|
------------------------------------- BrainStorm Bulletin e-newsletter of BrainStorm Group December 19, 2002 -------------------------------------
==================================================== --
INTRODUCTION and San Francisco
Summary
==================================================== As 2002 ends, the economic forecast for 2003 is looking a little brighter. In this issue of the newsletter we will interview Norbert Mikula of Intel. Norbert is the Web services trends and technologies person at Intel, and also serves on the WS-1 and Oasis boards. He gave a very interesting talk in San Francisco about the future of computing, web services and things that go beyond web services including the Semantic Web, a promise for more intelligent use of the Web. Read about what Norbert had to say in the full interview. Also interviewed is Jim Highsmith of the Cutter Consortium, who discusses why integration projects need agile project management. Many integration projects use leading-edge, and what Jim calls, "bleeding-edge" technologies. When you are out on the frontier of technology, traditional software development models don't work very well, and new, agile processes need to be used in order for these projects to be successful. Jim used a quote from Dee Hock, who wrote "The Birth of Chaordic Age" to illustrate the differences between agile and traditional decision making. "Simple, clear purpose and principles give rise to complex, intelligent behavior. Complex rules and regulations give rise to simple, stupid behavior." Isn't that a great quote? Jim said the important role of decision-making is not covered in most of today’s BPM work, a critical oversight. Read the detailed interview with Jim to follow. BrainStorm wishes you a wonderful holiday season and a happy and prosperous New Year.
Jon Huntress |
| ================================================= BrainStormU.com: The eLearning Center for Forward Thinking Executives ================================================= As a service to those who can not attend our Conferences due to travel and schedule constraints we have archived sessions from our events. We are pleased to bring these sessions to you for only $149 (an onsite value of $1495). They will be available to you on-demand for a full 60 days once you place your order. You may revisit the sessions as often as you like, simply log-in each time to gain access. Register today – your satisfaction is guaranteed! Use Priority Code BB1202 when you register.
|
| ================================================= FROM THE ANALYSTS: Why Integration Projects Need Agile Project Management Featuring: Jim Highsmith by Jon Huntress ================================================ Jim Highsmith is a Cutter Consortium senior consultant and author of “Adaptive Software Development: A Collaborative Approach to Managing Complex Systems,” and “Agile Software Development Ecosystems.” He has 30 years’ experience as a consultant, software developer, project manager, and writer. Jim has published dozens of articles in industry publications and his ideas about project management in the Internet era were recently featured in ComputerWorld. Jim began his recent talk BrainStorm’s San Francisco conference by summarizing how traditional software has been developed. Traditionally and today, software is created with extensive planning and architecture with limited customer involvement at the front end. The customer is told that they will get the package in x-years and everything will work. Jim thinks that most projects are still following these ad hoc rules. Today, it is more important to deliver some kind of value to the client every few months. Jim compared it to the style of the venture capitalists: Here’s some money; in a couple months show us how well it is being spent. Agile project management accomplishes this, and according to Jim, two-thirds of all corporate IT organizations will use some form of an agile software development process within the next 18 months. Agile Project Management (APM) focuses on innovation and delivering faster and better results in volatile, demanding situations. Innovation drives companies today. In 1990, it took an average of 35.5 months to bring a new product to market. Now it is 11 months. In 1990 it took six years to develop a new car from concept to production. Now it takes less than 2 years and Renault wants to get this down to nine days! Jim compared agility to the lessons of the oil fields, and the difference between exploration drilling and production drilling. Exploration drilling is an exercise in risk management, since it costs $100 million to explore a new area, and there is just a 1-in-10 chance of success. Creating a new product is similar to this. It is imperative to determine if your idea will succeed as quickly as possible to avoid a cash hemorrhage. In IT, project exploration differs depending on how far out on the curve you get. If you are using well-known or familiar technologies, this is easy. When you get into leading and bleeding-edge areas, it gets harder. Jim defined leading edge as technologies known to some but not all and bleeding edge as ideas so new that nobody knows or does them well. Any project in these areas needs extra time and flexible deadlines since you are making it up as you go. Exploration and innovation are especially important to eBusiness integration projects. Exploration type projects are managed differently and their success is measured differently as a result. Jim said that agility is defined as the ability to create and respond to change. It is characterized by the ability to balance flexibility and structure. Agility is also alert and improvisational. In planning and managing an adaptive project, three things are important: speculation, collaboration, and learning. He added that if a project goes according to a preconceived plan, it is usually an accident. Jim contrasted adaptive versus traditional predictive plans. Traditional plans are:
Adaptive plans, on the other hand are:
Jim pointed out that waiting until the eleventh month to find out that a 12-month project isn’t going to meet the changed criteria is highly ineffective. It is far better to begin the deliverables in the second month, and change them according to the needs of the users. If all you have to do to fulfill the plan is deliver documents, then you don’t have to prove anything until later in the project. Agile project management assumes that constant change is a certainty, and works with it. Jim said that HOW we make decisions on a project is very important, and yet it is virtually ignored in almost all the business process material that is currently available. Jim advised people take a close look at their organization. If it is a traditional command/control type of organization, it might not be able to be agile. Collaborative leadership is necessary. He posted a great quote by Dee Hock, who wrote “The Birth of Chaordic Age.” “Simple, clear purpose and principles give rise to complex, intelligent behavior. Complex rules and regulations give rise to simple, stupid behavior.” Jim said that what agile management attempts to do is identify a few key practices (rules) and let them evolve to meet specific problems through individual and group feedback. The steps include:
In the final analysis, eBusiness integration projects need agile project management. Here’s why:
Jim believes the large interest in agile project management is a result of the current emphasis on innovation, speed, and quality. You can hear Jim’s talk
and view his presentation at
BrainStormU.com. ============================================ Intelligent Enterprise is
the only IT publication focused exclusively For a FREE subscription,
logo onto:
www.IntelligentEnterprise.com today! Norbert Mikula is Web Services Technology and Vice Chair, WS-1: Web Services Interoperability Organization, Intel Labs, where he leads research of Web Services-related trends and technologies. Norbert is also Vice-Chairman of the WS-I Board of Directors and Vice-Chairman of the OASIS Board of Directors. An internationally recognized expert on Internet and eBusiness technologies, he often sits as conference and track chair of national and international industry conferences. Norbert began his presentation at the BrainStorm San Francisco conference with a short history of how we got where we are, starting with the first draft specification for SGML in 1986. Then he mentioned the myths, still prevalent, about XML; that it is self–describing, that XML/WbS (Web Services) will solve all the data integration problems, that it will have unlimited re-usability, and there will be no more coding. But while none of these myths have materialized, we are moving in that direction. In the evolution of computing, we have gone from mainframes to client-server, to N-tier, and to fully distributed computing, which includes all the devices such as cell phones, PDAs, and more. Now we have dynamic distributed applications that draw services from virtualized resources. We can work at the office, at home, on a plane, car, bus or train. We are entering the age of the pervasive Internet, where everything is connected to everything, everywhere, SOME of the time. Norbert mentioned that even the new John Deere tractors are heavily computerized so farmers can know the weather, soil conditions, and receive other information in the tractor cab to help them produce a crop. In order to make this work, we need:
Where this all comes together today is with Web Services (WbS). Norbert added that there are many problems with the term. Whenever it is used, nobody is sure just what is being talked about or described. He said he has been in panel discussions where time was wasted because panelists were talking about completely different things without realizing it. Web services can be simple, lightweight B2B messaging at the lowest level; the next level is a dynamic eMarketplace. At a higher level, it can enable dynamic composite applications, drawing components from dozens of different sources to create something entirely new. And, at the highest level, WbS can connect everything to everything else, everywhere, some of the time. Norbert said that beyond web services, fascinating things are beginning to happen. Intel is doing something on the leading edge of connectivity, which is the next thing beyond web services. In its Philanthropic Peer-to-Peer program, Intel is harnessing computers worldwide to accelerate the search for new treatments and drugs with the potential to cure diseases by using computer modeling to match drug candidates to binding sites on proteins. This is heavy-duty number crunching stuff, but by using PCs around the world during their down time, Intel has created a virtual supercomputer with 110+ Teraflops, to work with projects on cancer, Alzheimer’s disease, and protein folding. No centralized services are required for this, and it is adaptable and immune to single points of failure because the applications switch from computer to computer as they open and close transparently. This leads into Grid Computing, which is coordinated resource sharing. Grid computing is flexible and secure, coordinating resource sharing among dynamic collections of people, institutions, and resources. Another step beyond web services is the Semantic Web. According to Norbert’s source, W3.org, “The Semantic Web is an extension of the current web in which information is given well-defined meaning, better enabling computers and people to work in cooperation.” And what does that mean? It allows “agents” to access metadata repositories and annotated web pages, using ontologies and ontology construction tools. The concept of the Semantic Web comes from a May, 2001 article in Scientific American by Tim Berners-Lee, James Hendler, and Ora Lassila. It is based on RDF standards and is being developed by the W3C, in collaboration with researchers and industry partners. Basically it is a higher and more intelligent means of searching the Web using personalized “agents” that use standardized definitions to present much more meaningful data to the user. For example: you need a new prescription. Your “agent” would search for providers in your area, compare prices, make sure they are on your medical plan, and go into your scheduler to find the time when you could get the prescription, schedule a time, remind you when it was time to go, and provide a map to the pharmacy. Another example would be if you put a meeting in another city on your schedule; your “agent” would automatically find the airline connections and hotel information for you. The agents use ontologies to know what to look for and how to display it. Norbert said that the definition of ontology is a “formalization of a conceptualization,” which I think means that it is a highly structured search built around your particular requests and needs. At this time, we are about halfway to the Semantic Web. The W3C is working on the ontology vocabulary, RDF- (Resource Description Framework). But the logic, proof, and trust issues are still unspecified. There are challenges with the theoretical arguments around completeness, scalability, and expressiveness of RDF, and practical concerns about database extraction, text analysis, and data mining. Norbert related that he has attended meetings with people from different disciplines trying to agree on definitions for the schemas, and they generally do not. But, all the effort that is going into Semantic Web development just illustrates a major point in the evolution of computing; the constant push for personalization and increased intelligence. The question about intelligence is, “Do we have to completely and explicitly describe the Web in order to make it more intelligent?” One area that has many possibilities is location-based services such as enhanced 911, where your cell phone will also pinpoint your location in an emergency, along with letting you know the nearest doctor, gas station or Starbucks. (Norbert emphasized that not having quick access to a Starbucks can be a real emergency.) Another aspect of this is GPS, now used with the OnStar system in some cars, which will become more generally available in the future. In business, intelligence can become more process-driven. Norbert used an example of a manager who needs to approve purchase orders. If he needs 30 minutes to process a purchase order, instead of just sending him the PO, we can attach all the information he needs to decide whether to approve it, also searching his schedule to find the time for him to do it, all appearing on his desktop when needed. This kind of intelligence is hard to analyze and capture. Often the processes need to be changed to fit the existing intelligence, and exception handling is problematic. Norbert related how computers were first taught to play chess. Chess masters were interviewed about how they played, and the rules were written around what they said. Now, applied statistics and search algorithms to calculate the best moves have been created and computers are much smarter. So instead of trying to explain to computers what the rules of the world are, we can use the computers to tell us what they observe the rules to be. This is being used today at Amazon, which will inform you that many of the other people who bought the book you selected, also bought certain other books. Another example is the eBay rankings based on community standards. A problem with this increase in intelligence is the modality. The information presented to your computer can’t be presented in the same way to your cell phone. It has to be streamlined to fit the device that receives the information. The challenges ahead are:
The Interoperability standard WS-I, in which Norbert is a contributing member, was launched in February of 2002. It focuses on Web service interoperability, promotes common, clear definitions for Web services, and provides practical help for developers. Security and performance are ongoing issues in all areas of computing. Security continues to evolve. Performance problems such as the latency of http and XML being verbose and string-based can be solved with alternative binding and binary representation. But Norbert added that you must be prepared for difficulties whenever you try to talk to XML experts about the advantages of binary representation. They don’t like it. He ended his talk by saying that there is a lot of interesting research going on in all these areas now. You can hear Norbert’s talk and view his
slides (46 of them) by logging on to
BrainStormU.com. ================================================= Business Process Management (BPM) is a core component of dynamic e-business applications. Gaining a competitive advantage in the constantly changing business environment requires process agility. Evolving IT architecture requirements and standards necessitate adaptive BPM technology that grows and changes with the e-business infrastructure, leveraging existing IT investments. With Interstage i-Flow, Fujitsu Software Corporation provides a leading solution for managing dynamic business processes. Using Interstage i-Flow to manage business processes improves productivity and drives down costs, increasing profitability. Interstage i-Flow enables businesses to react quickly to customer needs, changed market requirements, and competitive challenges. That is why enterprises around the globe have chosen Fujitsu’s award-winning BPM solution to streamline and optimize their operations while providing the flexibility that today s business and technology environments demand. Fujitsu s solutions help enterprises work smarter and inspire innovation. Fujitsu Software has 11 years of experience in developing software to solve business process problems and was the first to introduce a Java-based process engine to the market. Fujitsu Software offers an unmatched level of technical support and works closely with its customers from development through deployment. To download a whitepaper from the Delphi Group on Business Process Management, go to: http://www.i-flow.com/campaign/ebusiness/registration.asp or for more information on the i-Flow solution, visit www.i-flow.com. BrainStorm Group's Conferences are the ONLY executive forums featuring:
BrainStorm Group
Conferences:
Nearshore and Offshore Outsourcing Enterprise
Content Management Conference Series Mobile Enterprise
Strategies Conference Series BrainStorm Group
Seminars: For additional
Information visit
BrainStorm Group ==================================================== The BrainStorm Bulletin is a service of the BrainStorm Group eBusiness Integration Conference Series, Northboro, Massachusetts: Sponsorship Information: Let BrainStorm EMS Assist
you in Planning Your Next Users BrainStorm Group's Event
Management Services Division will By providing both
strategic vision and an execution excellence, For our clients, we bring
to the table a set of core competencies and To submit a request for
proposal, please email details of your event ==================================================== If you have been forwarded this newsletter from someone else, please sign up to receive your own copy. Send an email to subscribe@brainstorm-group.com with the word SUBSCRIBE in the subject line. Copyright 2001 The BrainStorm Group, Inc. Rights are granted to forward this newsletter via email to others on an individual basis as long as it is forwarded in its entirety. To unsubscribe an email
to
unsubscribe@brainstorm-group.com with UNSUBSCRIBE in the subject
line.
Return to Newsletter Index |