[CI] Paper theme

[Hideaki Kimura]

Samuel Madden wrote:
> - Besides index compression, do we think what we are proposing is novel 
> (e.g., are soft dependencies plus our cost model and search techniques 
> new, even if just running on unclustered indices)?
I think the query cost model aware of data correlation is new. 
CORDS/BHUNT suggested to utilize data correlation, but no detail 
analysis of query execution was given. As for actual query optimizers, 
neither PostgreSQL nor SQLServer cost model could detect the difference 
b/w PK-sort and correlated-sort.

Having a detailed query cost model is essential to integrate the 
technique into real DBMS. So, I think we can claim this as our novel 
contribution.
-- 
Hideaki Kimura <hkimura at cs.brown.edu>