[Openflow-dev] Task triggering

[<jm...@so...>]

I have been thinking a bit on how to trigger events from openflow.

In fact, openflow is a passive map of the process, and tokens are moved by
tasks.

I am asking myself if it is a virtue or a defect.

The fact is that when you don't execute the tasks, you don't have to
bother about what happens if they fail.

Is a mixt model good ?
In a kind of task (passive task) tokens are moved by the task (calling
activity.moveToken, like it happens in the example) and in other (active
task) the token is moved by an intermediate unix process which launch the
"pending" tasks (the tasks that have activated the "automatic" flag). It
could even be possible to define what to do if the task has failed (two
methods, do_the_task, undo_the_task).


Uhm, is there some plans at this respect ???



--=20

      __o
    _ \<_
   (_)/(_)

Saludos de Juli=E1n
EA4ACL
-.-