Category: standards

Info on XPDL 2.1

I received an email last week from Nathaniel Palmer of WfMC on the upcoming XPDL 2.1 specification:

We are now in the process of developing the specification for XPDL 2.1 with the final list of changes to be completed by the next WfMC meeting on October 11-12, 2007.

It is expected that XPDL 2.1 will incorporate changes coming out of XPDL 2.0 as well as those required for BPMN 1.1 compatibility.  In addition, however, we are soliciting further input, particularly from those with experience working with XPDL 2.0 or earlier.

Interested parties are asked to please respond directly to XPDL Working Group Chair Robert Shapiro with one or both of the following:

1) Proposed Changes for XPDL 2.1 Specification – please be as specific as possible providing details about what to change in the existing specification;

2) Volunteer to Help Re-Write the Spec and Update the Schema – please  
be willing commit two or more days over the next three months to assist in this endeavor.

Please contact Robert Shapiro directly via email at [email protected]

The time for the release of XPDL 2.1 is as follow:

  • October 12 – Finalize List of Proposed Changes
  • November 15 – Finalize Details for Identified Changes in BPMN 1.1
  • December 15 – Draft Specification for Internal Review
  • January 15 – Updated Specification For Public Review
  • February 20 – Vote on Adoption of Final XPDL 2.1 Specification

XPDL will continue to be an important standard for the serialization of business processes (i.e., the file format that you use to save it, once you’ve modelled it in BPMN) for some years to come, and it still remains to be seen what impact that the new BPDM format will have on the use of XPDL.

BPM Think Tank Day 2: BPEL Roundtable

The second roundtable that I attended on last Tuesday’s sessions was on BPEL, headed up by Ismael Ghalimi. It was great to finally meet Ismael in person: we’ve been corresponding by email and blog comments for quite a while, and have even done a webinar together, but this is the first time that we’ve been in the same place at the same time.

We started with a discussion of BPEL4People, and how it’s changed from the original specification (which proposed implementing human-facing tasks as web services rather than changing BPEL) to the current specification (which proposes extensions to BPEL for human-facing tasks).

The title of the roundtable was “is BPEL relevant”, and we covered several aspects of that. First, a few people around the table (which included a few vendors with a vested interest in BPEL) stated that BPEL is relevant in the same way that SQL is relevant: as a standardized language that allows a separation of the design/development environment from the execution environment. Based on the lively discussion, some of these guys have spent a lot of time thinking about the BPEL-SQL analogy. My argument (I have no vested interest, so could have easily argued the opposite way) was that maybe it *should* be relevant in that way, but really isn’t in the consolidated model-design-execute environments that we see in BPM today. The real question may be, at what level is BPEL relevant: model, design, code or execution? Everyone agreed that it’s not relevant to business users or analysts, but it’s not clear where the line of relevance lies.

We also discussed how native BPEL execution providing code monitoring during execution, such that any code faults will have more semantic information included without having to build a monitoring stack on top of it. What remains to be seen is if BPEL4People will provide some level of business-relevant monitoring, or if that still has to be built on top of the execution layer.

What we’re seeing is that for the most part, it’s the larger vendors that are adopting BPEL — possibly as a common language to glue together all the BPM pieces that they’re acquiring — whereas the smaller vendors provided a consolidated (and therefore closed) suite environment where the execution language doesn’t matter, and in fact, their engine may be a competitive differentiator.

More on the Metastorm-Proforma acquisition

Metastorm CEO Bob Farrell gave a briefing this morning about their acquisition of Proforma; not much that wasn’t covered in the press release except the following:

  • Their customers were demanding BPA functionality that obviously went beyond what they offered as part of their BPMS.
  • They’ll be using Proforma’s CIF as an interchange format — a major mistake for process models in my opinion. Although CIF does provide support for a number of other enterprise architecture model types, they’re really going to be using this for getting process models into the Metastorm BPMS execution environment. In a post that I did at the end of the Proforma user conference last year, I noted that in spite of their lip service to standards, they only support BPEL export via a CIF remapping, XPDL wasn’t on their roadmap yet, and I’m not sure that they had started to think about BPDM. Of course, if they abandon the hope of selling ProVision to users of other BPMS’, then this may not matter much since it will just be an internal interchange standard between their own tools.

BPM Think Tank Day 2: BPMN/BPDM Roundtable

I’m just getting to the last of the BPM Think Tank sessions, namely, the roundtables and one lunch session that I had documented on paper. The three sessions of roundtables spanned Tuesday and Wednesday afternoons, and were some of the best conversations that I had at the conference. I’ll cover each of the ones that I attended in a separate post, then the summaries of the others in another post, just to keep things from getting too long. These were fairly unstructured, general sessions so the notes might be a bit fragmented

The first roundtable that I attended was BPMN and BPDM, with Stephen White of IBM and Antoine Lonjon of MEGA.

There are insufficient books and tools for educating the community on how to use BPMN for different purposes. There is a requirement for a reference document to educate end-user organizations that is smaller and more understandable than the specification (possibly both a business-oriented primer and a technical reference). Stephen stated that additional reference documents will be available within a few months. There is an HTML version of the specification online at ModelDriven.org.

Small consulting organizations and independents can’t realistically get involved in standards creation so we’re always “users” of the specification. I didn’t raise this point, but do agree with it — paying my own travel expenses and missing out on days of revenue to attend standards meetings several times each year is just not in my budget.

BPM vendors are unlikely to replace their own internal model formats with BPDM, but will translate to/from BPDM. Vendors need to review and understand BPDM and how it maps between different representations. There is a need for BPMN/BPDM conformance testing and certification of BPA/BPM products.

BPDM gives BPMN credibility as a modelling format since the specification is now “complete”. There was a great deal of discussion, both in this session and at other times during the think tank where this same point was raised, namely, that BPMN was rushed out without a serialization format, and that may have been a short-term mistake. One person at the table was concerned that combining BPMN and BPDM, and thereby increasing complexity, may be a mistake.

A comment that Phil Gilbert made on my TIBCO webinar Q&A post made a valid point about how there’s two main use cases for BPMN: non-executable process mapping and analysis by business analysts, and “visual coding” to create an executable process. We discussed this a bit at the roundtable, particularly around how business analysts could use the basic shapes (i.e., skip some of the internal graphic symbols that distinguish between different flavours of the shapes) and hence might benefit from a much simpler training program to get started. There was some discussion about how far up the chain that BPMN will or can be used for modelling businesses, e.g., whether it can be extended to strategy and goals or whether that’s more the mandate of BMM (Business Motivational Metamodel)

I had an interesting side conversation with Antoine after the roundtable ended about adoption patterns for BPMN and BPDM. Although standards organizations tend to have the “if you build it, they will come” attitude towards standards adoption, I believe that there needs to be some good reasons put forward for why BPDM provides benefits to the end customer and for the BPM vendors before we can expect to see widespread adoption.

BPM Think Tank Day 2: John Alden

Replacing the scheduled Bill Curtis (who had to cancel due to a family emergency), Chief Process Officer at McAfee, John Alden of Capability Measurement (which he co-founded with Curtis) gave the second day keynote on the role of the Chief Process Officer in business process improvement.

Responsibilities of the CPO:

  • Champion enterprise process discipline: quantify the need, articulate the vision, and infiltrate the mindset. It’s important for the CPO to be an internal evangelist for process improvement. Low process maturity organizations are spending 30-40% of their total time doing rework due to errors earlier in the process. In mid-maturity organizations, processes are stabilized but not standardized. In mature organizations, processes are standardized to reduce variability.
  • Drive process measurement: quantify business cases, support local management needs first, and mature the measures with the process over time. Derive the measurements, or KPIs, from strategic goals. Measures in immature organizations are unreliable, inconsistent and inaccurate; in mid-level maturity organizations, they become more standardized; and in mature organizations, they become strategic.
  • Coordinate enterprise process integration: represent cross-functional interests, establish enterprise process capability, and coordinate enterprise improvement projects. This requires optimizing the workflow rather than the functional performance of any given group, that is, focussing on the end-to-end process rather than silos: moving from siloed improvements, to coordinating functions through cross-functional processes, to enterprise processes that draw on functions as roles. There needs to be some sort of enterprise infrastructure to support these efforts, possibly through a centre of excellence.
  • Develop local process improvement capabilities: build business unit BPI capability, support local BPI activities, and establish enterprise improvement assets.

Alden talked a bit about BPMM, and how it needs to start at the local level and have the right type of leadership from the local managers, and finished up with a brief look at the CPO function and the related process improvement structures that are in place at McAfee.

I would have love to hear Curtis talk about this himself, since I’m sure that he’d bring more passion and hands-on knowledge about his role at McAfee to it, but Alden is very knowledgeable in this area and was a reasonable substitute.

BPM Think Tank Day 1: BPM Standards Panel

The final session of the day was a panel with representatives from four standards organizations: Fred Cummins of OMG (BPMN, BPDM), Charlton Barreto of W3C, Keith Swenson of WfMC (XPDL) and John Evdemon of OASIS (BPEL), moderated by Fred Waskiewicz of OMG.

The first question was on the focus or mission of each organization:

  • OMG has a focus on modelling business processes by business people through specifications such as BPMN and BPDM.
  • The W3C’s focus is on “keeping the web from fragmenting”, which seems a somewhat lofty goal. Relative to BPM, it’s about providing a testable architecture.
  • WfMC’s mission is to help the process design ecosystem work together better by providing process-specific standards such as XPDL. XPDL stills reigns as the de facto interchange standard for process models, with over 60 different systems and open source projects supporting it, and I can’t imagine that WfMC is going to back away from that for quite a while.
  • OASIS is involved in both orchestration, through BPEL, and choreography, through BPSS. Evdemon acknowledged some of the shortcomings with BPEL, such as the lack of support for human-facing tasks, looping and subprocesses, and discussed how they are addressing those with upcoming extensions.

So far, so polite. Everyone seems to be acknowledging everyone else’s position, and it seems like one big happy standards family. The question is, of course, is there room in the landscape for all of these? If so, how do they fit together? Is there One Language To Rule Them All?

Barroto called for more workshops to bring together the standards bodies to work out the issues of overlap and compatibility, and Evdemon agreed; I understood that that was part of the reason for the BPM Think Tanks in the first place, but they seem to be suggesting something different, like a technical skunkworks of some sort where the engineers just get together and hack it out. Since they come from the two organizations that produce much more technical standards, that’s not a surprising view, but I’m not sure that it is sufficiently sensitive to the nature of all of the types of people (including business people) who might need to be involved in the development of BPM standards.

Great question from Phil Larson of Appian: since BPMN is the only thing that everyone can agree on, why is OMG messing things up by including BPDM into BPMN, when there is still an active competition (on some level) between BPDM, XPDL and BPEL (when it’s used as an interchange standard rather than an execution language)? Phil Gilbert responded that BPMN wasn’t right in the first place since it was missing the representation part, and the inclusion of BPDM fixes that somehow. I’m not sure that answers Larson’s question, and he came back to ask if WfMC and OASIS see BPDM as a threat, and if not, why not? Swenson reinforced the big happy standards family idea that all of these can exist together, but mentioned something about migrating from XPDL to BPDM; Evdemon also said that if something better than BPEL comes along, then you should be looking at it. To me, it sounds like both WfMC and OASIS are in a wait-and-see mode to see what BPDM will turn out to be in reality, and both seem to be making noises about how if a superior standard emerges that handles all the use cases, then it would make sense to consider moving in that direction. As if they could say anything else.

That’s it for day 1. We’re all headed for the bar to see which vendors buys the first round.

BPM Think Tank Day 1: Modeling Notations/Metamodels

For the next two sessions, we’ve split into two tracks, business and technical, and I’m in the technical track where Stephen White of IBM (the “father of BPMN”) is talking about modeling notations and metamodels, namely, BPMN, BPDM and UML.

White started out by listing all of the process-related standards both within OMG, and those external to OMG, such as BPEL, XDL, ebXML BPSS (ebBP) and WS-CDL. I’m starting to think that they missed a great opportunity at lunch with the vegetable soup: a few letter-shaped noodles and we would have had alphabet soup for lunch as well as the dose of it that we’re having now. 🙂

He then focussed specifically on the three key process standards within OMG: UML, BPMN and BPDM.

UML’s been around quite a while; I know it primarily as a way to model software development concepts, and have never been happy with the attempts to shoehorn it into business analyst usages since it is difficult to explain the visual syntax of some UML diagrams to business users when they need to review them. UML activity models were added as a variation of state diagrams, and were beefed up with some business semantics to allow for use by business analysis to model business processes, but they’re not as functional as BPMN diagrams for business process models and have pretty much been replaced in that area by BPMN; I expect that they’re used mostly for modelling flows within software (by developers) than for business processes these days.

BPMI first developed BPML (an XML process execution language) which was later replaced by BPEL, and realized that a graphical notation standard was required, leading to the development of BPMN. BPMN was developed to be usable by the business community, and to be able to generate executable processes (through mapping to a language such as BPEL) by providing not just graphical elements, but the attributes for those elements. BPMN is intended to be methodology-agnostic, and to allow a business analyst to model a process in as simple or as complex an amount of detail that they deem suitable for the application.

White covered the basics of BPMN: the four core elements of activities, events, gateways and connectors represented as shapes, then variations on each of those such as the border type and thickness of an event to indicate if it is a start, intermediate or end event. I covered some of this in my recent webinar on business process modelling, or there’s tons of more detailed BPMN references around the web including Bruce Silver’s BPMN course.

This is turning into a bit too much of a BPMN primer, but I’m hanging on for the BPDM part. Also, I’m in the middle of the second row and can’t exactly sneak out unnoticed.

There seems to be a huge point of confusion with some of the audience members over pools and lanes in the swimlane elements of BPMN, and when to use a pool versus a lane; this seems like one of the most obvious things in the standard, so I’m not sure why this is a problem. Pools, in general, represent separate spheres of control; a business process starts, ends and has all of its elements within a single pool. Pools are often used to represent separate organizations in a B2B process representation. Lanes are sub-partitions of pools, usually used to indicate an organizational role or department, or even specific systems that participate in the process in some way; elements of a business process will be in different lanes of the same pools to indicate where (or by whom) that element is executed. Only message flows pass between elements in different pools, which implies a level of asynchronicity; whereas sequence flows are used to connect activities, gateways and events within the same pool, whether in the same lane or not.

BPMN 1.1 was just completed, with a few notational changes:

  • Signal, a new event type, denoted by a black triangle within the event circle shape. A signal is used to broadcast a specific state to other processes outside the immediate scope.
  • Reduction in scope for link events, because of the inclusion of the signal event; these are now basically goto events within a process.
  • Visual distinction between events that throw and catch, indicated by whether the internal icon is filled or an outline.
  • Rule event is now called “conditional”.
  • Icon for multiple events is a pentagon rather than a 6-pointed star.

Nine minutes after the session was supposed to end, we finally start on BPDM, so I think that this is going to be quick. I’m starting to understand by standards are never released on schedule…

BPDM was started in 2003 as a metamodel of business processes, initially without a notation and later aligned with BPMN. It’s intended to support the specification of multi-party choreography (think of message flows between pools) as well as process orchestration: basically, orchestration is what goes on inside a pool, that is, internal business processes; choreography is what happens between pools, that is, B2B interactions. BPMN 2.0, which will include BPDM, will update how choreography processes are modelled.

BPDM, as the metamodel, defines the meaning of the notation and provides the standardized structure behind it that allows for translation between different modelling languages. In BPMN 2.0, the meaning of BPMN will be changed to Business Process Model and Notation to indicate the inclusion of the BPDM metamodel into BPMN 2.0.

BPM Think Tank Day 1: Derek Miers

Derek Miers started the afternoon with an overview of OMG’s BPM standards.

Adopted specifications

  • BPM (Business Motivation Metamodel)
  • BPMN (Business Process Modeling Notation)
  • BPDM (Business Process Definition Metamodel)
  • SBVR (Semantics of Business Vocabulary & Rules)
  • BPMM (Business Process Maturity Model)

Specifications in progress:

  • OSM (Organizational Structure Metamodel)
  • BPRI (Business Process Runtime Interfaces)
  • BPMN 2.0 (Merged notation and metamodel)
  • PRR (Production Rules Representation)

BPMM is concerned with the evolutionary improvement path that organizations take from immature, inconsistent processes to a more mature set of business processes, and is targetted for use by the business side within organizations.

BMM is an integrated approach for deciding, documenting, communicating and managing key elements in business design, intended for use by business managers.

SVBR is more of a vendor-focussed standard for business rules: like other serialization and interchange standards, this is something that is built into products but isn’t much seen by business managers/analysts.

BPDM is an interchange format for moving between process model types: a metamodel that allows, for example, XPDL to be translated to BPDM and on to BPEL. Interesting to note that OMG puts XPDL in the same layer as BPEL, WS-CDL and ebBP, that is, execution models, with BPMN and BPDM being up in the business models layer. I’m not sure how much of this representation is political in nature, to allow BPEL and XPDL to be included in a BPDM world; since BPDM does serialization for the purposes of interchange, XPDL would only be used for interfacing with systems that don’t support BPDM, and BPEL would be used only when it was required as an execution language. This is a technical standard: business analysts won’t really care except that it provides them with interchange between tools as required.

BPMN we all pretty much know about: a graphical process notation standard, used by business analysts and others who need to create or understand process models. Version 1.1 will be released within the next few weeks, and 2.0 (which rolls in BPDM to provide notation, metamodel and interchange in a single standard) is targetted for the end of 2008.

OSM is a standard for modelling (essentially) the organizational chart: organizational units, position, authority, responsibility, relationships, contact information, organization rules, and matrix structures. A current de facto standard that covers part of this is LDAP, for example. OSM is targetted at business managers, but will need adoption by the modelling tools and BPMS vendors. There was a decision made to separate the process (BPMN/BPDM) and resource (OSM) architecture, although they are closely related: OSM will tie into BPMN concepts such as swimlanes.

BPRI is a standard providing an interface to process execution artifacts in order to extract information for monitoring and analysis. I think that this is a prime target for RSS feeds as a standard way of extracting runtime data from executing systems; that would at least provide a standard transport, leaving still the development of the standard for the payload. Derek indicated that this is one working group that is having trouble getting participation and traction in the industry, so if you’re interested in participating in this (or in any of the other OMG BPM standards), contact OMG.

On his summary slide, he has one line that says “Don’t leave it to the vendors to control your destiny”, yet the room is full of vendors with a sprinkling of independent like Bruce, Brenda and I, and I’m not sure that the end-user organizations that this is obviously targetted at are even in attendance.

OMG specification documents are freely available to the public, although I’ve always found it difficult to find them on their website. Apparently there’s a link from the OMG home page to the specifications catalog, and all of the specifications are linked from there.

BPM Think Tank Day 1: Phil Gilbert on OMG & BPM

Phil Gilbert was up next for a brief talk on the role of OMG’s BPM steering committee, which he chairs, as a lead-in to the more detailed discussions of OMG’s BPM standards to come.

He discussed how the steering committee started as BPMI.org in 2001, which released the first BPMN specification prior to BPMI being merged into OMG in 2005. The steering committee provides a platform around which input to standards can be gathered — specifically BPMN, BPDM and SBVR — but doesn’t actually vote on passing the standards. Ever since I was involved in satellite image standards back in the 80’s, I’ve never really understood how standards bodies work in their entirety, but am thankful for people like Phil who get in there and do the heavy lifting.

He covered their roadmap:

  • BPMN 1.1 (available now)
  • BPDM 1.0 (just released)
  • Merging BPMN and BPDM into BPMN 2.0, so that if you support fully support BPMN, you will support BPDM as well
  • Working on the relationships between the OMG standards and UML, XPDL and BPEL
  • OSM 1.0 for organization modeling (later this year)
  • SBVR 1.0 for rules
  • BPRI 1.0 for the runtime interface to processes

He finished with a great point: as more of these standards become baked into process-related products, the idea that you have to work with a single vendor — much less a single product — will become obsolete.

BPM Think Tank Day 1: Jim Adamczyk

The second keynote of the day was Jim Adamczyk of Accenture on how standards play a critical role in creating value with BPM. He said that they have about 40 current projects that are focussed on BPM — the discipline of creating process-centric business and IT architectures — in addition to those doing “low-level workflow”, although it’s not completely clear where the distinction lies. They’re early enough with all of these projects that he couldn’t even list client names, which means to me that Accenture is a bit late to the game here.

He moved on to talk about the value of standards, both notational and serialization, covering much the same territory as I did in a webinar recently: notational standards like BPMN allows users to move between different modelling tools more easily, and serialization/interchange standards make is easier to move processes from one system to another.

He made some great points about how changes are specified: the tendency is for business to actually specify the system change (e.g., add this function to this screen) rather than focus on their business process and KPIs — I struggle with this constantly with my customers, and have to constantly remind the business side to state their requirements, not try to design the system. The problem is that IT has been letting them do this for years, either because it’s easier to not have to learn enough about the business to do the specifications and design based on actual requirements, or because it effectively passes the buck for any mismatch between requirements (what the business needs) and specifications (what the system does) to the business side. But I digress.

Adamczyk stated that a client’s need for standards depend on their entry point to BPM, although I’m not clear what he meant by this. He said that IT almost always drives standards and that business rarely wants to implement standards; I completely disagree with this in the case of BPMN, since there are significant tangible benefits to the business side from having everyone use the same modelling notation, and I have few business-side clients who don’t recognize this. I agree that the business side doesn’t explicitly care about XPDL or BPDM or BPEL or whatever is being used for serialization, but they will start caring if it means that they can’t do round-tripping between the modelling and execution environments. However, he’s deep in the weeds talking about WSDL and LU6.2, so I think that he and I have different views of BPM standards. Since he went on to talk about how Oracle Fusion is one of the most commonly used BPM platforms amongst their client base, I think that we have different views of BPM, too.

Then he made the comment that if you use of the BPM suites (like Lombardi or Appian) that you probably don’t care about BPM standards, which couldn’t be further from the truth. Many companies use a separate process modelling tool even though they use a BPMS, so both notational standards and interchange standards are critical. They’re even important between tools from the same vendor, such as Lombardi’s Blueprint process discovery tool and their TeamWorks BPM suite, which uses BPDM for process model interchange. And there’s the advantage hiring a new analyst who already knows BPMN, even if they don’t know the particular BPMS that you’re using, because that particular standard has become so widespread, and the reduced training requirements that result.

He does have a future view for the perfect world enabled by standards that includes federated orchestration, consistent policy and governance, dynamic and predictive infrastructure, and consistent methodology/training/tools; ranging from consistency on one platform to coverage of all platforms.

Although an engaging speaker, Adamczyk seemed to spend a lot of time apologizing for things that might be missing or inaccurate in his presentation: according to him, he doesn’t really know a lot about BPM standards, nor about the utilities vertical (the industry of his unnamed client example). He also said that he’s here as a proxy for CIOs (this is billed as a “CIO keynote”) rather than as an actual CIO. Enough, already: tell us what you do know, not what you don’t know.