Re: Stupid Database Tricks

ML: Marchall
GW: Gene Wirchenko
BB: Bob Badour
RH: Roy Hann
DC: David Cressey
CD: Cimode

How about this summary (based on the above comments in the thread, these are known issues reminded in this thread)...I thought about 4 categories (I guess I am going to mount some web site soon):

CATEGORY 1: Little Design Mistakes that can be easily avoided

--> CONCATENATED KEYS VS ADDITION OF COLUMNS (CD, ML)
--> MISTAKING NUMERIC ID's (Surrogate key wannabe!) with PRIMARY KEYS
(BB, ML)
--> LETTING THE APPLICATION *HANDLE* THE BUSINESS LOGIC(BB)
--> AVOIDING SUBJECTIVITY INTO DESIGN(RH)
CATEGORY 2: Making your data administration life (under direct image systems) less miserable

--> GETTING RID OF DATA REDUNDANCY (BB)
--> TESTING YOUR BACKUPS(DC)
--> USING VIEWS(DC)
CATEGORY 3: Straightforward fallacies (Could be called Snake Oils's salesman argument)

--> DENORMALIZATION IMPROVES PERFORMANCE(DC)
--> BULLSHIT IS BEST PRACTICE (RH)
CATEGORY 4: Patterns of Behavior for Identifying the Ennemy and Defeating him (How to kick snake oil's salesman B*TT)

--> ON HOW TO DEFEAT A SNAKE OIL'S SALESMAN ARGUMENT FOR KEYS(RH)
Received on Wed May 23 2007 - 10:41:56 CEST