Well, I don't claim to be the world's greatest authority on
transactional theory. And I haven't yet seen the printed edition of
the column that people mentioned about MySQL vs. PostgreSQL. However,
my original manuscript did include the following paragraph:
Thanks to transactions, you can be sure that a transfer of money from
one account to another will not accidentally leave you with too much
money or with two little, even if the power fails in the middle of the
transaction. Until the transaction is finally "committed," the
database pretends that none of it has happened.
FYI, I'm expecting to review ACS and OpenACS within the next six
months; I'm going to be briefly distracted by the pluses and minuses
of n-tier systems and SOAP, and then I'll get onto a year-long
comparison of various application servers and frameworks.