Cities of Text: Some Notes On Some Notes on Intranets, Knowledge Management And Urban Planning

Marc Demarest

June 1997



The intranet market is, for good or ill, a supply-driven market: IT vendors and not end-user organizations define the market, drive the market, make the market. What end-user organizations, in the name of intranets, make, most often, are data junkyards: unplanned, managed and unmanageable tangles of web servers loaded down with the same over-elaborated, inscrutable junk that littered the predecessor "sharing" technology-of-choice: file servers.

This document takes a page from the work of Howard Rheingold, suggesting that intranets Ė like the Internet itself Ė are not technology ensembles, but a new form of human settlement, a new kind of community. It explores the notion that, by drawing some parallels between the way we manage complex human settlements in the real world Ė through urban planning and management disciplines Ė we can begin to understand what it takes to design, build and run an intranet environment that underpins, rather than thwarts, a corporationís desire to (a) build virtual communities and (b) manage and develop their human and electronic knowledge assets.


Intranets: Neat Technology, Or New Kinds Of Human Settlement?

Since it has taken more than five thousand years to arrive at even a a partial understanding of the cityís nature and drama, it may require an even longer period to exhaust the cityís unrealized potential. [But] we need a new image of order, which shall include the organic and personal, and eventually embrace all the offices and functions of man.

Lewis Mumford, The City In History[1]

Intranets are cool. Intranets are in.

Intranets are what we do with Web technology now that the Internet software marketplace, loaded down by the performance expectations that come from outrageously inflated initial public offerings (IPOs), and thwarted at every turn by Microsoftís economic lock on the market, has decided it canít possibly push the era of electronic commerce into being with glossy ads, catchy phrases, and dire predictions of impending disintermediation, forward integration and the collapse of the logistics cost curve.

Been There, Done That

Intranets, in practice, turn out to be in most cases nothing more that the office automation revolution, redux. Instead of WYSIWYG word processors, with dozens of fonts and clipart libraries from which we can create gigantic, over-designed data-poor reports, we get HTML, multi-image GIFs, Active/X, JavaScript and Java.

Instead of file servers Ė byzantine hierarchical mazes into which we dump inscrutable containers of chartjunk called blarvitz.doc and blarvitz.ppt and blarvitz.wks Ė we have Web servers: byzantine, hierarchical mazes into which we dump now-scrutable containers of chartjunk called blarvitz.html.

In short, weíve traded one generation of junk-making tools for another, one generation of data junkyard technology for another.

If all intranetting is, is the second wave of the office automation revolution, why bother? If all Web servers do is give us the same old way to store the same old stuff, what possible returned business value can they deliver to the firm?

Letís not have a theoretical argument. Letís not talk about what, in theory, non-linear hypertextual environments could do for the organization of large bodies of heretofore unwieldy semi-structured information. Letís not talk about what, in theory, access to the assembled knowledge of their colleagues could do for the productivity of teams of knowledge workers. Letís not talk about what, in theory, collaborative conversation arenas, work spaces and corporate memory repositories could do for the management problems of the twenty-first century virtual company.

What The Firmís Intranet Really Looks Like

Instead, letís talk about the real, typical, average intranet implementation, as it is today in the typical commercial firm. Some salient characteristics of this installation are:

  • Fragmented, parochial content. A hooverville [2] made up of literally, dozens of web servers (provided free-of-charge by Microsoft and nearly free of charge by other firms), each with its own parochial content, designed according to the personal idiosyncratic design standards of its orginators, organized in heaven-knows-what fashion by marginally-competent unofficial webmasters who, until six months ago, were a marketing managers, or controllers, or customer support representatives.

  • Perpetual software purchasing. As scattered, disjunctively organized Web servers become unnavigable, search engine technology is purchased to pave over the islands of Webformation with search engine results. As search engine results themselves become unnavigable, meta-search engine technology, or "intelligent agent technology" is purchased, to make sense of the search engine output. And when these fail, other technologies appear, to remedy the shortcomings of the previous technology generation and introduce new dysfunctions into the environment. Layer upon layer of software: a stratigraphy of the failure of information technology.

  • Huge hidden operational costs. Hundreds or thousands of employees, formerly productive members of their groups, now spend significant portions of their work day manipulating HTML through increasingly elaborate software toolsets, honing their skills as Web publishers and neglecting their Ďday jobs,í because making multi-image GIFs is more fun than closing out the quarter or doing employee reviews.

  • Frantic IS organizations. Finally on the verge of re-corraling the rampant proliferation of PC and LAN infrastructure that made their lives a misery in the late 80s and early 90s, the IS organization is now floored to find their networks Ė long the stable heart of the corporate IT infrastructure Ė are slowing to a crawl as they move bootleg full-motion video of the Finance boating trip from Seattle to a hotel room in Myanmar, or the full sound recording of the last session of the marketing department social committee from Seattle to a dozen desktops in southern France. Not to mention the unmentionable material being dragged through the firewall in gigabytes from the capital-I Internet late at night by that odd fellow in strategic planning.

Sound familiar? Even if you havenít helped build one of these in your own firm, you surf through such a data junkyard regularly: we call it the Internet.

The IS organization hasnít complained yet; they know the CEO read about the Web in the latest airline magazine, and has been told by nearly every management consultant sheís talked to in the last year that the Web is "strategic technology." Complaining looks like a politically incorrect move.

The end-user organizations havenít complained yet, because (a) theyíre in control of the technology, for the most part, and (b) they are still in the early stages of thralldom: the technology has seduced them. "Where do you want to go today?" Somewhere where the GIFs are pretty and the stock prices scroll across the bottom of the screen just like they do in the movies.

The knowledge management groups havenít complained yet, because they desperately need the Web to succeed; they need a power base within the organization from which to work.

The CFO hasnít complained yet, because itís too early, in most firms, for the hundreds of hours poured into HTML coding and bootleg Web server administration to show up in the per capita revenue measurement.

Time To Get It Together

"Now, now," the virtual community theorists tell us, when we carp about how badly the Web-enabled Internet works relative to its character-oriented predecessor, "we have to give the emerging worldwide virtual community time to organize itself: find some notion of building codes, standards of community decency and practice, some unifying organizing principles that every member of the community abides by because the clarity these principles bring outweigh the loss of expressive freedom the members experience in submitting to them."

And theyíre right, in several ways:

  • What web technologies create is, really, a new kind of human habitation: a space people build in, and then inhabit, in many ways as rich a space as the physical spaces we inhabit every day.

  • There are no examples, in human history, of long-lived communities that had no social contract, no agreed-upon standards of behavior, organization and navigation. Sooner or later, all communities become unified through a common set of "codes," or disappear from the landscape.

  • The Internet has time to sort itself out. Years, decades maybe. And signs that the first, early flush of rampant I-got-mine expressiveness are giving way to a cooler, more community-oriented emphasis on ease-of-location, ease-of-navigation, ease-of-use are, finally, beginning to appear.

But the firm is not the Internet. The firm is a commercial entity that exists to get and keep customers, sell products and services and produce wealth for its owners and stakeholders. The average commercial firm doesnít have the time that the Internet does to sort its intranet out, nor does it have the money or the patience Ė or frankly, the talent in many cases Ė to allow its community to evolve, through a half-dozen technology generations, into a self-policing virtual habitation.

Cities Of Text

But thereís something undeniable in what the Internet teaches us: the notion that the Web is a new kind of human habitation. Those of us who operate out on the Web regularly feel, viscerally, the framework of the habitation, the digital spaces we inhabit and sometimes construct. We employ building materials Ė text for the most part, despite the ballyhoo about multimedia Ė and we use designs, of one sort or another[3], to build edifices: buildings of one sort or another, strung together according to some plan or other. In fact, we routinely borrow from the built environment to describe structures in cyberspace: home pages, billboards, stripmalls, and electronic storefronts[4] are part of our working Web vocabularies.

Question is, what kind of electronic habitation do we want to build inside the firm? Clearly, intranets are closest in intent to cities: complex configurations of buildings used by various merchants and citizens at various points in time to facilitate various kinds of transactions, from the quotidian to the spiritual. But what kind of city makes sense in an intranet environment? We have a variety of forms to choose from:

  • Cities like London and Paris charm us, but their variety comes at great cost: difficulty of ingress and egress, complex and counter-intuitive navigation, fragile infrastructure[5], great variance in custom and behavioral rules as one moves from district to district, and relatively huge maintenance costs.

  • Cities like Los Angeles and Mexico City are, despite their planning and design regimes, fundamentally unhealthy because one critical dimension of planning or another has gotten out of control.

  • The 19th century grid-plan cities of the Western United States and the UK offer superiod ingress and egress, ease of navigation and relatively stable behavioral models, at the cost of aesthetics: often the most easily-walkable cities are the least interesting to walk in[6].

Other forms present themselves as well: the polynucleic city model favored in Asia, with multiple concentrations of activity of particular defined types, connected to one another by high-speed transport, or the concentric circle city model favored in some parts of Europe for example[7]. Each city form Ė city form pattern, one should say Ė is a combination of two factors:

  • A model of physical distribution of infrastructure and superstructure materiel and resources, coupled with

  • A model of logical distribution of control over what sorts of activities and buildings may rightfully exist in the urban environment.

In other words, cities are designs constantly trading off physical distribution and logical control: precisely the core problem with intranets in the firm today. Some city patterns opt for more physical distribution, some for less, and all have associated models of design and implementation control. Some trade-offs have been decidely more successful than others. All attempt in some fashion to prevent the erection of substandard buildings, the selling of dangerous wares, the conduct of illicit activities, and the misuse of public property. All city forms, in short, have some idea of self-maintenance and self-regulation built into them: some sense of what they need to be now, and in future, and how to get there from here.

Unfortunately, the average firmís intranet is typically based on none of these real-world forms, with their histories of successes, failures, limitations and advantages. Instead, the typical intranet looks, if anything, like Tombstone, Arizona: a boom-and-bust ,clapped-up, clapped-out frontier town, a buildings-now-rules-later, caveat emptor shoot-em-up wild west town.

Around the town, a deadline[8], and inside utter chaos: sidewalks that donít meet, mud pits in the middle of the road, drunken cowboys and cow pies everywhere, and outhouses and brothels where you least expect them. The sheriff selectively enforces the law with a deadly hand, the mayor runs the casinos, and no one is safe on the streets on Saturday night.

Tombstone was, they say, a really exciting place to be. Until it became a ghost town.

Figure 1 Ė Were Warm Closes: A Childís Epitaph For Tombstone

What we Ė people who believe that Web technologies deployed inside the firmís boundaries can create new ways for people to work with one another, new systems for the construction, embodiment and dissemination of the firmís knowledge, new kinds of human habitations Ė ought to be asking ourselves is:

With four thousand years of urban planning experience and practice to draw from, why on earth would we willingly, knowingly, allow Tombstone to be built? Why would we do to our firm what the web technologies did to the Internet?

Ten Myths About Intranets

The world of information systems is an "immature" world. One has to be careful using that word in public because it normally has a negative connotation. But from a historical perspective, it is true....the information processing profession is historically immature because it has existed only since the early 1960s. One of the manifestations of the information processing professionís youth is the insistence on dwelling on detail. There is a notion that if we get the details right, the end result will somehow take care of itself and we will achieve success. Itís like saying that if we know how to lay concrete, how to drill, and how to install nuts and bolts, we donít have to worry about the shape or use of the bridge we are building. Such an attitude would drive a more mature civil engineer crazy.

W.H. Inmon, Building The Data Warehouse [9]

We have argued that the typical intranet is Tombstone-in-the-making. Why? Mostly because people donít think before they build, particularly when the building materials are exotic and intrinsically interesting. They just build, and try to keep competing interests from building at the same time.

In no particular order, here are some of the critical fallacies about intranetting that people trade in regularly: fallacies that build Tombstones.

  1. Intranets are just a matter of technology.

      By far the most common intranet myth, if industry publications are any reliable indicator of popular opinion, this myth suggests that intranets are just a technology architecture: TCP/IP, Internet messaging, and the WWW software and protocols. In other words, intranets are an IS problem, well within the technology domain: leave it to the professionals.

      Unfortunately, intranets are not, finally, about technology. They succeed, fail or flail based on how they are used by the people who are invited to use them.[10] Intranets populated with peopleís "personal pages," the cafeteria menus and the photographs from Joeís vacation in the Yucatan Peninsula do nothing for the firm, in terms of reducing operating expenses, increasing revenue or managing the firmís commercial risk, which are the only top-level criteria against which intranets should be judged. After all the technology infrastructure (see #10) is installed, the cost curve is just beginning to rise, and the returned business value proposition of the intranet (what offsets its rising costs) depends entirely on the nature of the human settlement that the users of that technology build together.

  2. Intranets have nothing really to do with technology.

      The oppositional position to #1, this myth is usually fronted by end-user organizations and knowledge management functions afraid that the IS organization is about to hijack their pet technologies, drag them into the data center, and then go dark. A sophisticated version of the "all power to the people" cry, this argument suggests that we should stop talking about the technology, which is really completely unimportant, and get on to the important issues, which of course vary considerably depending on which group is taking this position.

      This position is patently incorrect: 10 minutes with Netscape Navigator and Microsoft Internet Explorer is sufficient to make any thoughtful person understand that a supposedly irrelevant technology issue like "which browser should we use?" is really a tremendous strategic problem. Netscape is dead-set on using its current lead in the browser market to get out of the browser business: taking a page from Lotusí book, they are using their browser to drag into the firm their messaging products, web server software and collaborative computing infrastructure, seat by seat by seat. Microsoft, on the other hand, as a franchise owner in the messaging, web server and collaborative computing market segments, is using its browser technology to block Netscape, cement its dominance of the desktop office automation market, and take control of the next-generation application development paradigm away from Sun.

      Choosing one or the other has profound effects on how IS does its work in the firm for the next decade; choosing both is inviting chaos to play as it will in the firmís network. Such choices canít be made without advisory-and-consent from the IS organization, which has the responsibility for managing the firmís IT infrastructure Ė which intranets share but donít own Ė for the good of the firm as a whole.

  3. If we give people an intranet, they will spontanously share knowledge.

      This myth is a particular favorite of the soft-side knowledge management community, which believes (or perhaps needs to believe) that people are inherent sharers who have been waiting patiently (and one notes, quietly and alone) for the magic technology that will free their sharing instinct and give it full expression. The web is a sort of technological pixie dust: sprinkle over the firm, and stand back while the knowledge flows.

      Absurd. True, people will share stuff as soon as they get their hands on a browser: if they can write HTML or manipulate a WYSIWYG word processor, and if they can find a Web server, manipulate an FTP application to post their stuff to that web server. Whether the stuff shared is knowledge Ė that is to say, actionable, directed, contextualized information Ė is another matter entirely. Experience suggests it is far more likely to be just stuff: my tips on Airedale breeding, my views on the latest episode of the X-Files, my angry commentary on the decision of some other manager in some other part of the organization.

      And, true, people will consume that stuff. If they can find it on the dozens of web servers scattered across the corporate landscape in fifteen minutes or less: conventionally, the outer bound of the human "search space". If they can wait for the 256-color 28-frame animated GIF of the Airedale to load. If they are, at that moment, in need of advice on Airedale breeding.

      Sharing doesnít matter. Quality sharing among the right communities with the right commercial ends in view is the objective, and that, like most things, takes thought and explicit design.

  4. Intranets deliver huge ROI almost immediately.

      A necessary claim, if one is selling intranet software, and a variant of the sprinkle-the-pixie-dust argument, this myth suggests that in addition to the magic knowledge-sharing properties of the Web technologies, these technologies possess a magic wealth-creating capability: sprinkle and grow rich.

      When you scratch the surface of the glossy claims about instant ROI, what you discover beneath the technology is at least:

      • a strong, well-conceived and centrally-managed design for the technology infrastructure and the logical web space

      • a clear business objective behind the deployment of the technology and

      • a firm operational hand guiding who uses the intranet for what, and how.

      In other words, underneath the HTML, we find sound business and technology management practices executed in harmony according to a well-thought-out plan: a discipline that, unlike technological pixie dust, has been known to produce significant ROI rather more often than not. One suspects that these sound business practices would have produced the same results with paper cups and waxed string. Or paper. Or Lotus Notes.[11]

  5. Intranets are a new kind of IT system, and donít require the definition, design and implementation discipline we employed with OLTP and DSS systems.

      This is an odd myth, because it seems to come from IT organizations: the very groups one would expect to be arguing that Web technologies require the same kinds of definition, design and implementation disciplines weíve applied for thirty-odd years to OLTP and DSS systems. Gather requirements, define returned business value objectives, design the logical architecture, design the physical architecture, unit test, integration test, system test, deploy, measure, rework: thatís the pattern for conventional usage model design and implementation. Instead, IT organizations seem to be rushing to throw years of hard-won methods and practices experience out the window, and just do it: install the bits, get the browsers out, see what happens.

      We know what happens: the Internet happens. Data junkyards happen.

      Of course, the same arguments were made about data warehouses initially.[12] And about previous IT usage models. Junkyards were built, the innocent were caught and punished, and the guilty went looking for a design and implementation methodology to make sure they did it right the next time.

  6. The IS organization can run the firmís intranet.

      See #1. This is the operational version of that argument, which suggests that no business oversight of the firmís intranet, or its applications to the firmís knowledge dissemination and management problems, is required.

  7. The IS organization has nothing to do with the firmís intranet.

      See #2. This is the operational version of the argument, which suggests that the IS organization has no vested interest in how its carefully-designed and maintained network infrastructure is employed by the business.

  8. Intranets are self-organizing.

      This is absolutely true. There is never a "design-free" human construction: design is inherent in all human activities at one level or another. Intranets will self-organize: into parochial, competing, incompatible cliques. This is the default human design model for most large-scale design problems: segment the problem, optimize for local conditions, and the system be damned.

      This is

      • good news for software vendors who derive revenue from each clique

      • bad news for the IS organization and the knowledge managers who have corporate responsibilities for helping the firm to function as a single integrated healthy sociocultural entity, and

      • terrible news for the firm, which will find itself out-maneuvered by firms that donít make this design mistake, unable to connect its knowledge bases in any coherent fashion to those of its suppliers, trading partners, channels or customers, and spending significant portions of its IS budget to paper over the clique Intranets with complex, under-functional software in (largely vain) attempts to restore order to the network of ghost towns the firm has created.[13]

  9. Intranets are knowledge management.


      Knowledge management is a human discipline focusing on identifying, observing, instrumenting, managing and intervening in the knowledge economies at work in the firm. Technology can help knowledge managers with these activities, particularly by providing instrumented infrastructure that brings the currently-hidden knowledge economies out, into the light of day, within a "marketspace" infrastructure that is inherently instrumented. But that is all intranets can do: provide a necessary, but insufficient, infrastructure for knowledge managers to maintain and enhance the knowledge economies for which they are responsible.

  10. Intranets are cheap.

      This is a local variant of the "it doesnít cost anything so there is no downside if we fail" argument, common in organizations for years.

      The first order costs (hardware and software) of intranets are indeed cheap when the firmís lower-level infrastructure (particularly its network and transport infrastructure, its interpersonal messaging environment and its overall data management architecture) are already sorted out in a way that favors the deployment of large-scale Web-based infrastructure: TCP/IP across the firm, Internet-based interpersonal messaging, open relational database management systems.

      However, the first order costs of ripping up the 3270 highway inside the firm, moving from PROFS and All-in-One to SMTP and Exchange, and transitioning from IDMS to Oracle or Informix are extremely high: prohibitively high, in fact, when considered in any timeframe under five years in duration.

      And, when you consider the second and third costs of intranetting, including software costs, training costs, operations and management costs, increased network bandwidth purchases or installation, lost productivity due to obsessive use of Web content creation tools, lost productivity due to fruitless forays in the firmís web junkyards, and so forth, intranets become as expensive as any other enterprise-wide system.

      There are intranet designs that minimize these costs and drive down the total cost of enterprise-wide intranets, but these designs must be explicitly chosen: the default design model Ė hundreds of dissociated web cliques Ė is the most expensive design paradigm, and the one most likely to be chosen.

If you examine your firmís environment closely, youíll find some variant of most of these myths cropping up in nearly every adversarial intranet discussion, as well as myths we havenít listed here. Ultimately, the appropriate approach to enterprise-wide intranet design is fairly simple:

  • The design problem belongs to IT and the business together: both have a fundamental stake in seeing the technology infrastructure implemented properly, and a fundamental stake in seeing that healthy, productive habitations arise on top of that infrastructure.

  • The final decisions about how to employ particular technology components belong to IS, which has primary responsibility for firm IT. IS does not however have the charter, in theory, to specify what technologies are employed, at least not unilaterally in an organization with a formal knowledge management function (which should be staffed by people who are highly IT-literate).

  • The final decisions about habitation, community and functionality belong to the business operations, which inhabit the spaces the technology creates.

  • The two camps -- IT and the business -- can work effectively together as soon as they agree on an overarching metaphor that:

    • explains what their common mission is

    • provides a history of design and technology trade-offs to draw from and

    • provides a useful common language for expressing seemingly "new" problems brought on by adoption of intranet technologies.

This common metaphor, we would argue, is the habitation metaphor: intranets are cities of words, and the IS organization and the business are designing and managing the evolution of a city.

Intranets As An Urban Planning Problem

Impersonal forces do not transform human settlements. Or at least they do so only on rare occasions and these are natural disasters: fire, flood, earthquake and pestilence. Otherwise, the modification of settlement is a human act, however, complex, accomplished for human motives, however obscure or ineffective.

Kevin Lynch, Good City Form[14]

In this section, I offer some parallels between actual urban planning concepts and behavior in real-world built environments, and desirable (normative) behavior of intranet designers and implementers inside the firm, in order to make a single fundamental point:

    Urban planning has a 4,000-year history of design and implementation, has tried most of the design models and implementation techniques IT professionals consider for technology, and has arrived at some stable, nearly-universal methods and practices that seem to work when applied. Why reinvent the wheel?

The parallels between urban planning and intranet planning are rich and instructive, and we canít possibly cover them all here. This section serves merely as a set of pointers to areas of fruitful investigation and elaboration.

Note also that, underneath this metaphor[15] -- the intranet is structured like a city -- is yet another metaphor Ė the firm is a set of knowledge economies Ė that helps make the choice of the city model of planning and implementation even more appropriate for our problems, since cities are the most complex system ever designed by human beings, and serve both as places of habitation and market spaces for economic exchanges.

City Form

Urban planning concept

Use in the built environment

Analog in intranet environments


What is under the road, under the building and underground. What makes cities work. Infrastructure returns no business value in its own right. It has to be used by superstructural elements to return value.


In and of themselves, intranets are plumbing, wiring, road-building. What matters is the superstructure Ė the city Ė that is erected on top of this infrastructure.


What we see when we look at the city: buildings, public edifices, monuments, parks, office complexes, suburbs. The above-ground spaces people inhabit, collect in, and behave within

The knowledge management environment build on top of the intranet. It should have analog for all the private, semi-private/commercial and public spaces of the cities. There should be places on top of the intranet to change money, buy goods and services, register with authorities, meet friends, and feed the pigeons.


The notion of a defined space with a defined purpose within the city limits, to which all or most behavior of a particular sort is confined. Districts serve to contain kinds of behavior and to concentrate them so that they are easier to locate and use, and more likely to be successful.

In traditional intranet environments, these districts are the components of functional organizations. This is a design model equivalent to grouping all the grocery stores in the grocery store district, all the movie theatres in the movie theatre district, and other nonsense of this sort. A far more effective "districting" structure for intranets is one based on the firmís internal value chain, where districts are formed for core business objects like "suppler," "customer," "market,." "competitor" and so forth.

Neighborhoods and subdivisions

The notion of a defined space with a particular look and feel or character that is worth maintaining. Also, the notion that services of particular sorts ought to be close to where people live.

This is a notion that ought to be applied to the functional organization in intranet environments. Cities support both districts and neighborhoods. Functional organizational elements are neighborhoods, not districts. Districts have primacy.

Optimal distribution of resources

Distribute resources only insofar as that distribution creates optimal use of the city, not in such a way that it fragments the city into multiple competing villages.

Distribute the intranet physical infrastructure only insofar as it is necessary to promote full use of the intranet by all employees, and no farther. A web server on every desktop is over-distribution. A web server in every department is over-distribution. A single web server for the corporation may be under-distribution if the QN between North America and Asia is underfunctional.

Optimal centralization of control

Centralize control only insofar as that centralization promotes effective, healthy growth of the city. Distribute all non-essential design and implementation decisions to the lowest possible level.

Keep only the control you need to keep.

Allow service providers and consumers in the intranet environment every degree of freedom consonant with the business objectives of the intranet environment.

Incorporation and annexation

Incorporate or annex any largish settlement concentrations on your borders: make them part of the city, and subject them to the same codes and structures as the rest of the urban environment to keep the city growing well.

Incorporate or annex any largish rogue Web servers or intranets officially, and have the owners of those sites comply with the firmís guidelines concerning intranets and their use to avoid Web sprawl.

Growth boundaries

Place reasonable boundaries on how far out the city will grow to prevent sprawl and encourage urban renewal, and to keep the city manageable.

Place limits on how and where intranet infrastructure may be deployed, and force organizations that set up web infrastructure to maintain it effectively.

Urban planning

Plan 5-10 years in advance of the cityís growth to make sure that infrastructure capacity is not exceeded and that superstructure continues to serve constituencies as the cityís purpose and constitution changes.

Plan intranet development formally 1-3 years in advance to make sure user communities are led to new functionality, services and capabilities at the right time, and that firm IT infrastructure is not compromised by intranet usage.

Community policing

Make the community police itself and provide resources to help it do so, instead of hiring more and more violent law enforcement officers.

Make the community police itself, and provide resources for it to do so, rather than using knowledge managers or IT personnel as police and National Guardsmen.

Prevailing standards of decency

Be explicit about the prevailing standards of decent behavior and root out behavior that runs counter to those standards.

Have a well-thought-out policy regime for the intranet covering content, acceptable use, forbidden materials and activities and penalties for non-compliance before you deploy the intranet.

Core Constituencies

Urban planning concept

Use in the built environment

Analog in intranet environments

City planners

Responsible for planning and designing the city within a bounded future and setting policy that limits and directs growth

The knowledge managers in the firm

City managers

Responsible for keeping the city in its current form optimally well-maintained, hygenic and useful

The IS organization


Responsible for providing the product and service superstructure on top of the cityís infrastructure

The functional organizations and business teams within the firm


The people cities are designed for: the people who buy, sell, talk, and act within the urban environment.

All firm employees


Urban planning concept

Use in the built environment

Analog in intranet environments

Zoning codes

Determines where buildings and businesses of particular types may be located; cordons off certain kinds of economic activities (e.g., "red light" businesses) from the rest of the city

Helps prevent people from publishing, for example, personal content in corporate webspace if they have a zoned area for personal expression.

Building codes

Determines minimum allowable standards for methods, practices and materials, and in some cases requires specific kinds of functionality in buildings built by others (e.g., handicapped access)

Determine minimum allowable standards for webspace construction and population, standard "blueprints" for various kinds of intranet buildings, and allowable tools, languages and applications.

Building permits

Canít build a building or add on to a building without one. Canít get one if you havenít demonstrated minimum competence, or if you have a history of violations of the building codes.

Canít build a webspace inside the intranet without one. Canít get one if you havenít demonstrated minimum competence, or if you have a history of misusing the IT infrastructure.


What city managers do to make sure builders and merchants have complied with code.

What knowledge managers and IS professionals do to make sure Web space in the intranet complies with code.


What happens when you donít comply with code and are detected. Typically, the violator has some time to rectify the situation before his operation is shut down, but sometimes the violation is so egregious that the operation is shut down until it regains compliance.

What happens when publishers in the intranet web space donít comply with code. Typically, the violator has some time to rectify the situation before his operation is shut down, but sometimes the violation is so egregious that the operation is shut down until it regains compliance.


What happens to abandoned buildings when they become unsafe.

What happens to poorly-maintained web spaces when they become unsafe.


What happens when abandoned buildings are sold or given to people who commit to maintaining them.

What happens when abandoned web spaces are given to people who commit to maintaining them.

Building materials

Governed by code: what can and canít be used for what kinds of building activities. Technology. Shoddy materials are frequently the source of code violations. Wood, steel, stone, glass, plastic.

Governed by code: what can and canít be used for what kinds of building activities. Technology. Shoddy materials are frequently the source of code violations. HTML, Javascript, HTML editors, server-side includes, and the like.

Urban Sociology

The city form produces a kind of concentration of human activity Ė not unlike that produced by the firm as an economic organization Ė that has both desirable effects like economies of scale, cheap access to wide varieties of resource, efficient use of infrastructure, and undesirable effects like crime, poverty and aesthetic and infrastructural decay.

Intranets do the same thing.

As in the built environment, the desirable and undesirable behaviors and effects are as much a product of the environment as they are a produce of the individual and collective actors. One of the primary functions of the knowledge managers in a firm with a large-scale intranet environment is that of intranet sociologist: observing the development of communities and habitations on top of the web infrastructure, making informed judgements about the desirable and undesirable behaviors inside those communities, and making policy, implementation or enforcement recommendations that optimize the desirable behavior and ameliorate or eliminate the undesirable behavior. Some times this involves use of force: in the real world, the exercise of eminent domain or the calling out of the National Guard, and in the intranet world the revocation of a license to operate a web space or the termination of access privileges.

This implies that knowledge managers and IS professionals have

  • sufficient organizational power to use force when necessary to maintain the integrity of the intranet environment

  • sufficient restraint, given such broad powers, to use them as infrequently as possible, and in proportion to the behavioral problem in question.

Our experience suggests neither is the case in the main, making this area of the metaphor particularly problematic. Knowledge managers and IS organizations are either unable to enforce any building codes whatsoever, or are engaged in heavy-handed tops-down dictation of nearly everything.

There are no real-world "fully distributed cities" where city managers are just the streetsweepers and there are no real-world urban totalitarian city-states where city managers dictate where clocks are placed on peopleís mantelpieces.

The combination of the urban form and either

  • abrogation of centralized control over growth and development, or

  • autocratic exercise of total organizational power

on the part of city managers and planners is an unstable pattern; it doesnít last for long.

A Note On Organizational Form

The relationship between the four urban constituencies

  • urban planners and designers

  • city managers

  • merchants

  • citizens

parallels fairly exactly the desirable relationship between

  • knowledge managers and the CKO

  • IS professionals and the CIO

  • functional organizations and teams within the firm

  • users of the intranet environment.

Knowledge managers and the CKO are responsible for doing the work necessary to keep the city growing and vibrant and well-managed, but they do not do most of the actual management or the actual development of the infrastructure: that is done by the IS organization and the CIO.

Both groups Ė knowledge managers and IS Ė place strictures on merchants: the knowledge producers in the firmís knowledge economies. These strictures govern where and how producers may erect buildings (sites on the firmís intranet and spaces within the firmís webspace), what kinds of knowledge in what packaging they may sell in those buildings, how those buildings must be constructed and how they must be maintained. These strictures are not optional: to fail to abide by one of these strictures is to lose oneís license to do business in the city.

But these strictures are designed to facilitate commerce, not impede it, and the merchant still has wide latitude about how buildings look and feel, what products and services the merchant sells and what techniques. Merchants should have all freedom possible to advertise, entice and connect responsibly with their customers, but the city should not become a caveat emptor boomtown.

All three groups Ė planners, managers, merchants Ė are focused on making the city effective, livable and beautiful (in that order, in many cases) for the citizen, who is the center of the urban design problem as an individual and as a member of various civic groups. No urban planner would reward herself for producing a beautiful city no one inhabited; no city manager would place the technical sophistication of his urban infrastructure over the usability of that infrastructure by citizens; merchants do not reward themselves for producing intricate, complex stores no one enters.

The city declines; the population moves to the suburbs or another area of the country; the merchants go out of business: those are the consequences of planning, design, implementation and service practices not focused on the use of the environment by the citizenry.

How Weíll Know Weíre Doing Well

Kevin Lynchís urban performance criteria are a remarkably a propos instrument for judging our success at effective intranet design and implementation.

Kevin Lynchís Urban Performance Dimension

Lynchís Definition for Urban Performance

Intranet Architecture Performance Analog

1. Vitality

"The degree to which the form of the settlement supports the vital functions, the biological requirements and capabilities of human beings Ė above all how it protects the survival of the species."

The degree to which the intranet architecture explicitly supports the firmís stated strategies to deliver a defined level of returned value to shareholders or owners, employees and customers, as well as other stakeholders.

The degree to which the intranet architecture, as implemented, supports the day-to-day work activities of the employees required to make good on the firmís stated business strategy and objectives.

The degree to which the intranet architecture, as implemented, supports the product, service and process innovations required by the firm or its employees without substantial lags in time between (a) the recognition of the need for innovation and (b) the ability of the IT architecture as implemented to help produce such innovation.

2. Sense

"The degree to which the settlement can be clearly perceived and mentally differentiated and structured in time and space by its residents and the degree to which that mental structure connects with their values and conceptsÖ"

The degree to which the firmís intranet architecture, as implemented, reflects the firmís cultural values.

The degree to which the intranet architectureís dominant navigational metaphors provide simple access to required resources, arrange those services and resources in appropriate contexts, and reflect the firmís cultural values.

3. Fit

"The degree to which the form and capacity of spaces, channels and equipment in a settlement match the pattern and quality of actions that people customarily engage in, or want to engage inÖ"

The degree to which the firmís intranet architecture, in design and in implementation, explicitly support the firmís business processes and the collaborative work of process teams as opposed to the automation, routinization and scrutiny of the tasks of isolated individuals.

4. Access

"The ability to reach other people, activities, resources, services, information, or places including the quantity and diversity of the elements which can be reached."

The degree to which the firmís intranet architecture, in design and in implementation, guarantees all firm resources (subject to security considerations) are accessible to employees when needed as needed, in the context of a unified navigational model.

5. Control

"The degree to which the use and access to spaces and activities, and their creation, repair, modification and management are controlled by those who work, use or reside in them."

The degree to which the firm can explicitly control all resources made available by the intranet architecture as implemented, including the passive monitoring of those resources, active management of those resources, and explicit control over access to those resources.

The degree to which the firmís intranet architecture, in design and in implementation, assimilates new resources and services that are introduced without substantial rework or extension of the firmís transport and network infrastructure, systems architecture or data architecture.

The degree to which the firmís intranet architecture, in design and in implementation, guarantees that new resources and services, once introduced, collaborate with existing resources and services without substantial rework or extension to either new or existing resources or services themselves.

6. Efficiency

"The cost, in terms of other valued things, of creating and maintaining the settlement, for any given level of attainment of the environmental dimensions (1-5) listed above."

The extent to which the intranet architecture predicts accurately the first, second and third order costs of its implementation; the first, second and third order benefits that will accrue as the result of its implementation; the timeframes required for benefits to accrue; and the risks associated with accrual of those costs, benefits and time scales.

The extent to which the intranet architecture contains within it a comprehensive model for evaluating the returned business value (RBV) of any new project considered as an extension to the architecture.

7. Justice

"The way in which the environmental benefits and costs are distributed among persons, according to some particular principle such as equity, need, intrinsic worth, ability to pay, effort expended, potential contribution, or power."

The extent to which the intranet architecture provides the mechanisms for the firm to allocate access to and control over resources based on its value system.

The extent to which the intranet architecture, in design and in implementation, strikes the appropriate balance between the empowerment of the work force and the fiduciary and legal needs of the firm.

The extent to which the intranet architecture, when it infringes upon an employeeís rights or privileges, does so explicitly and in the context of an explicit policy that existed before the act of infringement.

All of these criteria are of course, equally applicable to other kinds of IT projects: in fact, to IT enterprise architectural planning itself. But they sum up, quite effectively, what a dynamic, productive intranet environment focused on knowledge management would look like, and serve to set an agenda for intranet designers and knowledge managers alike.

Coda: Is The City Ever Built?

One last aspect of the "urban planning model of intranet design" requires some thought: the notion, fundamental among urban planners, that planning is a continuous process of staying ahead of an increasingly large and complex form of habitation that never stops changing.

Unlike other kinds of IT systems, which are redesigned at fairly long intervals and are stable between redesigns, an intranet environment is reshaped by every single habitation and activity it supports. Intranets are, in short, never built and always in-building. That implies that the IS and business resources devoted to the "urban planning function" of intranet environments have made a standing commitment to design and implementation/reimplementation, and further implies that a firm committed to intranets as a means to create knowledge-exchanging communities is also committed to allocating the human, technical and capital resources required to keep the intranet city vibrant on a perpetual basis.


  1. Lewis Mumford. The City in History. New York: Harcourt, 1961, 1989. ISBN: 0-15-618035-9. Pp. 3-4.

  2. A term used during the Depression to refer to the shanty towns of the migratory unemployed that sprung up outside of major metropolitan areas.

  3. There is a notion, quite prevalent on the Web, that (a) design isn't relevant or, if it is (b) design is a problem for graphic artists. This is tantamount to saying, in the built environment, that (a) design isn't relevant, or, if it is, (b) design is aproblem for interior decorators. Where are the architects, the urban planners, the urban sociologists of hyperspace? Nowhere to be found, for the most part, inside the average firm.

  4. It is tellingly naive that the only elements of the urban metaphor that have crept into our vocabulary are the elements of the private citizen/consumer (home) and the uglier merchant arrangements (stripmall, billboard, storefront). Where are the non-commercial communal spaces: the other meanings of agora?

  5. Anyone who has been in any major European city during a public transit strike of any kind is well aware of the incredible fragility of the infrastructure of these cities. When the trains don't run, London doesn't run - a fact that, in and of itself, suggests the "unplanned organic growth" city form is suboptimal as a metaphoric basis for building intranets.

  6. Portland, Oregon is generally held to be an exception to this rule, due largely to the consistent and thoughtful emphasis on the part of the city managers on the importance of urban planning for manageable growth.

  7. See Hatt and Reiss, eds. Cities and Society (Glencoe: Free Press, 1951) for useful background information on city patterns and forms.

  8. If we spent, collectively, 25% of the design time we spend on firewalls designing intranets, we could skip two generations of search-metasearch-agent software.

  9. W.H. Inmon. Building The Data Warehouse. New York: Wiley-QED, 1993. p.1

  10. We need look no further than the large numbers of unsuccessful Lotus Notes implementations to see a historical example of the results of the "collaborative computing is just a technology issue" mentality.

  11. The argument that the Web is technically superior to Notes is pointless. Notes is a dangerous knowledge management infrastructure not because of its technology, but because it is owned by a single vendor (IBM) that has shown itself (a) historically incapable of achieving franchise in software markets other than the mainframe (remember OS/2?), (b) more interest in cementing its control over its installed base of proprietary technology that attaining market franchise and (c) susceptible to megalomania. 100,000 Web companies; 1 IBM. Who wins? And then, of course, IBM says this argument is irrelevant as well, since Notes is or will become the Web. So the battle is already over.

  12. In fact, the parallels between the early years of the open data warehouse movement (1988-1993) and the early days of the intranet movement are striking, not only in that the same myths were common currency, but that the same results accrued: in the early 1990s, many firms built expensive and entirely unuseable data warehouses, until these firms got the "denormalization" and data mart religions, and, in the 1990s, many firms are building highly-distributed unuseable web junkyards.

  13. We once had a Webmaster tell us that centralized implementation was against "the spirit of the Web" in that Web technologies were developed for the Internet, a distributed community. When asked if he intended to reproduce the Internet inside his firm, the Webmaster replied, "Absolutely. Except it will be secure. "

  14. Kevin Lynch. Good City Form. Boston: MIT Press, 1981. (ISBN: 0-262-62046-4) p. 5.

  15. Properly speaking, this is a simile.


In addition to the texts noted in the footnotes, people interested in pursuing the parallels between urban planning and intranet planning (or between urban planning and enterprise IT planning and design in general) should consider the resources listed below.

Christopher Alexander. Notes on the Synthesis of Form. Boston, Harvard University Press, 1964.

Christoper Alexander, C., M. Silverstein, S. Angel, S. Ishikawa, & D. Abrams. The Oregon Experiment, London: Oxford University Press, 1975.

Christoper Alexander, C., S. Ishikawa, & M. Silverstein. A Pattern Language. London: Oxford University Press, 1977.

Christopher Alexander. The Timeless Way of Building. London: Oxford University Press, 1979.

Christopher Alexander. A New Theory of Urban Design. London: Oxford University Press, 1987.

Edmund Bacon. Design Of Cities. New York, Penguin, 1974. ISBN: 0-14-004236-9.

Francoise Choay. The Modern City: Planning In The 19th Century. New York,: George Braziller, 1969.

Mark Girouard. Cities and People. New Haven: Yale University Press, 1985. ISBN: 0-300-03968-9.

Jane Jacobs. The Death and Life of Great American Cities. New York: Random House, 1961, ISBN: 0-679-74195-X.

Le Courbusier. The City of To-Morrow And Its Planning. New York: Dover, [nd]. ISBN: 0-486-25332-5

Kevin Lynch. The Image of the City. Cambridge: MIT Press, 1960. ISBN: 0-262-62001-4.

John Ratcliffe. An Introduction To Town And Country Planning. London: Hutchinson, 1981. ISBN: 0-09-144021-1. This text is part of a Hutchinson series on the built environment, which includes some 20 texts all of which are extremely useful for understanding the methods and practices of urban planning and design.

Howard Rheingold. The Virtual Community: Homesteading on the Electronic Frontier. New York: Harper, 1993.

Frank Lloyd Wright. The Living City. New York: Meridian/New American Library, 1970. ISBN:0-452-01035-7.


Last updated on 06-26-97 by Marc Demarest (

The authoritative source of this document is