Explicit OR-Split [message #2438] |
Thu, 04 February 2010 21:04 ![Go to next message Go to next message](/theme/default/images/down.png) |
ljkbrost
Messages: 59 Registered: April 2006
|
Member |
|
|
Hi,
I'm playing around with the workflow and would like to emulate the following:
[Add Request] -> (Start) -> [Triage] --{status=assigned} -> (Work)
--{status=open} -> (Start)
--{status=cancel} -> (End)
Essential the Triage task will have three possible settings. If the status is 'open' send it back to the (start) place. If the status is 'assigned' send it to the (work) place. If the status is 'cancel' send it to the (end) place.
When I try to validate this I get an error message "More than 2 arcs of type 'OR-split (Explicit)' at transition Triage. Is my thinking wrong or is the system limited to 2 or-splits? If it is, how would I implement my 3 splits that I need?
Thanks,
Kyle Brost
----
|
|
|
Re: Explicit OR-Split [message #2440 is a reply to message #2438] |
Fri, 05 February 2010 11:36 ![Go to previous message Go to previous message](/theme/default/images/up.png) ![Go to next message Go to next message](/theme/default/images/down.png) |
AJM
Messages: 2361 Registered: April 2006 Location: Surrey, UK
|
Senior Member |
|
|
I have modified the system so that it will allow any number of 'OR-split (Explicit)' entries. They must all have a pre-condition except the last one (when sorted by place_name), as this will be treated as the default if none of the other conditions evaluates to TRUE.
The attached file also contains a small fix to 'std.table.class.inc' which prevented the case from being marked as closed when a token was moved to the END place.
Let me know if this works OK.
Tony Marston
http://www.tonymarston.net
http://www.radicore.org
|
|
|
|
|
Re: Explicit OR-Split [message #2447 is a reply to message #2446] |
Wed, 24 February 2010 10:41 ![Go to previous message Go to previous message](/theme/default/images/up.png) |
AJM
Messages: 2361 Registered: April 2006 Location: Surrey, UK
|
Senior Member |
|
|
There is one instance where sorting on pre_condition would not be a good idea, and where sorting just on place_name would be better. Just suppose there are several conditions which could be true, but they need to be evaluated in a particular order so that the first one found to be true is executed - it is possible that by sorting on the contents of pre_condition the wrong condition could be evaluated first. Sorting on place_name would give more flexibility as YOU would decide the order in which the conditions were evaluated. All you would have to do is prefix the descriptive place name with an ascending variable (such as '1-' or 'A-'). This would give you much more control over the evaluation order.
What do you think?
Tony Marston
http://www.tonymarston.net
http://www.radicore.org
|
|
|