See ObjectRelationalImpedanceMismatch, ObjectRelationalImpedanceMismatchDoesNotExist, ArgumentsThatTheObjectRelationalImpedanceMismatchDoesNotExist, TablesAndObjectsAreTooDifferent, ObjectRelationalMappingCostsTimeAndMoney
I thought it is good to put here some useful links to serve as a background for the discussion. Please help me to add, probably I'll miss a lot of useful links.
Naphtali Rishe's : SemanticBinaryModel.
ChrisDate and HughDarwen: TheThirdManifesto.
ChrisDate: AnIntroductionToDatabaseSystems
FabianPascal : Practical Issues in Database Management Systems
http://www.dbdebunk.com, "The Forum Where Database Matters Are Set Straight"
Links on transaction processing
A very good book on basic concepts of concurrency control, trnsaction management and recovery is found at: http://research.microsoft.com/pubs/ccontrol/
The famous isolation levels are discussed (aside from vitually every DBMS documentation) at http://www.pmg.lcs.mit.edu/~adya/pubs/published.pdf
and
http://www.cs.umb.edu/~poneil/iso.ps
TransactionProcessingConceptsAndTechniques
On OO side of the story
KyleBrown and BruceWhitenack: CrossingChasms
ScottAmbler: Mapping Objects to Relational Databases http://www.ambysoft.com/mappingObjects.pdf. This paper is a little old and rusty and suffer from obvious flaws. A more elaborate essay you can download from http://www.theserverside.com, see the book in progress there, complex persistence chapter.
EnterpriseJavaBeans is a very good example of how things got derailed because of a psychological impedance mismatch. You can see the quasi-majority of patterns related to them are supposed to solve a thing that is broken by the specification.
RelationalWeenie Rants
http://www.geocities.com/tablizer/oorelndx.htm