peter.koch.larsen_at_gmail.com wrote:
> DA Morgan skrev:
>
>> peter.koch.larsen_at_gmail.com wrote:
>>> DA Morgan skrev:
>>>
> [snip]
>>>>>> --
>>>>>> Sybrand Bakker, Senior Oracle DBA
>>>> Actually Peter my instinct here is to agree with Sybrand. Everything
>>>> you've written flies in the face of good practice in an Oracle database.
>>> If I were to develop a new application, naturally I would take the
>>> idiosyncracies of the target platform into consideration.. Instead I am
>>> porting an existing one that runs on a wide range of platforms -
>>> including several databases of which I've already mentioned a few.
>>> This should be evident had you read the first post on this thread.
>>>
>>> /Peter
>> It is evident. It should be evident to you from my response and those of
>> others that have be involved in this thread that we are all consider
>> your products design, at least as you have presented it, as far from
>> being best practice.
>
> Right. Certainly you have been reading between the lines, making dead
> wrong assumptions about our software.
>
>> And, further, that Oracle will not change its
>> concepts and architecture to make what you are trying to do work as you
>> seemingly wish.
>
> I never had that in mind. I just hoped that Oracle had grown up and
> enabled some "a transaction really is a transaction" somewhere in its
> product, but it obviously has not.
>
>> If you want to work with Oracle. And you want to engage in best
>> practices rather than using duct tape you will need to re-examine the
>> underlying design decisions.
>
> I see you are an employee of the University of Washington. I've had a
> brief look at your Universitys homepage and see that there is a
> Database group there (interesting research, some of it that i will look
> into later).
> I will recommend that you have a talk with one of the people from that
> group (I assume you are not part of it) and ask one from the staff
> there for a motivation to have a transaction that contains
> DDL-statements not to be a transaction (at least it violates at least
> one of the ACID rules). You will find there is none - except perhaps
> for a pragmatic "it is to difficult to implement" or "it is
> historically so" argument. (Those arguments will come from Oracle, of
> course).
>
> Kind regards
> Peter
>
>> --
>> Daniel A. Morgan
>> University of Washington
>> damorgan_at_x.washington.edu
>> (replace x with u to respond)
>> Puget Sound Oracle Users Group
>> www.psoug.org
I run my own program here. Contact me off-line if you wish.
--
Daniel A. Morgan
University of Washington
damorgan_at_x.washington.edu
(replace x with u to respond)
Puget Sound Oracle Users Group
www.psoug.org
Received on Wed May 10 2006 - 18:11:21 CDT
