Hi folks,
We have a problem that I'm sure many other have or will come to have. Let's say in a line rule we look up some smartobject data. In most cases it returns a valid value but occassionally we sometimes get bad data back. It would be nice if the process would not act on this data because if it does, it will fail in unpredictable ways and even if we reexecute the smart oject call, we might get the correct data this time, and thus not know why the process failed or followed the wrong path.
I have been thinking of a couple of ways to mitigate this.
1. We could copy the data with a SmartObject Event Wizard to a process data field. This could then serve as an audit for the actual data value received and acted upon. I would suggest this is something for consideration in your best practices.
2. Another idea with the the coming of Inline functions in 4.5 would be to wrap the SmartObject call in a custom inline function that tests the returned data against a set of parameters.. let's say we expect the call to return a number between 1 and 100, but it returns -1.. the inline fuction would then throw an exception, the process would fall to error state, process level exception handler would notify support, suport would correct the data, the error would be retried and hopefully the process continutes successfully - result... happy business people :)
I have some more radical solutions to this problem also but they can wait for later.
We're already implementing option 1 in our processes where applicable but I thought I would throw this open for wider debate.
Ashley
