processi

about processes and engines

“How to get the workitem of a newly launched flow”

This question comes often. The last occurrence came in this OpenWFE help forum thread.

But why would you need this workitem ? You already have the launchitem.

Consider that flow :

<process-definition name="yourflow" revision="0.1">
    <sequence>
    </sequence>
<process-definition>

There’s no workitem : the flow terminated immediately.

<process-definition name="yourflow" revision="0.1">
    <concurrence>
        <participant ref="manager1" />
        <participant ref="manager2" />
    </concurrence>
<process-definition>

Which workitem do you want ? The one of ‘manager1’ or the one of ‘manager2’ ?

What will you do with that workitem ? The launcher has the opportunity to fill the launchitem with values and then that workitem is sent to the worklist/engine for immediate flow launch, the work of the launcher is done. If you want him (‘it’ in some cases) to play around with the workitem, then simply add it as the target of a <participant/> expression, and the (a) workitem will come in.

This isn’t a flaw in OpenWFE’s architecture : it’s a strength. If someone understands the basic philosophy of the tool then it can make it do whatever, this idea applies to all the software packages around.

Such forum questions are exhausting, you know you’ll have to explain the world to your interlocutor and you are pretty sure that a few words won’t be sufficient. You could simply say ‘STFF’ or ‘RTFM’, but you don’t want to lose a potential user.

(side note : a way to bring down a [true] open source project is to hire people to ask lots of [semi-elaborated] questions on its forums. The people replying to the questions will get exhausted, become very rude, stop replying… The rumor will follow : “there’s no support”, “those guys are rude”, “they don’t answer to the questions”…)

Written by John Mettraux

June 21, 2006 at 8:18 am

Posted in openwfe, technical, workflow

%d bloggers like this: