Portal progress and KM: the Plumtree portal system

By Joseph M. Firestone

If it may be said that a single company originated the enterprise portal space, Plumtree Software is that company. Its first portal release in 1998 preceded the Merrill Lynch report defining the enterprise information portal (EIP) space, and brought to the marketplace a Yahoo-like hierarchical, navigation-enabled interface to content and data stores for both private and public enterprises.

Since then, Plumtree has implemented a vision of a portal framework integrating a variety of content, structured data and applications at the level of the portal interface (portal interface integration). In the course of pursuing its vision, it has introduced, or popularized the following innovations in portal technology: (1) personalization; (2) portal plug-ins or gadgets (called by others, widgets, portlets, etc.); (3) content and application syndication; (4) parallel processing for portal information and application delivery; and (5) multitier portal architecture for load balancing and scalability.

Distributed information management architecture

Most recently, Plumtree has emphasized an architectural framework called "The Enterprise Web," along with a strategy of creating composite applications using Web services-based integration. That emphasis is a critical departure and expansion of Plumtree's vision, because it commits the company to transcend portal interface integration and to provide for logical and process integration of distributed information and application islands in the enterprise. Significantly, it commits Plumtree to providing such integration using an open standards-compliant portal server relying on HTTP, HTML, XML and SOAP to implement Web services and comprehensive, but loosely coupled integration.

That server is complemented by a set of additional application servers providing foundation services including: a just-released content server for publishing to the portal and managing documents; a new release of Plumtree's collaboration server for project management and document sharing; Search Server for indexing and searching all content in Plumtree's directory including content and collaboration server-related content; and Plumtree Single Sign-on Server integrating Oblix's Netpoint and Plumtree's Authentication Web Services to produce identity management, security and access services. In addition, a new business process engine and an updated portal server (Plumtree Corporate Portal 5.0) that will allow creation of business processes integrating diverse data and applications into composite applications are both in development.

The Enterprise Web initiative, being implemented with Plumtree's allies Cap Gemini Ernst & Young, Documentumk and Oblix, is Plumtree's answer to so-called platform vendors who base their portal offerings on J2EE application servers. Those platform servers can only integrate among the varied resources available over the Internet developed in Java. Efforts to modify them to transcend that limitation produce non-standard proprietary servers that defeat the purposes and advantages of J2EE architecture.

On the other hand, while Plumtree's Web services approach implies some performance penalty in processing, the advantage of integration of all resources through use of Internet standards is gained. When Plumtree's new platform is completed, it promises to provide a technology layer that can use business process and workflow modeling to integrate data, content and applications into composite applications. That layer, called The Enterprise Web, may also be called a distributed information management system (DIMS)—a name much more descriptive of what the system does. (See my "Enterprise Information Portals and Knowledge Management," KMCI Press/Butterworth-Heinemann, 2003, for a discussion of requirements.)

Plumtree's delivery of an expanded portal systems framework in this form will represent a notable advance for a company fighting hard to stay ahead of the platform giants that have entered its space. And it certainly provides a substantial advance in portal capability supporting information processing and management. But where does it leave us with respect to knowledge processing and knowledge management? Will this advance, and Plumtree's portal system offering more generally, provide support for those processes and their outcomes? Knowledge processing and KM

In considering the answer to those questions, recall the knowledge processing and knowledge management frameworks presented in my column last month (see November/December 2002 KMWorld). First, nothing in The Enterprise Web provides distinctive support for knowledge claim evaluation. Although it provides generalized support for activities such as information acquisition, information broadcasting, information sharing and information search and retrieval, nothing in the portal platform or foundation services supports tracking the performance of knowledge claims in testing and evaluation.

Second, the platform provides no distinct support for distinguishing knowledge from information. If the portal platform has no way of distinguishing knowledge from information in the Distributed Organizational Knowledge Base (DOKB), how can it possibly support knowledge integration, further knowledge production or knowledge management?

Third, the platform certainly supports information acquisition through its search function. Fourth, in the area of knowledge claim formulation, support is provided by the platform through its collaborative and content sharing, and also through its categorization (search) capabilities. That support, however, is in the area of non-formal knowledge claims. To support formal knowledge claims, analytical modeling, simulation, statistical and semantic networking analysis gadgets or other Web services would need to be added to the basic platform.

Fifth, since individual and group learning sub-processes are themselves composed of knowledge life cycles (KLCs) at levels of analysis below the organization, the comments just made about information acquisition, knowledge claim formulation and knowledge claim evaluation apply here as well. Sixth, the capabilities of The Enterprise Web support information broadcasting, information search and retrieval, and information sharing, through its collaborative, content management and search services. Since the sub-processes of information integration and knowledge integration differ only in what is being integrated, the Plumtree platform would support knowledge integration if it had the capability to distinguish knowledge from information. Seventh, teaching is not supported by the basic portal services, but e-learning applications could be integrated into the platform as Web services.

My assessments of the support provided by the Plumtree platform for knowledge processing sub-processes are presented as ratings comparing Plumtree support with the ideal Enterprise Knowledge Portal in the table. I've done the ratings by splitting 100 points between the Plumtree and ideal EKP platforms in relation to each sub-process of the KLC. In future columns I'll use the same method to rate other portal products against the EKP standard.

Eighth, among the nine KM processes distinguished in the framework, none are specifically supported by platform functionality, though all are supported to the extent that they require content management, collaboration or the other foundation services. Translating that general level of support into a "split 100 points" rating, I propose a 15 (Plumtree)/85 (ideal EKP) split for all KM processes except KM knowledge production and KM knowledge integration. I rate knowledge production as 2/98 because knowledge claim evaluation is a necessary step in moving from information to knowledge, so the knowledge claim evaluation rating determines the rating for the whole. I rate knowledge integration 25/75, the average of the ratings provided earlier for the integration sub-processes because those work in parallel in the KLC model.

Finally, the assessment I've just given makes it clear that Plumtree's progress in developing its portal platform does not directly translate into support for knowledge processing and KM. That conclusion is not a criticism. Nor, as we shall see in future columns, is Plumtree atypical of portal products in its lack of support for KM. After all, Plumtree's vision is to develop a platform that will provide for comprehensive integration of data and applications through The Enterprise Web and Web services. It is not to develop an enterprise knowledge portal to provide comprehensive support for knowledge processing and KM.

Nevertheless, Plumtree's platform is what it is, and KM's needs are what they are. My task in this column is to match what vendors offer with what is needed from the viewpoint of KM as a discipline. And while Plumtree's product line offers an interesting and flexible platform for further development toward the EKP, in its generic form it does not provide specific support for knowledge processing and knowledge management.

..............................................................................................................................................................................

Joseph M. Firestone, Ph.D, is co-CEO and executive VP of the Knowledge Management Consortium International (kmci.org) and CKO of Executive Information Systems (dkms.com), e-mail eisai@comcast.net.