How Not to Build an Oracle Application

1

• How Not to Build an Oracle Application
• (Intuitive Problem Solving)
• Kerry Osborne
• Senior Oracle Guy

2
Who am I?

• I added this slide because everyone else had one!
• Started working with Oracle in 1983
• Developer and DBA and Trainer
• Became a consultant around 1987
• Started a little company, grew it, shrunk it,
  sold it
• Started another little company
• Never worked directly for Oracle
• Not certified in anything
• Dont even know what the Oak Table is?

3
Basis for this talk?

• My Totally Unscientific Survey
• Hundreds of Oracle projects
• Hundreds of Oracle professionals
• Selective memory (I dont remember all of them)
• You can observe a lot just by watching. Yogi
  Berra

4
How This Paper Came About

• So you know how delivering these papers things
  work, right?
• 4 months before the symposium you come up with a
  title
• 3 months before the symposium you write a
  paragraph (abstract)
• 1 month before the symposium the organizers
  remind you time is short
• 1 week before the symposium you decide its time
  to get serious
• 5 days before the symposium you discover that
  the paper has a taken on a life of its own
• 3 days before the symposium you discover that
  the paper doesnt really bear much of a
  resemblance to the original title

5
Tale of Two Topics

• Bad Things To Do
• Intuitive Problem Solving

6

• Part 1

7
Top 10 List - Things Not To Do

• 1. Hire Developers that Dont Know SQL
• Get good XYZ programmers let the DBAs tune the
  SQL

8
Top 10 List - Things Not To Do

• 2. Write a db access layer to isolate the
  developers
• Now your in the tool business!

9
Top 10 List - Things Not To Do

• 3. Generate a new session for every sql statement
• 10,000 traces in a single batch window

10
Top 10 List - Things Not To Do

• 4. Issue a commit after every statement
• (even better do it more often than that)
• JDBC, ODBC have a built in option to commit
  after each statement autocommit
• Log File Sync turned out to be the number one
  event at one client

11
Top 10 List - Things Not To Do

• 5. Dont allow Cost Based Optimizer to do its
  job
• Use Rule Based
• Hint everything
• Use bad statistics

12
Top 10 List - Things Not To Do

• 6. Use an unpopular tool/language
• Get one that no one has experience with
• Or better yet, write your own

13
Top 10 List - Things Not To Do

• 7. Dont use tools you already own
• Spend a lot of money on a sexy tool
• Spend a lot of time learning how to use it
• Use it for everything

14
Top 10 List - Things Not To Do

• 8. If it isnt fast enough, throw hardware at it
• I suggest starting with disk
• Then CPUs
• Then memory

15
Top 10 List - Things Not To Do

• 9. Dont Worry About Concurrency
• Oracle is really good at it
• So why should you worry

16
Top 10 List - Things Not To Do

• 10. Do everything one row at a time
• Why do something once
• When you can do it 100 times

17

• Part 2

18
Intuitive Problem Solving

• What makes a good problem solver?
• High IQ
• Hard Work
• Technical Background
• - All important, but none guarantee success

19
Problem Characteristics

• Complex
• Multiple Solutions
• Solutions vary
• Cost
• Time to Implement
• Risk

20
Basic Problem Solving Steps

• Defining the Problem
• Gathering Data
• Postulating a Reason for the Problem
• Listing Possible Solutions
• Ordering the Possible Solutions
• Attempting the Possible Solutions

21
Two Approaches

• Intuitive Methodical
• Both use the same basic steps
• Vary in time spent on steps
• Vary in order of steps
• Vary in practitioners view of steps

22
Hallmarks - Intuitive Approach

• Mental Discipline to postpone Step 3
• Willingness to quickly abandon an option
• Flexibility to re-order options

23
Hallmarks Methodical Approach

• Starts Step 3 too early
• Less creative in Step 4
• Doggedly pursues options in defined order
• Rarely changes order of options

24
Results (my opinion)

• Intuitive Approach Delivers Better Results
• More creative solutions
• Less Time

25
Improving Your Odds

• Draw a Picture
• Use Analogies
• Change the Problem
• Say it Out Loud
• Question the Necessity
• Look for the commonalities in Proposed Solutions
  eliminate them
• Take a Break

26
Simple Test
Noel
Mind Over Matter
MATTER
27
Improving Your Odds

• Longer Term Suggestions
• Collaborate
• Question the Conventional Wisdom
• Embrace the Ambiguity
• Relax

28
Negative Personality Traits

• Keep Us from Being Effective
• Ingrained Patterns of Thinking
• Common
• Easy to Spot

29
The Ive got a hammer andeverything looks like
a nail Syndrome

• Characteristic
• Using the wrong tool for the job
• Example
• SQLNet Wait Event

30
Spy vs. Spy

• Characteristic
• Sneaky / Secretive
• Example

31
Non-Gorilla Spotters

• Characteristic
• Too focused on details
• Example
• 1999 Harvard Study

32
Houdini

• Characteristic
• Blinded by assumptions
• Example
• Escape from Cell

33
Stressed Out Sam

• Characteristic
• One big twitch
• Example
• 1 minute test

34
Never say oops!

• Characteristic
• Always say Ah, interesting!
• Example
• 3M Sticky Notes

35
The Ostrich

• Characteristic
• Head in the sand
• Example

36
The Ghost and Mr. Chicken

• Characteristic
• Afraid of own shadow
• Example
• Green Bag

37
The Teenage Brain

• Characteristic
• Thinks hes immortal
• Example
• Working without a net

38
The Urban Legend Crusader

• Characteristic
• Believes unsubstantiated rumors
• Example
• Lipstick -
• w/ cancer-causing lead

39
The Monk

• Characteristic
• Extreme Compulsiveness
• Example
• Shaving 10ms off 3hr batch job

40
Mr. Magoo

• Characteristic
• Extreme Nearsightedness
• Example
• Keys in the flowerpot

41
Fiddler on the Roof

• Characteristic
• Loves Tradition
• Example
• 5 monkeys

42
Wrap up

• Problem solving is part science, part art
• I think we can improve on our natural abilities
• By being aware of the process
• By being aware of our negative tendencies
• By developing the right attitudes
• By watching others
• By practicing
• By being curious

43
References
Did You Spot the Gorilla? - Random House New
Zealand ,2004, Richard Wiseman Blink - Time
Warner Book Group, 2005, Malcom
Gladwell Optimizing Oracle Performance -
O'Reilly Associates, 2003, Cary Millsap, Jeff
Holt Creative Problem Solving A Step-by-Step
Approach - Pyrczak Publishing, 2002, Robert A.
Harris
44
Questions?
Kerry.Osborne_at_enkitec.com