bpmNEXT – IBM, IMSC, camunda

Full bpmNEXT program here. The format is 30 minutes per speaker: 20 minutes of demo, 10 minutes of Q&A.

Day 1, second session

Data-Centric BPM, John Reynolds, IBM – process improv

Define attributes and rules of data elements, and create a data lifecycle model; similar to document lifecycles from products such as FileNet, but at a finer scale. When data changes, trigger the rules and lifecycle accordingly, which may create human tasks. A managed data interface (MDI) creates an integrating layer between business apps (including BPM) and the underlying connections to systems of records, which allows the full information context to be presented in the business application to support the knowledge work. Knowledge workers do unscripted work to complete tasks, within that information context, with the rules enforced; if they cause data to change, then other tasks (and users) may be notified based on trigger rules defined for those data elements. Adding data as a first class citizen in processes makes is easier to deal with unscripted work. First non-live demo of the day.

Extreme BPMN: Semantic Web Leveraging BPMN XML Serialization, Lloyd Dugan and Mohamed Keshk, IMSC US – searching for meaning in processes

BPMN modeling tool with RDF triple store, SPARQL query engine using Jena API, and BPMN-to-OWL generator for an ontological representation; based on ongoing research and being applied at US DoD and Veterans Affairs. Required the BPMN 2.0 specification for the full metamodel and serialization, but offers advantages over the XML form of a BPMN model with the addition of semantic concepts. Showed an interesting example of using a BPMN model as an architecture model for a system: collapse all pools except the system pool, the linkages in/out of the pool represent user interfaces, and the functions internally are the system specification. By translating the BPMN XML to OWL, it’s possible to do different types of queries on processes, such as identifying all services used in the model, or style conformance. The demo (a bit old skool, a text application in a command line window, and also not live) showed sample queries for finding all tasks assigned to a certain role, finding the task preceding a named task, and finding all diverging gateways. Can also be used to link process instance data with the original model to allow business process monitoring in a standards-based fashion, that is, without vendor lock-in to the specific log format and monitoring tool.

Model-BPMS Roundtripping, Jakob Freund, camunda services GmbH – a meeting of the models

Zero-code approaches to model-driven development (i.e., having a business user create their own executable process models, which can then be technically augmented by a developer) are too complicated, too restrictive, and have some proprietary aspects, making them problematic for both the business and IT participants. camunda BPM does not attempt to be a zero-coding environment, but is an embeddable framework intended to be used as part of application development. It uses BPMN 2.0 natively, which is exposed to the developer in their Eclipse environment, but the business user can use any compliant BPMN modeling tool. Traditionally, the problem with this is that if the developer changes the model (augmenting it for execution), the business user doesn’t see those changes in their model; camunda Cycle connects the business user’s BPMN modeler with the camunda Modeler for roundtripping. In Cycle, create a roundtrip and add a connection to the business BPMN model — demonstrated with the Signavio cloud modeler — to a specific model in the camunda Modeler; this does the forward-tripping to create the developer’s model in camunda from the Signavio model. The developer can now open and edit this in the Eclipse modeler, typically to edit technical attributes on activities and add service tasks. Cycle now shows that the models are out of sync since the model was changed by the developer in camunda; the model can be synchronized back to Signavio, including a commit message, which creates a new revision of the model in the Signavio repository. The business person can see the changed model (and the diff with the previous revision, since that functionality is provided by Signavio), and make further revisions. Cycle shows the last sync, as well as a thumbnail view of the process on either side.

Breaking for lunch and brain reboot. Have already seen more cool stuff in a half day than normally seen in an entire conference.

Leave a Reply