BEDL - Data 4 BPM do we need it?
This a mirror of my first post in Redux online.
IBM recently launched the proposal for a new business processes notation called BEDL - Business Entity Definition Language. To access IBM’S documents follow this link:
http://www.ibm.com/developerworks/websphere/library/techarticles/1004_nandi/1004_nandi.html
According to the team that developed the concept: “in most suites of tools for managing business process, data is treated mostly as an afterthought. The activities and their flows are the main abstractions and the data handled by the processes are essentially the hidden variables. The presentation and aggregation of data are treated outside the process definition and implemented through generic service calls. This unique approach of the process ignores the important data perspective during the analysis of business operation”.
When I started reading the article I’ve got caught by the idea of developing a standard that would help implement business processes in BPMS, representing the way data should lie down into a second layer in order to be manipulated during process execution and how to manage it.
The importance given to this issue is related to the fact that this is the fifth year since I entered in the world of BPMS. In the distant year of 2005, BPMS vendors promised for the first time the possibility to get rid of systems that oblige us to execute business processes in the way they here structured in an ERP or CRM, with the possibility to conceive the processes structured as they should be, executed without the need for shredding all the investment made in information technology until then. In this sense business processes data would continue to reside in the existing systems and could be accessed and used in BPMS by bridging existing gaps.
Well, the reality is that most BPM systems did not achieve the above promise. In some cases that was accomplished with lots of programming effort.
What is BEDL?
BEDL structure business entities, through a data information meta model for its lifecycle as well describes the possible scenarios and timings in which tasks can be invoked and executed on business entities.
Entities are represented as below. For this case I use the same entity in IBM’s article:
This meta model can be interpreted using BDEL, an XML language for describing business entities.
I am not going to lose time with the language description, yet what seems more important is I consider BEDL offers nothing new. Indeed it seems more a proprietary /custom notation rather the solution outlined in the article for the analysis and business processes data representation. For this we have already UML.
On the other hand some organizations continues to invest more in designing the tech side of business processes and data architecture, using non standard approaches, rather on to bringing some clear definition in these approaches to be used by everyone.
I think this increasing the silo effect between process teams, analysts and techies. Isn’t this contrary to BPM philosophy?
Do we need BEDL? I think not. It may be useful for anyone working with IBM's Websphere.
Finally a brief comment regarding transition states as showed in IBM’s diagram.
The transition states should be in my opinion the following:
Draft> Ready> Picked> Transit> Delivered> Done.
I do not understand how to ready state can flow to transit without being picked, or picked after being in draft state. On the other hand seems right to represent a recursive state in transit, but the same can happen in the draft, or in ready states.
IBM recently launched the proposal for a new business processes notation called BEDL - Business Entity Definition Language. To access IBM’S documents follow this link:
http://www.ibm.com/developerworks/websphere/library/techarticles/1004_nandi/1004_nandi.html
According to the team that developed the concept: “in most suites of tools for managing business process, data is treated mostly as an afterthought. The activities and their flows are the main abstractions and the data handled by the processes are essentially the hidden variables. The presentation and aggregation of data are treated outside the process definition and implemented through generic service calls. This unique approach of the process ignores the important data perspective during the analysis of business operation”.
When I started reading the article I’ve got caught by the idea of developing a standard that would help implement business processes in BPMS, representing the way data should lie down into a second layer in order to be manipulated during process execution and how to manage it.
The importance given to this issue is related to the fact that this is the fifth year since I entered in the world of BPMS. In the distant year of 2005, BPMS vendors promised for the first time the possibility to get rid of systems that oblige us to execute business processes in the way they here structured in an ERP or CRM, with the possibility to conceive the processes structured as they should be, executed without the need for shredding all the investment made in information technology until then. In this sense business processes data would continue to reside in the existing systems and could be accessed and used in BPMS by bridging existing gaps.
Well, the reality is that most BPM systems did not achieve the above promise. In some cases that was accomplished with lots of programming effort.
What is BEDL?
BEDL structure business entities, through a data information meta model for its lifecycle as well describes the possible scenarios and timings in which tasks can be invoked and executed on business entities.
Entities are represented as below. For this case I use the same entity in IBM’s article:
IBM's Courier Shipment Entity Type
This meta model can be interpreted using BDEL, an XML language for describing business entities.
I am not going to lose time with the language description, yet what seems more important is I consider BEDL offers nothing new. Indeed it seems more a proprietary /custom notation rather the solution outlined in the article for the analysis and business processes data representation. For this we have already UML.
On the other hand some organizations continues to invest more in designing the tech side of business processes and data architecture, using non standard approaches, rather on to bringing some clear definition in these approaches to be used by everyone.
I think this increasing the silo effect between process teams, analysts and techies. Isn’t this contrary to BPM philosophy?
Do we need BEDL? I think not. It may be useful for anyone working with IBM's Websphere.
Finally a brief comment regarding transition states as showed in IBM’s diagram.
The transition states should be in my opinion the following:
Draft> Ready> Picked> Transit> Delivered> Done.
I do not understand how to ready state can flow to transit without being picked, or picked after being in draft state. On the other hand seems right to represent a recursive state in transit, but the same can happen in the draft, or in ready states.
Comentários