Searcher
Vol. 9 No. 6 June 2001
FEATURE
Fear, Software Integration, and Religious Wars: Internet World 2001
by Joseph Helfer Project Lead, Enterprise Integration Portals KMRS
Table of Contents Previous Issues Subscribe Now! ITI Home
In all candor, some of the key themes for this article on Internet World 2001 started percolating in my brain a couple of weeks before the event itself. I became somewhat frustrated while climbing a learning curve in the Java programming language. During a session in the smoking lounge at Sierra Cigar in Woodland Hills, California, spent trying to write some practice blocks of code and, simultaneously, to break in a new meerschaum pipe, I encountered Chris the Mathematician. Chris, who favors cigars almost as much as improving the minds of undergraduates, mentioned that he was playing around with an open source, object-oriented programming language called Python. To this news flash, I responded, "Yeah, that sounds like something of high practical value a language that nobody has ever heard of, developed and supported by a bunch of programming aficionados." In response, and also to jerk my chain a bit, Chris bemusedly asked me if I thought that anyone was actually going to show up at this year's Internet World.

Well, people did show up. In addition, this was the Internet World at which a company, Digital Creations, showed a Content Management environment underpinned by Python, thereby putting a damper on my tendency towards instant analysis and creating a certain degree of self-doubt as well: I am now avoiding Chris so that I won't have to eat crow. The reasons behind Digital Creations' choice of technology, as well as its augmented product strategy, are symptomatic of the current state of Internet-enabled products and services. This became clearer as I talked to vendors, customers, and speakers at the show itself. Although these themes may seem somewhat removed from the day-to-day responsibilities and concerns of most online searchers, I hope to be able to at least give a flavor for how some of these issues have a powerful if indirect influence on the professional searcher's current world and future challenges.

Overall, five themes jumped out at me during the exposition:

  • Customer focus on tactical implementations
  • Integration complexity
  • Convergent evolution of Internet-enabled applications
  • Scott, Lou, and Larry versus Bill (Sun, IBM, and Oracle versus Microsoft)
  • Fear and uncertainty
The vaunted merger with Internet World of the ECRM annual exposition, devoted to electronic/enterprise customer relations management, into a single event contrasted sharply with the separation of the two vendor types into separate exhibit halls. Ironically, upon observation, it seemed as if the forces of convergence were evolving both groups into the same future species.

Customers walking the exhibit floors fell into two categories the "golly gee whiz" versus the "show me the money" types. ECRM floorwalkers seemed decidedly less naïve and more systems savvy than those on the Internet World exhibit floor. Early-stage products and services, as well as more mature offerings, all struggled to differentiate themselves from each other, rather than merely riding the Internet wave, on both exhibit floors.

Conversations with Internet World Media show organizers [http://www.internetworld.com] and L.A. Convention Center employees validated that attendance had dropped, with fewer exhibitors and more last minute no-shows, such as recently extinct companies, than last year. Vendors with specialized search offerings and syndicated or filtered/personalized online database services had smaller booths and were clearly running scared and nervous. None that I chatted with had any real idea (or were willing to confess) as to the magnitude of the tasks inherent in integration of disparate information sources and types into XML and object-oriented solutions. Vendors maintained (off the record) that these issues stretched well beyond the current state of target clients' real need sets or implementation capabilities. Even the traditional dot-com booths, such as Intel and Sun, were attempting to present themselves as solutions or gravitational centers for clusters of channels partners. Conspicuously absent were IBM, Apple, and the gamut of Fortune 100 systems integrators.

Functionality was the theme of Internet World and applications solutions the currency of ECRM; although even a cursory walk through the ECRM exhibit hall revealed that the range of solutions from content management to data warehousing and business intelligence to customer support centers went far beyond the notion of CRM in vogue even 3 years ago. Nonetheless, one vendor's total solution was another's toolkit, and the result could confuse or mislead the non-initiated or novice.

From the Neolithic to the futuristic, one could see a whole range of architectures displayed at the conference. In practical terms the revolution and convergence everyone touts seem to involve bundling information into client operations through XML or some other linkage via dynamically created views. That is where the future is going and what the product architectures are designed to do. EAI (Enterprise Application Integration) or B2B (business-to-business) transactions search catalogs and ancillary data and present it to the viewer in a seamless manner. This works in parallel with traditional document management systems (which are typically focused on individual departments and a limited set of pre-defined document types a pre-Web and pre-Internet problem set) and searching of unstructured text.

Sales pitches all seemed oriented towards the customer's relative degree of ignorance. But the products and services on exhibit extended from software platforms for Web-enabled systems and integration products to e-services offering logistics support, such as UPS support of e-commerce shipping to tactical departmental offerings packaging themselves as solutions: products that deliver a limited set of functionality for a very specific problem, while still being marketed as enablers for more complex challenges involving multiple departments and external stakeholders.
 

Interview with Digital Creations
Digital Creations' CEO, Brian Methvin, and VP of sales, Paul Everett, proved quite insightful regarding both technology and business issues. With infectious smiles and the enthusiasm of missionaries, these two held my attention for an hour expounding on their open-source approach. They focused particularly on speed to market for corporations or departments chartered with implementing a content management system. Such systems are of particular interest to online searchers since these systems not only offer mechanisms for managing the creation and distribution of internally generated content (unstructured text, captured e-mail, discussion group threads, etc.), but also address capturing and managing information from external sources, the bread-and-butter vendors of the Searcher community.

Paul assured me that if I were familiar with Java, I would find the Python language a real accelerator for putting together the content management system under ZOPE [Z Object Publishing Environment; for more information, go to http://www.zope.org]. There are three concentric circles in their play the inner is Python (the object-oriented programming environment); the middle is Zope (the Python-based applications server that does the back-end work supporting the generation of Web pages); and the outer is Digital Creations itself (which creates the libraries of pre-built solutions and service offerings).

The Digital Creations system [http://www.digicool.com] lets customers go to a Web browser and create an object that creates the SQL call to a database for adding content to the Web environment a much more dynamic tool than the usual scripting effort. In this approach, a techie writes a Perl or JavaScript and embeds it into a Web page or an applications server, which constructs an SQL call routine to the database.Discussing the current avalanche of XML schema and tags and the overwhelming number of initiatives, Paul worried about confusion among customers seeking to build a content management environment.

With this approach, Digital Creations aims to put the business process owners, rather than IT techies, in control. The Zope folks at Digital Creations believe that their environment hides the syntax complexity that can accompany traditional scripting and coding and intimidate large numbers of users. Organizations or departments without large and flexible IT staffs at their disposal attempting the construction of a Web-based environment that integrates external and internal data would fall within Digital Creations' scope especially if the clients hope to create up-to-date Web pages on demand and in real time.

Digital Creations wants to enable a broader range of people to author in a seamless manner. Class libraries (the fundamental warehouses for reusable programs that techies can invoke to accelerate the systems development process) are the key here. And Digital Creations claims to have a distributed object model another techie standard that enables a more open and flexible programming environment in ZOPE as an open-source application server to build dynamic Web sites. Changes can be rolled back to prior versions of a built environment. I'll have to take their word for that since I didn't have time for a full demo, but I didn't see the feature in competitors Epicentric or Plumbtree. The most robust enterprise integration portals can support federations and syndications to automate the embedding of online database information into portals that create individual views of the world based on the portal managers profiles of user types and privileges.

How the object-oriented database software does these things, as well as its compatibility with Java Beans and other encapsulation schemes, remained unclear to me. Information professionals might have quite a time figuring out how to integrate syndicated external databases into their systems. But, for the techies out there, they were talking UML and that's a good sign. unified modeling language (UML) comprises a toolset for graphically creating and managing an IT environment using object-oriented concepts to facilitate reuse and modularity. The market leader in this space is Rational Technologies. Digital Creations' executives and I chatted about the UML as a toolset and not as a solution, per se.

With the focus on advanced rapid prototyping and JAD (joint application development) work with the customer, Digital Creations still seemed up front regarding the danger of "back-end loading" of integration projects. In this traditional ploy, the vendor-integrator submits a low initial bid that will hopefully entice the customer-client into believing that they have the "best bang for the buck." However, profit margins are built into the project in terms of hidden botches unknown to the client: ("Oh, you didn't tell us that you needed a.... By the way, you are aware of the fact that you need a ... in order to get this to work?")

The folks at Digital Creations offered some insightful remarks regarding XML and its many variants and Web service permutations. They felt that the sheer number of competing schema for data typing and standards for XML tags in various markets and disciplines creates a moving target of complexity that makes XML less a solution to the Content Management challenge, but rather an added degree of complexity for architects of solutions.
 

Interview with MicroStrategy
The 11-year-old MicroStrategy [http://www.microstrategy.com] is a prototype of the "pure play" dot-com. With a multitude of systems integrators and its own fully mobilized integration services arm, MicroStrategy represents a middle-tier Internet software enterprise. Its software falls into the decision support category. This marked MicroStrategy as clearly worth investigation, though I must admit giving away battery-powered Yo-yos that lit up and flashed dramatic red light didn't hurt their chances either.

First, consider the MicroStrategy positioning statements:

MicroStrategy is the first company to combine two essential components of CRM: analytical CRM and marketing automation. The end result of this integration is that marketers can now seamlessly analyze customer data and take action to create one-to-one relationships. MicroStrategy CRM Applications harness the most powerful customer segmentation and analysis engine on the market today to produce a complete view of the customer across the entire business.

The tide has turned. Mass marketing, with its great reach and equally great waste, is losing the efficiency game to a powerful new alternative: permission-based, one-to-one marketing. A convergence of technologies makes it possible to communicate and build relationships with millions of customers on a one-to-one basis. It's called CRM, customer relationship management.

These two quotes from the Micro Strategy product literature sum up its strategy and differentiation from a host of CRM competitors. The center of its product architecture is "MicroStrategy 7 Business Intelligence Platform." The software aims to fulfill the needs of businesses to extract relevant information from a traditional data warehouse of customer transaction-derived data. In past years, effective attempts to translate the embedded information in a data warehouse into hypotheses and insights of use to management operations required a good deal of analytic and statistical inference processing expertise. By Web-ifying this process and creating analytic wizards and a portal interface, MicroStrategy is also attempting to accelerate the process of creating, deploying, and garnering results for its core departmental users undoubtedly marketing and product management constituencies.

MicroStrategy also is attempting an enterprise-wide positioning in its literature and marketing pitches. The chatter in and around the booth included ample mention of integrating workflow and event-driven applications with the information and hypotheses generated by the Business Intelligence system, per se. Despite the chatter, however, MicroStrategy recently announced that it planned to scale back operations, reduce payroll, and stick to knitting. These days, a multiplicity of companies are disinvesting from enterprise-wide capabilities and focusing on tactical delivery. Overall, companies like MicroStrategy offer the ability to combine disparate sources of data and information, ranging from point of sales to government data to whatever, and to produce profiles that can form sources of new information for marketing campaigns and possibly product development efforts.

Informal conversations with the MicroStrategy marketing folks, as well as with other CRM firms, revealed a general frustration at the prospects on the show floor between those looking for tactical departmental solutions with immediate functionality and those looking to build a set of deliverables within a corporate architecture. These days, the goals of departmental functionality and near-term deliverables are defeating architecture and strategic integration, so to speak. Even while recognizing the strategic importance of deliverables, MicroStrategy's "big picture" diagram with interfaces to Dialog, U.S. Census data, and other syndicated services are designed to work in concert with ERP (enterprise requirements planning) packages such as SAAP, PeopleSoft, and Baan. ERP packages are large, complex systems that embed information flows into business practices and typically involve major, mission-critical IT initiatives. MicroStrategy also handles other in-house-generated data sets for marketing and manufacturing campaigns. Even MicroStrategy, like so many vendors in this arena, has scaled back from a "we can deliver this" model to a "it can be done" model, morphing from a promised solution to a suggested process. This means the vendor will put advanced features in the Requirements document but not the Statement of Work, thereby postponing the implementation until the day that George Bush (Sr. or Jr.) and Al Gore form a punk rock band together. Excuse my cynicism.

A concrete manifestation of this new reality recently hit home at MicroStrategy with particular vigor:

MicroStrategy to cut 600 workers and narrow software focus

After warning that its first-quarter financial results will fall below expectations, the struggling software vendor said it plans to lay off one-third of its employees and concentrate solely on its flagship business intelligence applications [Computerworld_Daily@computerworld.com].

The quest for the Holy Grail has been scaled back to a search for dishwasher-safe wine glasses. Compare the two quotes below to see the "quintessential" dilemma for the knowledge worker attempting to build an environment.

Today, the eBusiness Platform has become the foundation for a complete Internet infrastructure that will soon support applications as services (IDC).

But a number of corporate users expressed doubts that any of the initiatives will produce standards that meet with widespread acceptance. And until that happens, they said, companies trying to automate their supply chains will have to continue relying on a jumble of different technologies and communications channels not just the Web, but also electronic data interchange (EDI) systems, faxes and phone calls [http://www.computerworld.com/cwi/story/0,1199,NAV63_STO59259,00.html?RCKEY=0].
 

Interview with Core Intellect
Core Intellect [http://www.coreintellect.com] is a relatively new online service bureau and aggregator portal. [For background information, check the Newsbreak, "Core360 Applies Intelligent Agents to Traditional Business Data," by Barbara Quint; http://www.infotoday.com/newsbreaks/nb001204-1.htm.] Touching base with them at Internet World seemed a good way to gauge the impact of recent events on new players using data familiar to most professional searchers. With a focus on creating one's own ontology depending on the portfolio of information needs of target clients functional groups of line executives seeking an instant/portal view of relevant information Core Intellect also addresses the issues of schema, i.e., taxonomies and thesauri, and integrating enterprise data into critical business applications.

Core Intellect professed that APIs (collections of software modules that are object/component oriented) were forthcoming (late spring) for the standard enterprise-enabling packages, i.e., links to EAI software such as SAAP and PeopleSoft. Although Core has a proprietary portal software package (as opposed to reselling Plumtree or Epicentric, for example), it has built its services around the Oracle database. This should offer some investment protection and standards orientation to Core service.

When we asked how Core has differentiated itself from Northern Lights (an approach familiar to most searchers), three kinds of key differentiators were described:

  • Precision of filtering
  • Robustness of the natural language processing functionality
  • Content offerings, per se
Interview with Steve Arnold
Steve has written for Searcher often and we were most interested in his session on content management. Most of the attendees were "point people" for content management projects for individual departments or divisions of corporations. It was clear from the interaction during the Q&A session that, although their charters appeared fairly narrow at first glance, such as simplifying the intranet management problem or handling some sort of version control problem, most of the audience had significant strategic issues and IT complexity embedded within their deliverables.
Content Management Mind Map

Steve addressed the misleading nature of the " Content Management" issue with his customary insight and candor. I sketched out a mind map diagram during his talk (see "Content Management Mind Map" at left).

Steve duly noted that there appear to be something on the order of 75 content management products out on the market, and each has defined itself in a rather self-serving way. If you examine the implications of what most companies or even departments mean by a content management system and move beyond document management and version control of workers contributing unstructured text to intranets or corporate Web sites, the picture that emerges is quite intimidating strategic and enterprise-wide, rather than tactical, in scope. Consequently implementing a content management system is, typically, a 6-month project involving a team of perhaps half a dozen dedicated people or more. These projects are really equivalent to installing and implementing an ERP or EAI system and can cost on the order of $1,000,000. Vignette, for example, is marketed as such a production environment, but is actually an enterprise integration project in disguise. As can be seen in the "Mind Map" illustration, even if the originating framework is bounded (CRM, EAI, knowledge management, enterprise portal), the implicit complexity of building a content management environment is quite daunting.

About 50 percent of the audience seemed to understand the implications; the rest were technical implementers who saw their role as to install, configure, and deploy a bounded document management system, i.e., a system not architected to handle multiple data types from multiple sources nor to manage version control issues in complex networks of contributors, but, perhaps, able to handle the addition of a search engine without spending undo time or effort building a taxonomy or thesaurus.
 

Scott, Lou, and Larry Versus Bill
On the other side of the coin were the offerings from NCompass [http://www.ncompass.com], a Microsoft-architected platform.

Through the Resolution Content Connector for Commerce Server 2000, Resolution also closely integrates with Microsoft Commerce Server 2000 to deliver relevant and timely content for e-commerce applications. The Resolution Content Connector was the first product certified for use with Commerce Server 2000.
Convergent Evolution of Applications

In contrast to the J2EE (Java 2 Enterprise Edition), which enables the explicit reuse of business logic and business rules and associated platforms, such as BEA WebLogic, a premier applications server environment built on the Java platform and offering an alternative to purchasing and installing new hardware to support every product in the Microsoft catalog, NCompass took a rather proprietary approach. Sure NCompass has APIs to connect to legacy systems and other infrastructure elements. However, when you hear people talking about "COM" (no, not that "com," Microsoft's standard, called Component Object Model) rather than "DCOM" (the open standard competitor called Distributed Component Object Model), be prepared for a religious conversion and mandatory attendance at the church of Redmond, Washington.

You can pick and chose in the Sun approach, tapping multitudes of publicly accessible class libraries, JavaBeans, and other reusable software widgets. This is not the case in the Microsoft theology: One must build out the entire proprietary infrastructure and product set in order to benefit from the integrative functionality in the architecture from BizTalk server for XML functionality to SQL server, etc. Although the marketers and product managers in the Microsoft booth and its partners in the ECRM exhibit hall did not manifest a candid understanding of the implications of this, it did bear out Steve Arnold's warnings regarding a polarization between Oracle, Sun, IBM Websphere, BEA Weblogic, and DB2 on one side and the demons from Washington and .NET on the other.
 

Convergent Evolution at Work
The conference led me to conclude that the multiplicity of technological implements is at a local maximum, while the capacity of customers to both take advantage of the possibilities and change their business processes is at a local minimum. In plain language, the combination of fear and economic downturn, individual risk aversion and fear of job loss, coupled with the inherent risk of assembling the wrong combination of technologies and integrators, has led to a state of strategic inertia. Take a look at recent news releases illustrating the overreaching of product and platform companies attempting to deliver "total" or strategic solutions. Internet World 2001 provided a setting that illustrated this conflict between limited customer adoption rates and rapid technological innovation.
 

Key Insights and Suggestions for the Searcher Community
  • Many of the key Internet-enabled enterprise applications that corporations deploy have embedded within them the ability to at least link to and sometimes intelligently extract from external databases.
  • The architects and techies responsible for installing these applications will attempt to form technical and business relationships with online database service providers and may not solicit the input of the corporate librarian or searcher.
  • As opposed to the "on-demand" model of information acquisitions that professional searchers have lived with in the past, the new application paradigm has prepackaged queries and target information sources built in, so that the relevant questions are answered as the business professional performs his/her day-to-day work.
  • The thesauri and ontologies underpinning these efforts within the corporation are typically conducted without the input of the professional searcher.
  • Unless the searcher tunes in to signals that any of these applications are being rolled out, there is a strong threat that their participation will only be solicited after the fact, when it is too late to take advantage of their expertise and heed their warnings.
  • The IT folks will typically form syndications with outside databases based on ease of integration and conformance to standards, rather than quality of data or qualitative excellence and utility of the information source.
  • The new IT model is that if the user feels the need to explicitly ask for missing information, the application has failed to do its job.
  • The IT people do not have a business relationship with the online database providers and do not understand how to negotiate with them effectively.
  • Learn how to use XML-based applications and tools yourself or be prepared for relegation to the Mesozoic era and disenfranchisement from corporate, Internet-enabled application teams.


Death by a Thousand Cuts

For those not initiated into this form of technological combat, the scenario goes something like this:

  • A requirements document is constructed that represents the client's wish list.
  • A SOW (statement of work) setting the deliverables is submitted.
  • The client signs off on the various milestones of progress delineated in the SOW.
  • The client and solution provider engage in a mutual project management progress negotiation regarding the traditional "Waterfall" model (see graphic below).
In recent years many project veterans have discovered that as the project unfolds, both the customer and the solutions provider redefine the requirements as they become more knowledgeable and the original scope document becomes more a roadblock than a roadmap. The solutions provider's project manager attempts to reject customer needs and requests that are "out of scope," and the customer attempts to force the systems integrator to deliver functionality that was implicit at the time the contract was signed, but not clearly specified as a requirement for project success at the time.

The solution to this classic systems integration dilemma lies in the JAD or RAD3 model of the project: Requirements, Analysis, Design, Development, and Deployment all take place in parallel rather than sequentially. Digital Creations, for example, explicitly recognizes this reality and uses object-oriented technologies to enhance the adaptability of the time and effort to completion. This way most clients not having the coder resources and time to build out a Java or XML solution can still have their object-oriented cake and eat it too. The key is to use an object-oriented project database that allows for component re-use.

Joseph Helfer's e-mail address is helferjoe@kmrs.com.

Table of Contents Previous Issues Subscribe Now! ITI Home
© 2001