Page 132 -
P. 132

3  Advanced Synchronization                                     119
                           Chapter Notes


                           The formal semantics defined in this chapter represent a second attempt at defin-
                           ing the OR-join semantics in YAWL. The original semantics of the OR-join in
                           YAWL can be found in this paper [15]. The original semantics have been found
                           to be problematic when there are multiple OR-joins in the net at the same level and
                           in composite tasks. The reasoning behind the current OR-join semantics together
                           with the formal foundations of the approach can be found in the PhD thesis of the
                           first author [272]. The work on the current OR-join semantics was first published in
                           2005 [273] and the extended version with technical details of the restriction tech-
                           niques and empirical evidence of their improved performance using several YAWL
                           nets can be found in [275].
                              An OR-join formalization is proposed for syntactically correct EPCs based on the
                           notion of state (i.e., marking of tokens) and the context (i.e., wait or dead status of
                           an arc) in [162]. The context of an input arc to an OR-join is then used to determine
                           whether an OR-join should be enabled at a given state. The semantics proposed by
                           Mendling, however, will mean that if there is a deadlock preceding an OR-join (e.g.,
                           a situation where an AND-join connector upstream cannot propagate tokens due to
                           a deadlock), the OR-join connector downstream will continue to wait for a token.
                           Additional details can be found in Mendling’s PhD thesis [160].
                              The issue of “vicious circles” in EPCs is pointed out in [9]. The proposal by
                           Kindler for a nonlocal semantics of the OR-join in EPCs can be found in [132,133].
                              More information on IBM Websphere MQ Workflow can be found in [150]. The
                           Eastman manual that was referred to in this chapter is [85].
                              The definition of the OR-join by the Workflow Management Coalition is taken
                           from [270].
                              The reader is referred to [261] for further information about BPMN and to [263]
                           for a discussion on why BPMN’s inclusive OR-join gateway does not work correctly
                           in the context of unstructured models. The specification of BPMN 1.0 by the OMG
                           is described in [185]. More on the relationship between BPMN and YAWL can be
                           found in Chap. 13.
   127   128   129   130   131   132   133   134   135   136   137