Send the link below via email or IMCopy
Present to your audienceStart remote presentation
- Invited audience members will follow you as you navigate and present
- People invited to a presentation do not need a Prezi account
- This link expires 10 minutes after you close the presentation
- A maximum of 30 users can follow your presentation
- Learn more about this feature in our knowledge base article
Do you really want to delete this prezi?
Neither you, nor the coeditors you shared it with will be able to recover it again.
Make your likes visible on Facebook?
You can change this under Settings & Account at any time.
Five Classic Data Modeling Mistakes...
Transcript of Five Classic Data Modeling Mistakes...
& How to Avoid Them
Sr Project Manager
Mistake 1: Over Modeling
Signs You may be Over Modeling:
Lots of Entities with no non-key attributes
Lots of Modeling Objects with names no business user would recognize
You have a lot of trouble coming up with names for your Entities
Subtypes that have great-great-great-great grandchildren...and grandparents
Even you can't remember where a concept is supposed to be found in the model
How To Avoid Over Modeling:
1. Get experience implementing your own models
2. Remember why you are modeling
3. Get a second expert opinion
Mistake 2: Applying Surrogate Keys Too Early
Mistake 3: Making Unplanned Changes
Mistake 4: All Flags Are Evil
Mistake 5: Not Sharing
Early Use of Surrogate Keys:
Obscures business rules
Leads to redundant relationships
You may be Applying Surrogate Keys Too Early If You:
Apply them before creating new entities
Apply them before creating relationships
Think about performance before business needs
How To Avoid Early Surrogate Keys
Don't worry about multi-part PKs until Design Time
Focus on Business Rule confirmation first, then Performance
Apply to Physical Models
You Might not See All Your Changes:
Data Models are More than Diagrams
It's all Relative
We're Human...most of us, anyway
Compare & Merge is how you see changes
Don't Use Design Standards for Requirements Standards
Be wary of the guy who rates himself an "eleven" out of "ten"
Cut & Paste is not Compare & Merge
Be wary of the "Nexter" - the modeler who just clicks "next" until the dialog goes away..
How to Avoid Unplanned Changes:
Require a Compare & Merge before check in
Require a Compare & Merge report before releases
Use external tools, such as Change Manager, too.
Never doubt that a small group of thoughtful, committed people can change the world. Indeed, it is the only thing that ever has.
All Flags are Evil?
What are the possible values?
Often avoid recording the real data
Almost always redundant
How to Avoid Flags:
Create the "real" attributes
Question every flag
Ask why they are needed
~75% of all Flag requests I receive are already derivable from existing attributes
There are only two things a child will share willingly; communicable diseases and its mother's age.
-Benjamin Spock, “Dr. Spock's Baby and Child Care”
How to Avoid Not Sharing Models:
Make them available in a variety of formats
Make them available on demand/self service
Make them searchable
Make them transportable
Make them easy to use and incorporate into other works
Publish and provide HTML Model Reports
Use ER/Studio Entprise Portal
You May be Avoiding Sharing If:
No one asks questions about data models
You throw models (or just DDL) over the wall
You don't like taping printouts together
You don't get any model feedback
1. Model Just Enough to Support Current and Probable Requirements
2. Use Business Identifiers at Requirements Time
3. Ensure All Changes are Planned
4. Avoid Flags