Re: The C in ACID

From: David Cressey <dcressey_at_verizon.net>
Date: Tue, 22 Aug 2006 11:19:50 GMT
Message-ID: <qLBGg.760$Bu2.731_at_trndny02>


"David Cressey" <dcressey_at_verizon.net> wrote in message news:t6BGg.2560$HW1.600_at_trndny03...
>
> "Paul Mansour" <paul_at_carlislegroup.com> wrote in message
> news:1156178886.704203.83220_at_75g2000cwc.googlegroups.com...
> > I'm reading a new paper by David Lomet and Roger Barga, two respected
> > (I think!) senior researchers at Microsoft, and come across the
> > following in discussing the ACID properties of a DBMS:
> >
> > "However, these techniques [Atomicity, Isolation, and Durable], which
> > are commonly known in the database technical community, do not
> > entirely cope with the problem of consistency (C), which is primarily
> > the responsibility of a user transaction to preserve."
> >
> > If the meaning of this sentence is somewhat unclear, later in the paper
> > they state:
> >
> > "Recall that it is the user, not the database system, which provides
> > the "C" in "ACID" transactions."
> >
> > ...which seems to unequivocally indicate a complete confusion between
> > the concepts of consistency and correctness. Or am I missing something?
> >
> > The paper, "Recovery from Bad User Transactions" may be found here:
> >
> > http://research.microsoft.com/db/immortaldb/
> >
> > Scroll down to the bottom, its the first reference.
> >
>
> Quoting from the abstract:
>
> "However, user transactions can be flawed and lead to inconsistent (or
> invalid) states."
>
>
> It's unlikely that the authors conflate consistency with validity given
the
> above. I presume validity and correctness are both the same concept,
> namely consistency with the real world. Given the thrust of the article,
> concerned with correction of invalid states, the parentheses in the above
> quote were unfortunate, at best. If anything, it should have been
written:
>
> "However, user transactions can be flawed and lead to invalid (or
> inconsistent) states."
>
> This would, I claim, have more clearly indicated the authors' thought.
>
>

After reading other comments, I've decided that my correction was itself incorrect. Better would have been:

"However, user transactions can be flawed and lead to invalid (and possibly inconsistent) states." Received on Tue Aug 22 2006 - 13:19:50 CEST

Original text of this message