Processes need to be captured in a structured way in order to analyze them by using computer-assisted methods. This circumstance becomes more important as the process becomes complex. Although there are standardized formats, they do not capture semantics of input/output parameters, involved persons or references to external data sources. Existing solutions provide tools to capture processes locally and specify new properties to extend the semantics of process languages. However, a collaborative platform to capture, discuss and share information is more advantageous, because processes are usually used and maintained collaboratively. In addition, users cannot de ne own semantics for their use-case scenarios and the proposed semantics and processes are not published according to the Linked Data principles. To address these problems we 1) provide an open-source tool to capture BPMN processes graphically in a Semantic MediaWiki; 2) allow users to de ne own semantics and 3) publish the information according to the Linked Data principles.
Process modeling languages are mostly graphical representations, which capture work ows. However, processes must be available in a structured way in order to analyze and optimize them. Semantic information like input/output parameters, involved persons, conditions or references to external data sources can improve analyzes of processes. Although, there are standardized formats that represents the graphical representation of process modeling languages and allow to share them, they do not capture semantic information.
Existing solutions allow to capture semantics partly, however, they do not
publish the information according to the Linked Data principles1 and do not run
1 https://www.w3.org/DesignIssues/LinkedData.html
in a collaborative platform [
Processes are usually used and maintained collaboratively. Therefore, a collaborative platform, which captures processes and its semantic information, is preferable. The constantly availability and easy accessibility of the World Wide Web provides the basis for a collaborative platform, in which multiple persons can accumulate, share, discuss and edit processes, as well as information about them.
To this end, we present a graphical editor to capture Business Process Model
and Notation (BPMN) processes and semantic information about them in a
Semantic MediaWiki [
We want to use Semantic Web Technologies and already existing contributions to model, store and access information about processes for enabling people to use the information for analyzing processes.
We used Business Process Model and Notation (BPMN) as modeling language. The main reason is that BPMN is proposed as a standard by the Object Management Group (OMG) in 2008. Currently, the latest available version of BPMN is 2.0.2, published in ISO/IEC 195103. In addition, BPMN is not restricted to a speci c domain.
There are many ontologies for BPMN 2.0 available [5{7] that allow to capture
the semantics in processes and include meta-information. Among others, there
is a BPMN Ontology from the Data & Knowledge Management (DKM) research
unit4, which has a very detailed formalization of the BPMN 2.0 speci cations
in OWL 2 DL [
Once the information is modeled and structured according to existing ontologies, the process and the related information is available for di erent purposes. The graphical presentation of the process allows people, especially non-technical people, to understand the work ow of the process.
Besides this, analysts can use the available structure data and retrieve information from the process. Thereby, especially the semantics allow to retrieve information, which could not be queried before, and therefore help to analyze a process in very detail. The information can among others be used to perform statistical analyses on processes like correlation tests to indicate crucial parts 2 https://www.semantic-mediawiki.org 3 http://www.iso.org/iso/catalogue_detail.htm?csnumber=62652 4 https://dkm.fbk.eu
Title Suppressed Due to Excessive Length of the process, train models for predicting the outcome or perform similarity analysis for compare di erent processes. 3
We use a Semantic MediaWiki to store process models and meta-information and providing them as Linked Data. For modeling and annotating BPMN processes, we developed an extension to SMW, called Cognitive Process Designer5, that allows to model BPMN processes graphically in Semantic MediaWiki. Each element of the BPMN process (node and edge) has its own wiki page, and therefore unique URI, that stores information about the element. SMW publishes the information according to the Linked Data principles. Therefore, the information is published by using standard format like RDF, which can be accessed, queried and interlinked. Figure 1 shows an exemplary process, modeled with Cognitive Process Designer, and an extraction of the semantically available information.
Cognitive Process Designer provides an interface to Semantic Forms, which allows to capture and store meta-information about the BPMN process and its elements by using forms. The graphical user-interface and the provided forms make it also for non-technical users possible to enter process models and related information.
Cognitive Process Designer interacts with each modi cation of the diagram with the MediaWiki API in order to change the information on the wiki page of the corresponding element. For the communication with the MediaWiki API, we use an asynchronous communication so that users do not have to wait for the response, which makes the tool more user-friendly. Besides creating and loading existing BPMN processes in Semantic MediaWiki and modifying them, Cognitive Process Designer allows to import BPMN 2.0 XML6, which is a standardized format to represent BPMN processes. Moreover, Cognitive Process Designer allows to export BPMN processes in BPMN 2.0 XML and in SVG, which is an 5 https://www.mediawiki.org/wiki/Extension:Cognitive_Process_Designer 6 http://www.omg.org/spec/BPMN/2.0/ image format. Therefore, Cognitive Process Designer can handle proposed standard formats in order to share BPMN processes with di erent modeling tools.
The available information can be queried by using SPARQL. Thereby, structural information like the presence of deadlocks or loops can be queried, metainformation related information like the runtime of the process and number of involved persons, as well as the ratio of description for each task to see how well a task or process is described. 4
The demonstration of Cognitive Process Designer involves the modeling of a BPMN process, taken from the domain Internet of Things. During the demonstration, we will show how the tool is suitable to 1) capture BPMN processes 2) store semantic information about the BPMN elements 3) handle proposed standard formats in order to import/export process models and 4) query the entered information.