|
Critical Characteristics of Enterprise Portals Enterprise portals will become a pivotal tool of information architecture which is the framework for managing the meaning, organization and flow of information throughout an enterprise. An enterprise portal provides enhanced meaning by organizing content in context and by supporting collaborative flows of information between communities of common interest. It helps convert raw data into information and then validates that information to produce knowledge. Volumes of work have been devoted to defining the terms "data," "information" and "knowledge." For the purpose of this column, I'll offer these definitions: Data is a set of observable, measurable or calculable attributes. Data is isolated facts or observations. Alone and in the abstract, data does not provide us with information. Information is data in context. It is raw material plus conceptual commitments and interpretations. To create information, data must be extracted, filtered or formatted for presentation in some specific manner. Correlated sets of data produce greater meaning. Knowledge is a subset of information that has been subjected to and has passed validation tests. Tests of validity produce information that allows you to act. The essence of knowledge is that it is conclusive. Data is often differentiated into structured and unstructured types. Structured data is organized in a formal way (tables, columns, files and fields) to produce a database. Unstructured data is everything else including text, graphics, sound, images and combinations as found on the Internet. This distinction has always irritated me because it only serves to segregate domains that have historically been managed in radically different ways. The disciplines, and the disciples, have diverged to such an extent that they have no common language or tools. People who deal with structured data are "data people" while people who deal with unstructured data are "information people" or "knowledge people." Bah, humbug! This artificial perspective has carried over into the realm of enterprise portals. The vendors fight about the very term to preserve the distinction. "Enterprise information portals" come from data vendors, and "corporate portals" or "knowledge portals" come from the other guys. Their products do share one thing in common they are incomplete. The most essential characteristic of a true enterprise portal solution is that it manages all forms of content in a more uniform and integrated manner. Instead, I see a continuum of content that is differentiated in two more meaningful ways. The first is the definition of information as data in context and knowledge as validated information. The second is the distinction between raw and refined material. Raw material is an input and refined material is an output of an analytic or cognitive process. It is okay to think of data as raw material and information as refined material, but this is too simplistic. Anything can be raw material if it is collected by someone to initiate a decision-making or discovery process. The process of viewing, using, investigating, manipulating and/or reaching a conclusion produces refined material. This is sometimes referred to as refined results to drive home the output connotation. What is raw and what is refined is based on the eye of the beholder and the usage of the content. The most basic role of an enterprise portal is to be a delivery vehicle for diverse content. A more powerful role is to help manage the flow of the information supply chain1 of increasingly more refined results. The ultimate role is for the enterprise portal to allow the organization to provide context to generate and share information as well as correlation processes to produce validated knowledge. Content Stuff that is Organized for Access and Assimilation Content includes every type of data and information source that is used within an organization. This must encompass the whole gamut of data types such as organized databases (plain text to documents to Web pages), all AV sources (sound, video and animations) and links to content stored elsewhere. Messages in the form of e-mail, notes, forums, news and chats are a natural component of an Internet-based environment. The portal must provide assessors or viewers, at a minimum display, the content and ideally allow editing or manipulation of the material. This topic is covered later in both the connectivity and channel categories. In addition, content includes organized methods to present and deliver sets of material as a unit. These are critical forms of refined results that allow the publisher to group content that is interrelated or makes a specific case. These include reports, findings and packages all of which are terms that are given a unique meaning here. Reports are formatted presentations of material in the form of a document that may embed text, charts, diagrams and tables. Findings are the results of analysis or studies that incorporate the indicative data points, the conclusion and the rationale behind the results. Packages are sets of dissimilar data and information objects that are defined by the publisher as a single unit for delivery. Context That Which Surrounds and Gives Meaning to the Content Context is the essence of information. If you are given an isolated fact, it has no meaning. When you are told that the answer is 42, what do you know? Precisely nothing. However, when you learn that 42 is the "answer to the secret of life, the universe and everything," you now have information 2. Not terribly useful information, but information nonetheless. A portal defines context by embedding, linking or packaging related content together and by ubiquitous meta data and reference data. Embedding places content directly on a Web page. Linking is one of the defining features of the hypertext-based Internet and provides context by connecting to associated material. This can help clarify a point, explain a result or support a position. One of the most powerful features of a portal context setting is immediate and intuitive accessibility of meta data. Definitions, descriptions and examples can be one click away. Likewise, reference data can be made available to explore related topics in depth. Context is also revealed by learning what the popular topics are (top 10 hits, for example) and discovering who else has read or viewed the same content. The fact that Ralph Kimball's book, Data Warehouse Toolkit, is the number one data warehouse bestseller tells you more than an isolated review or testimonial. Learning that many other database analysts have read a particular article is likely to convince you to read it, too. Without this contextual information, you may have a harder time deciding where to begin your personal research on data warehousing. Connectivity How the Portal Gains Access to the Content A portal cannot serve the needs of information consumers if it can't access specific content they are interested in. Built-in capabilities, wrappers, application programming interfaces or external functionality can all provide connectivity. A superior portal has an extensible architecture allowing for a diverse blend of connection methods. A broad range of built-in capabilities provides more immediate capability outside the box. The native ability to open Microsoft documents or read Lotus Notes databases is preferred over being required to have third-party software installed. A full-featured portal will support all industry-standard and market-dominant application programming interfaces. APIs exist at the storage level, the application level and the business level. Storage-level APIs include, at least, all of the following: SQL, native DBMS interfaces, the newly evolving OLAP call-level interfaces, text databases and document management system APIs, as well as access to common messaging (e-mail, notes, forums, etc.) environments. Application level interfaces include access to business intelligence servers, ERP systems, other business applications, knowledge management products and other forms of application servers. Business-level APIs are a more obscure topic. They include invocation of business rule management systems, business-to-business interaction via EDI protocols and industry-specific implementations of XML to enhance semantic standardization of a business process. Channels How People Interact with the Portal Channels are the means made available to information consumers to send and receive content or to interact with the portal. Our definition is based on a communications metaphor for how information is received or transmitted.3 Receiving channels are methods that allow you to scan, fetch, order or subscribe to content. Scanning simply means to browse a set of pages on the portal looking for items of interest. Fetching requires you to go after something intentionally and includes following links and requesting downloads. Ordering is a broad concept that covers all forms of requesting specific content interactively. This includes everything from following a scripted process to guide you to a result to building an ad hoc query from scratch. All query and reporting business intelligence tools are included in the ordering feature set. Subscribing is the act of signing up for periodically released material from a specific source or on a defined topic. The result may be to have material sent directly to you or you may simply be informed when new stuff is available. Sending channels refer to all the ways that content is transmitted to information consumers. The sending channels include such methods as deliver, publish, share and refer. Delivering means designating specific content to be sent specifically to an individual or group. Publishing is a method of making content available by subscription. The publisher may not know who will receive the material, but he does define the rules for subscription. Action is taken to send the material, or a notice of availability, to the subscribers. Sharing is a method of designating material as being available to a defined population. No further action is taken by the one sharing information. Anyone who passes the qualification criteria can access the material. Accesses may or may not a tracked. Referring is an explicit action taken by one party to alert another to interesting or relevant content. The "refer this article" button on some Web sites is an example of this feature. Methods of referral include sending e- mails with embedded links or posting a link on the individual's private home page on the portal.Providing both interactive a delivery-oriented access is essential to offer a complete solution. Agents will play a major role in the future of information access. Some believe agent based processing will also play growing role in data analysis. Collaboration How the Portal Helps People Interact with People An enterprise portal must support collaboration to increase the value of the content. Collaboration comes in passive, active and autonomous forms. In all its forms, collaboration supports interaction between individuals to understand and interpret content. Passive modes include post, solicit and research. Posting material (in a forum, for instance) implies an intent to get responses from others. Everyone involved can see the entire chain of interaction. Soliciting refers to all forms of requesting something specific from the general population or an identified group. Classified ads are an example of targeted solicitation. The solicitor may be seeking specific skills or information or may be looking more generically for people with common interests. Active modes include contacting, conferring and interacting. Contacting is one-to-one communication in real time and includes traditional phone calls and Internet-based chat functionality. Conferring involves one-to- many or multiway communication that is often moderated and focused on a specific topic. Interacting is the most active form of collaboration in which multiple people can view and manipulate a work product. An example is remote CAD (computer-aided design) reviews where each participant can interact with the model to demonstrate a point or issue. Autonomous collaboration requires more intelligence to be built into the portal. It is both an exciting and scary capability of this technology. Autonomous collaboration involves the portal actively bringing together people of common interests or individuals that may be working in the same area unbeknownst to each other. This requires the portal to track and correlate activity to find patterns of affinity between the actions of portal participants. This offers enormous potential to make individuals and groups more productive. It also is a horrifically invasive process that will set off all kinds of privacy alerts. Correlation Sifting Through Content to See What is Useful Correlation implies a set of processes to help manage info- glut that helps the seeker find what is relevant. A sophisticated variant of correlation helps validate results, which is the means to the transition from information to knowledge. The subcomponents of correlation include search, sift, collect, evaluate and validate. All basic portal products will have some form of Internet-style search engine. Search engines are getting more sophisticated and more discerning in their approach. They are great at scanning the world of HTML pages. Unfortunately, inside the four walls of the enterprise, very little information is found in Web formats. An enterprise portal needs to provide generalized search methods that scan documents and databases where the bulk of enterprise content is recorded. They also need far more precision in the specification of search criteria. Sifting components help you rank the relevance of material gathered in the general sweep of a search engine. As with a search, these facilities must be more than the crude word/phrase equality and frequency test of Internet search results. There are already vendors in the market offering more sophisticated relevance and relatedness testing. Expect to see more products entering this space over the next two or three years. Collecting components allow you to automate the retrieval of relevant content once you have identified what you want. Then you can a invoke evaluating components to help you determine which material supports the case you are building. Searching finds a crude superset of material you may find interesting. Sifting helps you reduce down to the highest value material that is relevant. Evaluating looks at the relevant material to see what you can use to make a point or confirm a hypothesis. Validating, as noted, is more of a process to test information to determine if it is "true," "correct," "accepted" or "meaningful." It provides the seal of approval that allows you to add the material to your portal-accessed knowledge base. Communities Associations of People with Common Needs Allowing the establishment of multiple "communities of interest" within the portal creates a more focused and supportive environment. All organizations are a collection of teams and affinity groups, although each has widely divergent command and control systems. A portal can support both traditional vertical information flows as well as horizontal cross-functional flows. Communities can be specified, self-selected or virtual. Specified communities are established formally to support enterprise goals. They come in two forms: structural and functional. Structural communities are organizationally defined and encompass departments or work groups or teams. Functional communities include people with common roles or authorities. An example is a community designated for divisional controllers to share methods, techniques and results. You are a member of a structural or functional community based your position in the organization. Self-selected communities allow optional participation based on personal interest. For instance, you may establish a community for those people interested in supply chain dynamics. An administrator may establish self-selected communities, or the portal may present a method for individuals to propose or establish a new community. In either case, a self-selected community requires a mechanism for people to become aware they exist and to join the community. Virtual communities are based on who you are or what you do when you interact with the portal. You are not assigned to these communities and, in most cases, you do not intentionally join. In a virtual community based on characteristics, profile information about you is used to automatically relate you to an affinity group. Employees in dual-income households with young, school-aged children may be the target of a human resources-sponsored virtual community. An experienced-driven virtual community includes people with common behaviors on the portal such as people who regularly check the news feed on competitors and also tap company profiles in the market research document management system. They are the people who may be most interested when the company buys a new competitive analysis service. Customization - Adapting the Portal to Individual Needs The basic features of customization are the ability to manage the look and layout of your own home page in the portal. This includes the ability to select embeddable components such as stock tickers and calendars and to specify the sequence and positioning of sections. A more complete solution involves the ability to configure all the services of the portal. How content types are accessed and displayed may be modifiable. Features such as user-defined communities may be offered. Ideally, all modes of interaction will be customizable including configuration options for channel, collaboration and correlation features. Control - Managing the portal to suit the needs of the enterprise. Control functionality includes administrator and security features. A portal must support a wide range of control models from highly centralized through distributed to fully decentralized administration. It must provide detailed management of all usage domains including users and groups, authorized roles, default community assignments and much more. More preconfigured option sets allow more sophisticated assignment of authorities and features. The higher the degree of automation, the better the product. Security is clearly of dominate concern when you intend to offer an information-delivery vehicle that accesses all enterprise content and serves all individuals in the organization. All forms of Internet authentication and authorization services are candidates for inclusion in the enterprise portal. Access control issues are compounded when you use the portal to support customer or vendor extranets that reach way outside the four walls of the enterprise.
References 2. I apologize for the weak reference to the cult favorite," The Hitchhikers Guide to the Galaxy." This observation serves to define information as data in context through the information itself is inscrutable. 3. This use of the term "channel" is in conflict with the more typical definition that is evolving in the portal space. Many vendors use "channel" to mean what we call a category, which is just a grouping of content. Often the term "channel" takes on a "community of interest" connotation. People with shared interests either subscriber to or access a channel to see relevant content. The industry usage is ambiguous and conflicting. We choose to use the more precise terms described here. |
|||||||||||||||||||||||||||||||||
![]() |
![]() |
![]() |
![]() |
![]() |
home | about us | solutions | products | partners | clients | careers | contact us | search ![]() |
|||
|
|||
![]() |
![]() |
![]() |
![]() |