Newsgroups: comp.databases.object,comp.databases.theory,comp.databases,comp.object,comp.lang.smalltalk
Path: cantaloupe.srv.cs.cmu.edu!europa.chnt.gtegsc.com!news.mathworks.com!newsfeed.internetmci.com!news.sprintlink.net!cs.utexas.edu!milano!bromine.mcc.com!rsb
From: rsb@bromine.mcc.com (Richard S. Brice)
Subject: Re: What Are Truly Compelling Reasons To Use OODBMS ?
Message-ID: <DD3qGE.6yp@mcc.com>
Sender: news@mcc.com
Organization: MCC, Austin Texas
References: <3vcin3$rmo@news1.best.com> <7zka8qf5tq.fsf@setanta.setanta.demon.co.uk> <408is2$1von@news-s02.ny.us.ibm.net>
Date: Thu, 10 Aug 1995 15:45:50 GMT
Lines: 27
Xref: glinda.oz.cs.cmu.edu comp.databases.object:6861 comp.databases.theory:5126 comp.databases:48630 comp.object:36636 comp.lang.smalltalk:26996

In article <408is2$1von@news-s02.ny.us.ibm.net> bfelton@ibm.net writes:
>I think it is pretty easy to come up with application semantics such
>that it is benign to read data which is in the process of being changed --
>given a model that guarentees that each 'view' of the database is,
>and remains, consistent within its own boundaries, where's the problem?
>
>Bill F.
>
>


The comments above describe a current area of research in which
readers are allowed to execute concurrently with writers.   One
objective of the research is to find ways to bound the errors
the readers will get, e.g. if I'm summing up all the bank accounts
in some branch while debits and credits are ongoing, I may
not mind getting the wrong answer if I'm only looking for
an approximate sum.

As other posters have already said, there are cases where this
is unacceptable.

R. Brice
MCC Corp

_____

