Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Subject: Re: more concurrency information in the documentation
Date: Mon, 14 Mar 2005 14:15:16 -0500
Message-ID: <6EE64EF3AB31D5448D0007DD34EEB3412A763D@Herge.rcsinc.local>
Thread-Topic: [DOCS] more concurrency information in the documentation
Thread-Index: AcUovT2o03LwaUn7Q1KFqRLoXqk34wACtLUA
From: "Merlin Moncure" <merlin.moncure@rcsonline.com>
To: "Bruce Momjian" <pgman@candle.pha.pa.us>
Cc: <pgsql-docs@postgresql.org>

> Merlin Moncure wrote:
> > I was looking for some good documentation resources on dealing with
> > database concurrency issues online and I stumbled across this gem
from
> > Microsoft:
> >
> >
http://msdn.microsoft.com/library/default.asp?url=3D/library/en-us/dnbda/=
h
> > tml/concurev4M.asp
> >
> > In it is a kind of 'best practices' in dealing with various
concurrency
> > issues in the 'real world' i.e. a business environment.  Would a
> > similarly structured document covering concurrency issues be
acceptable
> > in the PostgreSQL documentation?
>=20
> I read the headings on the page and read the first line of each
section,

ok, I admit there is a fair amount of gook in there.  It's fairly laden
with meaningless or obvious terminology ('service model', etc).  However
there is some good stuff in there as well.=20

> and had very little idea what they were talking about.  What exactly
> would you like to cover?  It seems more like concurrency basics rather

1. Relative merits of various locking strategies.
2. Data conflict resolution.
3. When and why to use the different transaction isolation models.
4. How to deal with time differential between client and server.
5. How to organize your transactions such that they will not a. deadlock
b. not inadvertently invalidate one another.

The above all focused within the capabilities of the database, of
course.

> than something we would have in our docs.  Perhaps we should like to
> more outside documentation, but the URL you provided is so
content-light
> that I would not recommend it.

hm, I'm not so sure I would agree with that sentiment.  Good information
on how to deal with database concurrency issues seems to me to be pretty
hard to find.  Let me ask you this, do you think the current
documentation on database concurrency could be improved, and if so, how?
Right now, it is fairly abstract and technical, and that may be a good
thing.

Merlin