Re: Record Locking using web forms ?

From: <yocca_at_ppg.com>
Date: 1996/09/13
Message-ID: <01bba183$accfae20$3352bd8d_at_Win95PC.ppg.com>

Timestamping is optimistic concurrency. So, the real question is how do you send the timestamp to the form and get it back in the where clause. If the timestamp sent from the form does not match the key in the where clause you need a method of informing the user. This is not any different from a local connection. The concept of asynchronous update is the same internet or local.

Is it not possible to feed the timestamp into a hidden field in the form?

                            \\|||//
                             |^ ^|
                             (0|0)
/------------------------oOO--( )--OOo---------------------------\


| Michael Yocca, Data Analyst, PPG Industries, Inc.              |


| yocca_at_ppg.com (work) myocca_at_sgi.net (home)                     |


\---------------------.oooO--------------------------------------/
                      (   )   Oooo.
                       \ (    (   )
                        \ )    ) /
                              ( /

disclaimer "The opinions expressed are those of the writer and not of PPG Industries, Inc. nor any PPG-associated company."

Chris Kimpton <Chris.Kimpton_at_ubs.com> wrote in article <323874F3.206B_at_ubs.com>...

> yocca_at_ppg.com wrote:
> > 
> > Wouldn't this be the same as handling concurrency in a regular
> > system?  tsequal in the where clause against a timestamp field.
> > Or is there something more to it?
> > 
> 
> ...but in a regular system, either in pl/sql - you can use 
> transactions to ensure a consistent view of the data.  
> And in forms, it automatically handles the locking/concurrency
> problems, but on the web...
> 
> You send the user a screen, he may then send you updates back.  
> In between, you cannot lock the data, since you don't know when/if 
> he'll come back.  There are, currently, no oracle facilities to help
> you automatically.
> 
> So, we're back to doing the old fashioned way - by hand.
> 
> ...unless someone has an alternative...
> 
> TTFN,
> Chris
> 

--

Received on Fri Sep 13 1996 - 00:00:00 CEST