|Steven L. Telleen, Ph.D.
Director, IntraNet Solutions
Copyright (c) 1995 All Rights Reserved
IntraNet is a trademark and Amdahl is a registered trademark of Amdahl Corporation.
The Internet is based on a set of protocols developed to allow these distributed networks to route and pass information to each other independently, so if one site is down, the information can be routed through alternate sites to its final destination. The protocol developed for this purpose was called the Internetworking Protocol or "IP" for short. When you see the acronym TCP/IP, it is this same Internetworking Protocol that is being referenced in the second part of the acronym.
The IP protocol came into widespread use in the military community as a way for researchers to share their computerized information. Since the military had numerous research projects underway at universities around the country, and it provided an effective way to move information across diverse networks, the protocol quickly spread outside the defense community. It also spread into NATO research institutions and European universities. Today the IP protocol, and hence the Internet, is a ubiquitous standard World Wide.
By the late 1980s a new problem had emerged on the Internet. In the early days information consisted primarily of mail or traditional computer data files. Various mail protocols and file transfer protocols had evolved to handle these requirements. However, newer types of files were beginning to emerge on the Internet. These were multimedia files, that contained not only pictures or sound, but also hyperlinks that allowed users to jump around inside files in a non-linear way, or even jump to other files containing related information.
In 1989, the European Particle Physics Laboratory, CERN, began a very successful internal project that led to an effort to create standards for passing this new kind of information around the Internet. The basic components consisted of a standard for creating the multimedia, hypertext files, and a standard for serving these standardized files when they were requested. The file standard is called the HyperText Markup Language or HTML. It is actually a much simplified subset of another standard called the Standard General Markup Language or SGML. The server standard is called the HyperText Transfer Protocol or HTTP. A server running the HTTP demon will send HTML files over the Internet to clients requesting them.
These two standards provided the basis for a whole new kind of access to computerized information. Creating multimedia files in a standard way allows client software to be built that not only can retrieve the files from an HTTP server, but also open them and display them as part of the request. And, since the file can contain hyperlinks to other files (even when they reside on other computers), a user now has the ability to navigate information with a point-and-click interface from what looks to them like standard printed documents. This technology takes away the complexity of accessing information on distributed computers.
The multimedia files served and retrieved in this way are commonly called "pages" by those using this technology. The information sent to the client machine for each request is known as a page. This is because the information that normally would be a document on paper often is broken into smaller units that are hyperlinked together. This separation gives the user flexibility in deciding what she wants to see, and saves time and network bandwidth by not transferring a lot of information that is not of interest.
Most browsers can be set to automatically retrieve a specific page when they are started. Usually this page contains an index hyperlinked to other pages the user retrieves frequently. This page is called the home page, and if the browser has a button called "home page" this is the page it will return to when the button is activated. The hyperlink documents mentioned in the previous paragraph generally have a page that is the equivalent of a table of contents, or at least provides a common starting point for the set of information. This page is called a home page too. In this context, the term home page is relative and can be used for any index page that provides a starting point into a specific set of information. These relative home pages generally are preceded by the name of the set of information they are representing, e.g. the Amdahl home page.
Like the Internet, on which all this operates, a specific page might be found and accessed via several different routes. This global collection of pages and hyperlinks on the Internet is known as the World Wide Web (or WWW or W3). It is completely distributed, and an author (or publisher) of generally available information has no way of knowing about all the index pages that may have pointers linking people back to her page. It is important to note here that the server of that page can be set up to keep track of who accesses that page, just not what other pages are pointing to it. For those of us used to paper media this is almost a reversal of the kind of information we had available. In many research fields in the paper world there are periodical indexes that list all the articles referencing a specific article, but there is no way to tell when someone reads, that article. With this medium we can tell who is accessing (reading) the page, but not who is referencing it.
The other side to this is that it has become impossible to keep track of all the information available on the Web. New information appears and disappears regularly, independent of any central management. However, before we become too alarmed by this fact, we should consider that this is no different than the world of paper and print. If we find this problematic, then it probably comes from restricting our view to the sheltered world of computers where, in the past, information was so hard to get in and out that we had a clear gate and a limited set of information to deal with. This technology widens that gate and makes electronic information as easy to create and as easy to access as print information.
Browsers exist for most major client systems that support intelligent windows. These include MS/Windows, Macintosh, X-Window systems, and OS/2. There also are browsers for non-windowing systems that can display the text portions of the accessed documents. The availability of browsers across these diverse platforms is significant. The operating environments of the author, the server, and the client viewer are independent of one another. Documents built using the HTML and related standards and served by an HTTP server can be accessed and viewed by any client regardless of the operating environment on which they were created or from which they were served.
HTML also supports forms development and return functionality. This means that the user interface extends beyond point and click to both query and data input. A number of sites, including Amdahl, have written interfaces between HTML forms and legacy applications, creating a universal client front-end to those applications. This opens the possibility of writing client-server applications where the author does not have to be concerned with client-side coding. In fact, new applications already are emerging where the client is assumed to be a browser, Oracle's WOW interface which substitutes for Oracle Forms and Oracle Reports is an example.
While this technology is still very young, it appears to be having an analogous effect on information management to what the invention of solid state electronics and microprocessors had on computer hardware. It is modularizing functions and simplifying applications allowing us to move to a new level of integration that is closer to the business functions of the enterprise.
The most hyped early application for the Web was as a new channel for selling directly to customers. While a few companies are beginning to realize revenues directly from Web sales, at this writing, the majority of business uses still revolve around activities that do not involve the direct transfer of money (or credit). Many companies market their products and services on a home page, much like they might advertise in a magazine. Software patches and support information also can be dispensed successfully today.
One of the most successful business uses of the Web today is supporting informational queries that otherwise would have been handled by a customer services representative. These range from checking personal bank account balances and recent transactions via an encrypted transaction (Wells Fargo Bank), to checking the location and status of specific overnight packages (FedEx), to finding job opportunities in the State of Florida, to identifying the times and local theaters where a movie is playing. In a similar vein, a number of politicians have discovered the Web as a new channel for distributing their ideological products and public service information. More elaborate Electronic Data Interchange (EDI) for business to business use has been suggested, but if much has happened in this arena, it has not been widely shared.
But the external Web is only one potential use of the Web technology. Its use for improving communication and coordination inside large organizations has the potential to generate even more profound effects on the way we do things than the external commerce receiving most of the publicity.
Earlier technologies applied to these corporate problems were centralized in nature and difficult to keep current. More recent groupware technologies have attempted to provide a distributed approach, but users have complained that they are not intuitive to use, and if they want to use an application not already integrated with their proprietary groupware package, the custom integration takes a significant effort by an expert.
The Web technology is different. It is easy to create, publish, and access information; so easy that anyone with access to the TCP/IP backbone can publish information inside the enterprise that is accessible on a wide variety of client platforms. Additionally, many off-the-shelf applications can be quickly and easily added as helper applications. Even interfaces to custom legacy applications are much easier to create than with traditional programming approaches. An analogy that comes to mind is the explosion of desktop systems that occurred after the microprocessor revolution. In this case we are facing distributed information management rather than distributed computing.
The Web technology makes the creation and publishing of information easy. It also makes the retrieval and viewing of the information easy. What is not easy is finding the relevant information that is created in this independent environment. What Web technology is missing is a marketing and distribution channel for the information created. Without a marketing and distribution channel there is no efficient way for information to be found. Our paper systems have information marketing and distribution channels that have been refined over time. In most organizations it is so integrated with the way we operate that we don't think of it as an information marketing and distribution channel, it is viewed as the processes and procedures we use to do business. It is the management framework.
Often the Web usage in organizations began as a skunk works project started by technical experts. While today many of these projects enjoy higher visibility than during their inception, most still are managed as a skunk works, separate from the mainstream business. Much of the material on corporate home pages consists of traditional collateral translated to Web standards. The technology has not been incorporated into the business infrastructure, where it has the most potential to transform the way the enterprise manages itself.
An enterprise can treat Web technology as a separate or "extra-step" technology, or it can integrate the technology into its day-to-day operation. Once Web technology is effectively integrated into the internal operation, its effective use for external interactions becomes a natural and easy extension. Without the internal infrastructure, external interactions will always be strained and limited.
The advent of xerography machines made just-in-case information much worse. Instead of being limited to a small number of carbon copies, the copy list could be expanded. Electronic mail has carried this to greater heights. Today a "publisher" of information can store personal mailing lists and, with a single command, send a nearly unlimited number of copies "just in case" they might be needed. Some of these information loaders recognize that their lists are inappropriate. But instead of tailoring the list, they put an identifier at the front of the message that goes something like this: "If you are not interested in ... delete this message now." The message still clutters the mailbox, and the knowledge worker still has to spend time evaluating and disposing of it.
The alternative to "just-in-case" information is "just-in-time" information, or information on demand. This has been a promise of computers and networks that up to now has failed to meet its potential. Historically there have been two major approaches to just-in-time information delivery. The first left information distributed across applications and systems. To access the information the user had to learn and navigate multiple, complex, access procedures. Once accessed each application required a different interface. Faced with this level of complexity, users generally ignored just-in-time information. They may have learned how to access one or two applications, but the rest were left to languish.
To solve this problem, some enterprises attempted to collect all the distributed information into one master system. This gave the user a single access and single interface. However, because they tried to manage all requirements in the enterprise centrally, these systems tended to become large and complex. After more than a decade many still are not fully populated with information because the cost of inputting and maintaining it was too great. There were other problems as well. The complexity of these unified systems made them difficult to modify and to use. The tools to manage them were designed to support the discrete data of transactional processes. Over the past decade we have moved much more toward complex data to support informational processes. The shift in information needs combined with the difficulty of change has caused these large centrally managed systems to lag the enterprise requirements.
The Web technology offers a new approach for delivering information on demand. Because it supports distributed information authoring, publishing and management, it does not require the complexity of the old centralized systems. Information is authored and managed by those who create it, without having to rely on programmers to create data entry and reporting programs. With the new browsers, a user can retrieve and view information from distributed sources and systems using a simple, uniform, interface without having to know anything about the servers they are accessing. These simple changes in the technology will revolutionize our information infrastructures, and change our organizations.
Making this work requires not only a new information infrastructure, as discussed above, but a shift in attitude and culture. As creators of information we must retrain ourselves to publish without distributing. As users we must retrain ourselves to take more responsibility for determining and tracking our changing information needs, and actively and efficiently acquiring the information when we need it.
More detail on the roles required to support these changes will be discussed in the section on the IntraNet™ Information Framework
If information can be efficiently found and assimilated on demand, it changes the education and training model. Traditional training can focus on skills, while specific content can be provided on demand, as it is required.
The on-demand model has other advantages. Traditional information transfer technologies are batch oriented. A batch of information is created for delivery through documents or traditional education and training methodologies. Because the information is packaged as a batch, when one portion changes it often is not updated immediately. As time progresses the batch gets out of date. Some material becomes obsolete and new material is not yet included. To fix the problem, the entire batch would have to be re-released every time a change occurs, and the released batches all have to be updated. Several questions begin to emerge. How often can the enterprise afford to update the batches? How much time can the employees afford to spend in training? And, how much information can an employee absorb at one time?
As mentioned above, the Web technology supports distributed information creation and management, so each portion can be updated as it changes. If the proper infrastructure is in place, current information content can be found and accessed when and where it is needed. This reduces both the time needed for training and the amount of information an employee is required to absorb at one time.
Users access and view the information. There are many reasons why a user may be accessing the information, and the reasons will vary even from session to session with the same individual. The important point is that the user is where the value is created and the ultimate requirements are defined. If the difficulty of accessing the information exceeds the information's value to them, they will either not use it or will find another way to get it.
Authors create the information. In traditional media, authoring and structuring are intimately tied. The access paths are primarily linear, and all the related information is bound together and delivered together. If a different access path or a different combination of information is required, it frequently is more efficient to build the structure into a new document, replicating the common information, than to use external indexes to tie together pieces of different documents.
In the new world of hyperlinks, the structuring aspect of authoring will change. Several factors are driving this change. First, hyperlinks allow the user to pick and choose the order in which they want to access information. So while the author will necessarily continue to provide structure, the function of that structure will shift to helping users determine for their current need which information is valuable next rather than attempting to determine their needs apriori for them.
Second, when information is referenced or recombined, the original can be hyperlinked rather than copied. This concept will be covered in more detail in the Broker discussion below. However, it does have an impact on authoring. Much like the database world, the goal will become to reuse information rather than maintain separate copies that have to be continually synchronized. This means authors will become more concerned with structuring information into reusable modules than with creating specific linear structures.
Brokers are the key to finding information. This is true in the world of paper and will remain true as Web technology is adopted. Any technology that allows both prolific, independent creation of information and easy access to it will quickly generate inefficiencies for users trying to find information for specific needs. The information broker provides this service. Information brokers supporting our traditional information infrastructures are ubiquitous. We are using information brokers when we go to a phone book, to the TV Guide, to a librarian, or to an edited anthology. In our businesses we have people whose job it is to broker competitive information, to broker benefits and Human Resources information, and to broker product information. Many of the knowledge worker jobs are in fact information brokering jobs.
It is the brokers who will be the most affected by this technology. They can use it to search for and screen information for their constituencies. They can use it to deliver their results. But the most dramatic impact is the switch from delivering repackaged content to delivering information access pathways. This is an important distinction, because it changes both the deliverable and the focus of the information broker.
The deliverable in the past might have been a 300 page bound document. With the Web technology it might be a single page of hyperlinks. In creating the 300 page document, the broker spent a large portion of the time "cutting and pasting" and editing together information from other documents. With the Web technology this time will shift toward understanding the users' decision processes and structuring access paths through the content to better support those decisions. The Web technology replaces the time consuming effort of collecting and republishing text with single lines hyperlinked to the original text.
Publishers make information available. Organizations require information content to be managed, coordinated, and communicated in predictable and relatively efficient ways. Certain kinds of information are required for directed activity. Information becomes old and out of date. Some information requires approvals and validation. This is true of electronic files just as it is of their paper counterparts. A major difference is that we have 400 years of infrastructure built up to manage information conveyed on the paper medium. In fact, many of our bureaucracies and organizational processes are nothing more than infrastructures for moving information on paper, and many managers of white collar workers are in effect publishers of official information required by the enterprise.
Publishers have different information access requirements from users or brokers. They need a structure to help them efficiently manage the information content for which they are responsible. For this reason the IntraNet information framework supports two distinctly different kinds of pages and page structures. One supports users and brokers, the other publishers and managers. We call the second structure a Management Map, and the pages it contains map pages. Map pages are structured according to rules, and once created they do not require significant maintenance except when management changes occur. Even during these times the changes are straight forward, restricted to the areas changed, and not time consuming activities.
The diagram here provides an overview of the key components of
the IntraNet™ information framework. The roof depicts the business goals,
which provide the reason for creating the framework. The technical infrastructure
at the base provides the necessary foundation for the development, management,
and access of the information. It consists of the hardware, software, networks,
protocols and standards required to implement the remaining structure.
The middle three columns provide the support for the business goals. They
include both the publishing and the brokering functions mentioned above
and the additional function of the content approval processes that support
the publishing of "official" information. Content approval processes are
important for coordination and efficiency and to protect the enterprise
and its employees from liability and loss.
The implementation methodology used by the Amdahl team amplifies the
effects and quickly engages the client organization's personnel in the
management of the new framework. This process both shortens the implementation
cycle and encourages independent innovation so the benefits begin accruing
Comments about this material may be sent to the author at firstname.lastname@example.org