dau (gedau_at_mim.com.au) wrote:
: s9200765_at_westgate.vut.edu.au (Arshad Humayun) wrote:
 

: ]
 

: I have a bit more to say on this. Our academic design types all want 3rd normal
: form. Anyone who has tried to actually implement a system will want some
: redundancy - for performance or just for convienence.
 

: It is slow to have to re-calculate a total every time an order is displayed.
: You pay a bit for this when you change an order, but not many are changed.
 

: It is slow to have one record for each month, instead of 12 repeating fields in
: a record. OK, repeating fields can be a problem, but really, are the number of
: months in a year going to change over the 5 year life of this app?
 

: When it comes to maintenance, we wish we had stuck to normalized structures, and
: gotten them to buy better hardware.

I dont know if i get your point... but what you say strikes a nerve. I have implemented several large systems. I find that it is ALWAYS better to get as close to 3rd normal form as possible. I find it irritating to hear the objectionable phrase 'academic types'. This somehow implies that this thought (or process or methodology) is wrong by somehow being removed from practicality. What is the alternative? Use the ideas of the non-academic types? What are they? Uneducated?

Why do you assume that people who proclaim the benefits of 3rd nf have never implemented a system? Have you ever built a large successful database? If your apps aren't making it to 5 years, then maybe you should rethink your strategy.

Please don't poo-poo a elegant design methodolgy because you don't understand it or have never gotten it to work. There are many many instances where it does work, and work well. (better than the old non-normalized ways that it replaces). My opinion is that people who make comments like yours have never seen a good design much less ever came up with one. Keep trying...

--
    ..uu.                                     ----------------------
  .?$" '?i     .                              I  Randy DeWoolfson  I
 .T^M  ._at_"    d9    .     f   ,.un.  b,    i  I--------------------I
 "  Z :#"    M `8   U    <  .dP"``"# `M   _at_"  I  randyd_at_cais.com   I
    &H?`    Xl _R   $5.  $  ?*    _at_   'P,#"   I--------------------I
  ,d#^*L   :RP'~$b  f`$L:M  Xf  .f'    dH`    I        ,\//.       I
    &  'M ,P    `E  M   "$  Mux~      n!`     I        |o o|       I
   dk   `h"       ' j     " y"       *~       I====oOO==(_)==Ooo===I