Note: I'll be recording a version of this article for National Public Radio's "All Things Considered" in the next few weeks. I've already bought an atomizer and ascot in order to protect my precious instrument.

Of course it's not literally true that The Web Changes Everything (the mantra of Webheads everywhere), but it's the right mistake to make. It keeps us on our toes so we aren't blindsided by the correct version of the mantra: The Web Changes More than We Expect. (That's not nearly as good a bumpersticker. That's the problem with truth.)

But all of these changes can be traced back to a five-word description of the Web: Many Small Pieces Loosely Joined.

Yes, of course it's obvious once you think about it that way. But obviousness is usually a good sign since most true things aren't radically different from what we already know. (That's why my consulting company is called "Evident Marketing," by the way.)

Think of it as The Web Effect. The Web takes just about anything it touches and smashes it into small pieces, and then lets the pieces re-join themselves in loose, unpredictable ways.

That's exactly what it's done to documents, for example. Authors of documents off the Web work very hard at providing a strong hierarchical structure -- which readers recognize as the document's outline. This structure provides the context for any particular piece of the content, and it's a very big part of the value that authors provide.

The Web Effect on hierarchical document structure? Boom, gone! Web pages are rarely arranged hierarchically. Instead, the author puts in hyperlinks so the reader can jump all over the place. It's like ripping the cover off the book and throwing the pages up in the air. Whoopee! Documents unbound!

So, a book goes from being a carefully structured, long document with lots of assumed context, read in a sequence dictated by the author to being a loose collection of short pages read in any sequence that interests the user. Many small pieces loosely joined.

If you take a diagram of an outline and rotate it 90 degrees, you get an org chart. The Web is doing to business organizations precisely what it's done to documents.

Thanks to intranets, people discover that they can "hyperlink" themselves into virtual teams to address short term objectives without having to bother going through the traditional hierarchical business structure.

Here at the HWJ we call this "The Hyperlinked Organization."

A hyperlinked organization is much more responsive to customers and to shifts in the market because the people closest to the customer and to the market are able to aggressively target emerging pigeons and problems (whoops, I meant "opportunities and challenges" -- time to go back to Marketing Slime school).

On the other hand, with all these teams organizing themselves and scurrying around, the corporation can easily lose control of its overall direction. Managing hyperlinked teams is a real challenge.

You can derive all the characteristics of a hyperlinked organization straight from our five magic words:

• Decentralization. Because the pieces are loosely joined, a rigid, centralized authority has a lot of trouble exerting control.

   

• Self-organization. This is one of the key principles of the Web, and it's a consequence of the Web's decentralization -- the Web itself is the result of people doing things on their own initiative. People organize themselves into hyperlinked teams.

   

• Anti-authoritarianism. Like it or not, the Web encourages the view that external authorities are nothing but an impediment. What do the authorities add if I can hyperlink a team together of the best and brightest to address some opportunity? Zippo, and we don't mean the cigarette lighter.

   

• Democratization. Little d democratization occurs when the hierarchy is disregarded. Your position in the hyperlinked team depends upon your perceived merit, not your official status or business card title.

   

• Attitude. When humans can't position themselves in a group by means of formal roles, then attitude starts to play the role of social shoe horn. Unfortunately, so far it seems like male adolescent cock fighting is the "winning" attitude on the Web.

   

• Attention Deficit. Because the pieces are so small, the Web encourages a shortening of our already microscopic attention spans. In a hyperlinked organization, this works out to virtual teams taking on small, short term projects.

   

• Brokenness. When you have a "system" that consists of many small pieces loosely joined with no one in charge and dependent on the unscheduled support of many individuals, it's always going to be a little bit broken (as Tim Berners-Lee, the founder of the World Wide Web, says). But, you know what? The big hierarchical organizations of today and yesterday are also always a little bit broken. We just don't like to admit it.

And to think all of these changes are because the Web consists of (c'mon, sing along with me now!) many small pieces loosely joined.

Postscript

Because the Web is so influential -- the Web sets expectations even outside of the Web, which is really kind of weird -- we're seeing other aspects of our lives suffering from the Web Effect as well. We'll talk about this a lot in future issues, but let me give you just one small example. Last year when my son was in kindergarten in a moderately progressive public school, his class was given the assignment of coming up with a question about fish and then showing how you go about answering it. His question was "How do electric eels make electricity?" (and then a small colleague added: "And why don't they shock themselves," a pretty deep question).

So, this class of five year olds set out learning that they can get a question answered by going to a library, asking a librarian, faxing a question to the New England Aquarium, using the Web, or looking it up on a CD ROM encyclopedia. My point is not that they're learning to use the Web in kindergarten. It's that becoming educated is shifting from getting filled up with contents to learning how to find answers. And this is directly analogous to what's happened to documents -- the prime vehicles of education -- on the Web. Traditional documents are valuable because of what they contain. Web documents are valuable at least in part because of what they point to. Our children are learning to be good pointers, not just good containers. We're subtly redefining what it means to be an expert: not the person who contains the most knowledge but the person who knows how to find (point to) knowledge.

I don't want to make too much of this ... well, yes, actually I do. It's my role in life. (Pity me.)

("The Hyperlinked Organization" term is a trademark of Open Text Corp., whom the HWJ thanks for giving us permission to use it in perpetuity.)

Buzz Soup: Aurora and RDF

Netscape's announcement of Aurora (at the newly resuscitated Seybold conference), is truly interesting. Netscape is positioning Aurora as its answer to Microsoft's Active Desktop which lets your Windows desktop be your primary interface not only to your own hard drive but to the entire World Wide Web. With Aurora, your Netscape browser in effect becomes your new desktop. So, with Microsoft your desktop can be your browser and with Netscape your browser can be your desktop. If you find this distinction to be more confusing than worthwhile, you're right. Putting aside all the many important issues of comparative functionality and stability, your decision about which way to go (and the integration of desktops and the Web is just about inevitable) will have a lot to do with how widely a new standard -- RDF -- is accepted.

RDF (Resource Description Framework) is based on XML -- eXtended Markup Language -- a very important standard we'll discuss in a future issue. RDF lets an organization assign metadata to information. Metadata is information about information. For example, a card in a library card catalogue contain metadata about the book: who wrote it, when it was published, where it's filed, etc. Metadata is a crucial information enabler.

By providing a mainstream tool that can make sense of RDF, Aurora will encourage document authors to add metadata to their Web pages, which will provide the critical mass required for other applications besides Aurora to make use of that metadata. That's the hope, anyway. (Click here for a Netscape page listing RDF resources.)

An official W3C working paper lists some of this new standard's uses:

"RDF metadata can be used in a variety of application areas; for example: in resource discovery to provide better search engine capabilities; in cataloging for describing the content and content relationships available at a particular Web site, page, or digital library; by intelligent software agents to facilitate knowledge sharing and exchange; in content rating; in describing collections of pages that represent a single logical "document"; for describing intellectual property rights of Web pages, and in many others."

Overall, RDF will help businesses address one of the central problems of the hyperlinked organization: making the massive amount of information in a corporate intranet or on the Web actually useful. Users currently have to wade through acres and hectares of information to find the occasional grain of rice that's of real value to them. The info bogs are currently, at best, sorted and catalogued by a centralized authority that can't possibly understand the user's needs as well as the user does.

The key here is that information's value is always dependent on the user's interests, and those interests are very varied and awfully unpredictable.

Initially, electronic information has been organized using principles designed for the constraints of printed matter. A librarian takes the mass of books and articles and arranges them on bookshelves according to a set of categories deemed to be useful to the mass of users.

Traditional libraries, that is, freeze information in time and space: each piece is filed in one place, and that arrangement is permanent.

The Web, of course, busts this up. Electronic information can be filed in more than one place. And the cataloguing can be entirely dynamic, changing as the information environment and users' interests change. In short, the library can be organized differently for each and every user, even at each and every visit.

Perhaps that's soft of what you thought hyperlinks do. Not quite. There are problems with relying on hyperlinks as the sole way of browsing for information. Hyperlinks represent the set of relationships the authors see, but these may not reflect users' interests. And, browsing using hyperlinks remains a pretty random way of finding information. You may discover a lot of useful material, but you can have no confidence you're finding everything that's out there or even a meaningful sliver.

Of course you can, and should, use a search engine to see what's available. But search engines are very literal and actually quite stupid. They don't know what the documents they're searching are about and they don't know anything about the user's interests. The result is "hit lists" (lists of documents found) that can be 200,000 entries long -- the Web's most frequent type of white noise.

So, how does RDF help? The key to good searching is metadata. Metadata lets the author (or librarian) attach information that will make the document much easier to find. The metadata may also include important facts such as that a document has a security level of 3, has been rated TD (for "Too Dull For Words"), and will be out of date by last week.

RDF lets a user find all the information she cares about by searching through the metadata. Thus, librarians and authors can cross-file a single page under many categories, and can provide the additional information that can filter out the noise.

A search through metadata (possibly combined with a search through the contents of the documents as well) can be thought of as a dynamic bookshelf or a "channel." Users and organizations can create highly customized, dynamic "views" of data by searching through RDF metadata.

Now, metadata isn't fool proof. For example, the company that created the mail order rifle catalogue from which Lee Harvey Oswald ordered his rifle undoubtedly did not attach metadata to it saying that the catalogue was going to be of interest to people looking for information about JFK's assassination. That is, authors may know what a document is about but they can't know why every possible reader might be interested in it. So, metadata has to be supplemented with good content tools, as well as human intelligence. (Wait, how did humans sneak back into this picture? Sorry!)

And, of course, RDF by itself does nothing. Users need applications that know how to take advantage of RDF. That's what Aurora is about. And since Aurora is aimed straight at the the Microsoft Death Star, we can expect Netscape to invest heavily in RDF's success.

This is great news for the hyperlinked organization trying to keep it's head above the information waves by letting users get highly relevant information without having to wade through a centralized library centrally organized.

The key to success with RDF will be not mastering the standard -- that's easy -- but thinking through what types of metadata will be useful to your organization. Then you'll need to institute practices to get that metadata attached. From then on -- once we have applications that caan leverage RDF metadata -- it's pure benefit.

 
 

Middle World Resources A BiWeekly Compendium of Resources
Walking the Walk   From Eric Severson, of IBM Global Services:   Yes, this is the same write-up that was in the previous beta issue. I'm working on more. That's why this thing is called a beta, for jimminy's sake! You may be pleased to know that I have been personally inspired by your concept of a "hyperlinked organization" and have been touting this idea within IBM. I have in fact constructed a web of "technical competency leaders" within my immediate group who en masse cover the spectrum of converging industry areas (document management, workflow, imaging, SGML/XML, web, groupware, electronic publishing, knowledge management), technical disciplines (OO, VB, database, etc.), and specific products (Documentum, Open Text, ArborText, etc.). This sort of "thousand points of light" is on the face of it entirely unmanageable until you look at it not as a traditional hierarchy but instead as a set of hyperlinks. More concretely, it doesn't really matter who reports to whom, what matters is whether you know who to go see to chase a particular issue, and who else should be brought in since they deal with a closely related area.   Simply following these natural hyperlinks turns out to be an extremely powerful way of looking at the organization. It's also a wonderful metaphor for seeing how to leverage each individual's thin bandwidth into a much more powerful whole: like the web itself, each individual site is relatively "thin" but taken as whole with all the links, the net result is awesome.	Cool Tool For the Hyperlinked Organization  Email is a real enabler for hyperlinked organizations, but if you want to take it to the next step, then hie on over to www.mirabilis.com where you can download ICQ ("I Seek You"). This piece of shareware provides a wealth of functionality all focused on delivering messages faster, better and more interactively than email. With it you can send a message to any other registered user and will pop up immediately on her desktop (if she's turned ICQ on during that session). Coolest of all, you can easily initiated a real time chat, so the two of you can watch each other type (typos and all) in a split screen window. You can invite others on board, transfer files, and save the text that's been transmitted. This is a great tool for distributed groups trying to collaborate in real time, as well as for people on the front line of business (like customer support) who could benefit from being able to give the local genius a dingle when a question needs to be answered pronto. ICQ in no way replaces email. It adds a new way to collaborate and stay responsive -- and potentially to wear our nerve endings down to the very nubs.
Trivia InformationWeek (Oct. 6) reports that there are over 140 vendors providing tools for searching text. (This figure comes from Collaborative Strategies.) Meanwhile, Delphi Consulting Group says that revenue growth will slow to 5% this year, down from 21% last year. But Dataquest says that businesses will spend $4.5 billion on knowledge management products and services by 1999. Hmm, put them together and what do you have? A search market that's been commoditized and is ready for consolidation. But if search products are going to play such a small part in KM, where's all that KM money going? Maybe into studies expanding the definition of KM so that it gets gerrymandered into containing $4.5B worth of stuff...

 

New Proof: Web is 2,000 years old!

Well, let's get out the old calculator. One document every four seconds means 15 per minute, 900 per hour, 21,600 per day,or 7,884,000 per year. Not bloody likely. There are more Pam Anderson nudies than that added per day (and that's not counting the skimpy bathing suit shots).

In fact, we can use this figure to do a type of carbon dating of the Web. If the rate of document additions is one per four seconds, then if we say there are 100 million documents on the Web (a conservative estimate), then the Web is 12.68 years old. But, given that the growth of the Web has been at least geometric, then the Web must in fact be at least 33.27 years old. This places its birth date at November 22, 1963 ... precisely the time that JFK was shot.

Coincidence? I think not ...

Tell me your tales

Email, Comments, Suggested Lifestyles

In response to last issue's comparison of Knowledge Management (KM) and Business Process Reengineering (BPR) on the question of top-down vs. bottoms-up changes, Amie White of International Data Corporation writes:

it is nice to see someone else compare and contrast [KM] to BPR. However, another major difference between BPR and Knowledge management is that BPR tried to streamline the intellectual capital within the organization (I am not sure what happened to it after its streamlining) and KM tries to collect any and all information regardless of where it is coming from. And then KM infrastructures attempt to sift through this information and index it according to a pre-defined taxonomy. However, it seems to encourage the information glut rather than streamlining intellectual capital.

Oh yeah? Says you!

Sorry. I was just trying out a feisty Web attitude. What do you think? Did it work for you?

Unfortunately, I agree with you, Amie. Ultimately, the test of a KM system will be not how much money you have in your intellectual bank account but how much of it is put into circulation. The last thing a corporation needs is another high-maintenance repository of information gathering dust. Unlike money, information that sits untouched in a repository precisely doesn't gather interest.

In the previous issue of the HWJ, I was trying to confine the comparison of BPR and KM to just a single issue. And my only real reason for bringing it up was to get to use the phrase Business Process Deengineering.

Bogus Contest: Safe Weaknesses