The Soa Iceage

By: Max J. Pucher

'Oh, no!' you might think. 'Another one who wants to shove SOA as a means for corporate agility down my throat ...'.
Relax, nothing could be more wrong. In short, I am certain that SOA does not improve corporate agility, rather the opposite. The problem is that SOA represent a prevailing vision in the eloquent diction of Thomas Sowell's book 'The Vision of the Annointed'. A prevailing vision automatically provides a status of higher intelligence to its proponents without the need for empirical proof or more detailed analysis. Opponents simply have non-disclosed darker motives.
Service-Oriented Architecture (SOA) was first proposed by Gartner Group in 1996 in a SSA Research Document as a logical evolution of losely coupled, object-oriented messaging interfaces. But many paths lead to Rome. For example, in 2000 as enhanced with a freely definable service adapter for HTTP and FTP and in 2001 with MQ-Series. In the following years Papyrus' adapters were provided to support the most common messaging formats including SOAP and WSDL. Because Papyrus WebRepository is by design a state/event driven process engine with definable user front end, and it therefore provides the SOA core function - to create agile business services linked to encapsulated external services - since 2000.
Ok, interfaces are an important IT subject, but is it enough to invest in SOA? I am asking: "What can SOA do for the user?" The answer of the SOA visionaries: "Agile processes." I say that this is an unfounded claim with no sensible proof whatsoever. Agility is a capability of a living being that a computer infrastructure cam never posess. Agile manager and employees create agile processes with or without SOA and not the other way around.

Processes became popular in 1911 when Frederick Taylor proposed them as a means to organize a business scientifically. Rummler and Brache wrote in 1990, that "a business process is a series of steps designed to produce a product or service for a customer." This is an oversimplification that will only apply to at most two percent of all business processes. In 'Reeingineering the Corporation' Hammer and Champy enlarged on that by saying: "Not the individual task or process is important but only the outcome." That makes a lot more sense and leads to the question: "Is it even possible to analyze rigid processes to achieve a certain business goal?" I seriously propose that this is an illusion!
It is my experience that emplyoess are inerviewed for month just to figure out a few simple processes that still are not correct once he get used. I ask: ‚How are these processes then continuously improved to finally reach the desired outcome?" I tend to get pitiful smiles: "Obviously we use monitoring." Hold it, folks. Process monitoring only measures the service criteria of the analyzed process but not if it achieves the goals, right? "But we got Dashboards to monitor busines goals." Another one claims. True, some BPM products do, but even then it shows some value that maybe good or bad depending on the quality of the data source, but it DOES NOT tell you which process has to be changed how to get closer to your goal. You are back to square one. On top of that, you will find out that users will resist any change and a simple change needs a complete retest of all relates workflows because of the unknown dependencies.
Current BPM and SOA do not thaw the iceage that has befallen our applications. Java code is mostly frozen and thus dead. Business processes however have to be alive - call them agile if you want - but that has very little to do with SOA or what kind of interfacing technique you use. The people using those processes have to accept that it is them to create and modify the processes, which will never happen with any modelling tool that uses simplistic 2D step-by-step graphs, because of the hidden complexity. Only very innovative technology such as the Papyrus User-Trained Agent that can learn processes from user interaction, will provide the power of process tuning to the user.

Let's take a step back. How did we do dynamic agile processes before IT? Very simple ... by usng documents! So what we need most from BPM is an overview of all the states of all the documents of a particular business process. No one needs rigid processes that are monitored for inhumane perfection. Let me add one more: I propose that there are no ‚document related processes' but if a process does not require a document then the process is not needed! Yes, some dialogs in BPM systems replace those documents but that just proves my point. In my 35 years in IT I have learned that there are no fragmented process steps, but the propagation of a process is implicitly contained in the interdependent summary state of all its content (data, documents, controls). Because content is mostly irrelevant in process analysis so many process models are either wrong, incomplete or require substantial user input.
Another subject is even more controversial. When we interview those users, how does one model their decision making into simple IF-THEN-ELSE rules? One has to know WHY a user made a decision and encode that. I propose that this is actually not possible and my proposal is based on Antonio Damasio's research documented in his book ‚Descartes Error.' Humans are apparently incapable of being purely rational and need emotions to come to a sensible decision. (I know that this is true for me, but then I am an entrepreneur and don't have to be reasonable.) Human rationality as well as analytical fact is an illusion. We never know for sure and that's why a good feel about what is going on is much better than rational decisions. Now if that is true, each and every BPM system on the market today is a waste of time and money.

Conclusion: IT and its processes lack today because of fragmentation.
This is true for IT organization with analysis, development, test and production; missing change management for SOA interfaces, process steps, rules, documents, and GUIs; as well as the chance split of BPM, CRM and ECM. To solve the above the end-user departments have to be willing and agile enough to create and tune their own processes, but they won't succeed with 2D-graphs and rule coding. Only the Papyrus WebRepository manages all the changes from analysis, definition, and test to production, transparently across all operating systems all those process elements, including the SOA parts. A typical SOA, Java and XML application has data models and logic hidden away in the database tables, in the processes, in the Java development tools as well as in XSL, XSLT, DTD, XPATH definitions und finally in the SOA interfaces with no common change management mechanism.

Papyrus WebRepository even goes a step further. In December 2007, the User-Trained Agent will be generally available and enable the training of business processes from user interaction. The logic is much more powerful than simple rule coding because all content of the business process is considered in the transductive pattern matching learning process. The transductive training concept is patented and thus an ISIS exclusive. Yes, patenting and proprietary is good because it is innovative and functional. The time it takes to standardize kills innovation and solution orientation as the dramatically lacking BPEL 2.0 clearly shows.
Visions can't be proven (sic), but agility mostly means to have the mental flexibility to innovate.

"There are two key areas where IT has to focus: Strategic change - including architecture, business process management and change management - and innovation."
Barbara Gomolski, VP Gartner Group - Computerworld Opinion, October 2006

You can't lead following in someones footsteps. IT benchmarks pull everyone down to the same low level and have no other purpose than to be a pseudo-proof for the prevailing vision.
Innovation - to do something new - always bears some risk. Be brave!


Bibliography and References:
Damasio, Antonio (2005) Descartes' Error: Emotion, Reason, and the Human Brain, ISBN 0-380-72647-5
Draheim, D. & Weber, G. (2006) Trends in Enterprise Application Architecture, ISBN-13: 978-3540327349
Hammer, M. ; Champy, J. (1993), Reengineering the Corporation: A Manifesto ... ISBN-13: 978-1863735056
Rummler & Brache (1990), Improving Performance: How to manage the white space ... ISBN-13: 978-1555422141
Sowell, Thomas, (1996) The Vision of the Anointed ISBN-13: 978-0465089956
Taylor F. W. (1911) The Principles of Scientific Management, ISBN-13: 978-1434638205
Gartner Group SSA Research Note SPA-401-068, 12 April 1996, "'Service Oriented' Architectures, Part 1"
Gartner Group SSA Research Note SPA-401-069, 12 April 1996, "'Service Oriented' Architectures, Part 2"
E. Christensen et al., Web Services Description Language (WSDL) 1.1, (W3C) note, Mar. 2001; www.w3.org/TRwsdl
SOAP Version 1.2, World Wide Web Consortium (W3C) recommendation, June 2003; www.w3.org/TRsoap
Gomolski, B. (2006) Computerworld Opinion, Oct. 2006

Top Searches on
Technology
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 
 • 

» More on Technology