John Bossert wrote:
> Let me better articulate my request, and any comments on six sigma are
> non-sequitur...
>
> I need to model an event mechanism, where tasks will come into a queue
> and be processed. My modeling experience suggests to me that the
> "right" way to represent this is with some sort of EVENT table and an
> assortment of intersection tables associating this EVENT table with the
> multitude of other entities which might be involved with an EVENT.
>
> On the other hand, my counterpart on the application side wants a table
> similar to:
>
> {
> EVENT_ID NOT NULL,
> KEY1,
> KEY2,
> ...
> }
>
> so that the _application_ can determine what goes in the EVENT table and
> how to deal with its contents.
>
> This goes against everything I know about ER design, but I'm looking for
> real-world experiences from the forum.
>
> John Bossert wrote:
>
>> I am attempting to model jobs deposited into a queue for subsequent
>> processing.
>>
>> These jobs will possess a diverse range of relationships - some jobs
>> involving customers, some with outgoing email, some with workflow.
>> Some will be handled by specific users, others will be dequeued by
>> other processes. Some events will spawn other events.
>>
>> The object-oriented (java) developers want a wide-open set of tables
>> ("just give me a table with 5 nullable numeric columns and I'll figure
>> out what goes in each") which just makes me cringe. At the same time,
>> having a spider web of intersection tables between some "event" table
>> and (many) various entities seems cumbersome.
>>
>> Could anyone having implemented such a system give me some
>> suggestions? This will be initially implemented under 8i.
>>
>> Thanks,
>>
>
>
Look into Oracle Workflow - you said yourself it's a workflow
type of app. WorkFlow can handle email (incoming and outgoing),
decisions (if this condition, the start that process, else
start another)
It comes with an API and a web interface, and -isn't it
beautiful- uses queues. There's even a modeling tool.
--
Regards, Frank van Bortel
Received on Wed Mar 19 2003 - 13:32:30 CST
