Transcript slides - The Stanford University InfoLab

Shamelessly drawn from
“Research Principles Revealed”
Codd Award talk at SIGMOD 2007
Good Papers and Good Research
Jennifer Widom
Stanford University
My Recipe for Database Research Project
Pick a simple but fundamental assumption
underlying traditional database systems
Drop it
Must reconsider all aspects of data management
and query processing
Many Ph.D. theses
Prototype from scratch
2
Jennifer Widom
My Recipe for Database Research Project
Pick a simple but fundamental assumption
underlying traditional database systems
Drop it
Example “simple but fundamental assumptions”
 Schema declared in advance
Drop:
Semistructured data
 Persistent, relatively stable data sets
Drop:
Data streams
 Data elements contain known values
Drop:
Uncertain data
 Application provides the data
Drop:
3
Crowdsourced data
Jennifer Widom
My Recipe for Database Research Project
Must reconsider all aspects of data management
and query processing
Reconsidering “all aspects”






4
Data model
Foundations
Query language
Storage and indexing structures
Query processing and optimization
Concurrency control, recovery
Application and user interfaces
Implementation
Jennifer Widom
My Approach
To develop a new type of database system:
Data
Model
Query
Language
System
 Consider all of them
 In this order
 Solid foundations first, then implementation
5
Jennifer Widom
In Reality
Applications
Data
Model
Query
Language
System
As research evolves, always revisit all three
 Solid foundations first, then implementation
6
Jennifer Widom
Disclaimers
This “foundations-driven” approach doesn’t work
for everyone
Some prefer a system-driven approach
“My recipe for a database research project”
traditional and large-scale
7
Jennifer Widom
Disseminating Research Results
 Don’t wait
 No place for secrecy (or laziness) in research
 Every place for being first with new idea or result
• Post on web, inflict on friends
• SIGMOD/VLDB conferences are not the only
place for important work
Send to workshops, SIGMOD Record, …
8
Jennifer Widom
Disseminating Research Results
• Make software available and easy to use
 Clean, documented APIs
 Run-able over web
• Make experiments reproducible and extendible
 Provide code and data
9
Jennifer Widom
The Publication Process




10
Brutal
But we have to live with it
Unfair
(for now)
Random
Fill in your own negative adjective
Jennifer Widom
The Publication Process
My rule of thumb:
One rejection
Meaningless
Two rejections
Three rejections
But do heed
referee suggestions
Still meaningless
Maybe meaningful
Plenty of crazy stories:
Strong Reject  Best Paper Award (paper unchanged)
Desperation Venue  Top-100 CS Papers
Poor-Writing / Acceptance Correlation
11
Jennifer Widom
What’s a Researcher to Do?
1. Develop a thick skin
2. Consider alternative venues
3. Recognize that impact comes in many forms




12
Unpublished papers
Public software
Formal and informal presentations
Casual conversations
Jennifer Widom
Shamelessly drawn from
“Research Principles Revealed”
Codd Award talk at SIGMOD 2007
Good Papers and Good Research
Jennifer Widom
Stanford University