Pruitt-Igoe, a housing project built in the 1950s in St. Louis, Missouri, was a spectacular architectural failure. Neither its thoroughly modern design, nor its latest technologies, nor even critical acclaim could save it. The architects doomed the project by failing to take into account the way people actually lived. The massive scale of the project turned simple tasks like visiting friends or running out for milk into burdensome chores.The isolating expanses combined with the towers' huge galleries— envisioned as sidewalks in the sky—brought street crime to inhabitants' front doors. People left in droves. In 1971 a special commission, tasked with finding a solution, convened a town meeting with the remaining tenants. Shortly into the meeting the crowd began to chant, "Blow it up! Blow it up! Blow it up!" And they did. Within five years the entire complex had been reduced to a pile of rubble.

Click Here to Download the Complete PDF

Architecture Matters

It sets the environment in which people try to accomplish their goals. Information architecture, like physical architecture, sets requirements and restrictions on how users can interact with an application. It's often said that the demolition of Pruitt-Igoe was the death of modern architecture. More people-centered architecture followed. The advent of Web 2.0, though lacking a good demolition event, will be remembered as the death of modern information architecture. In its place we'll see a new usercentered approach that provides information tailored to the user's perspective—even as that perspective changes. But for this shift to happen, designers and developers must realize how three core ideas from physical architecture— place, path, and plan—imposed limitations of the physical world onto the digital one.

Static Places

In any physical structure there are individual destinations and waypoints.Think of an airport. A gate is a destination. Check-in, security, and junctures with more than one possible direction are all waypoints. Each physical area is defined by a sense of place—the appearance that you are in a distinct location. The shape, size, orientation, and physical markings of each area help you to know if you're on the right path, whether you've arrived, and once you have arrived, exactly where that is. In web apps, destinations are often specific articles, documents, or detail pages for a product or part. The waypoints are category and subcategory landing pages, plus search-results lists.

Pre-determined Paths

A path in the physical world proceeds like a train on a track, able only to travel a line or choose a fork—and we are all riding the same cars. It's the inverse of "if you build it, they will come." If you don't build it, there's no way they can get there. Modern information architecture perpetuates this tyranny of the perfect path. Human resource portals are a good example. Their deep menu structures force employees to specify a sub-sub-category of corporate policy before they can get to pertinent documents. Similarly, many customer-facing financial services sites bury the glossary of terms in the "Customer Support" section under an area marked "Resources." Users have to jump through the links the designers set up as if information in the digital world had to be in a specific spot. And the universal failover, search, offers paths that are little better. It simply demands that the user know the magic words to conjure up the perfect results set—a potentially shorter, but more error-prone path.

Monolithic Plans

Architects lay out a building's destinations, waypoints, and paths to get the optimal overall utility and beauty. But ultimately each thing has to go in one spot. And the risk is that people who don't know the whole system won't be able to use the piece they need. Just try to find the radiology lab at your local hospital. It's on some midlevel floor, in the back. Why? Because reception, the emergency room, and non-ambulatory care have to be on the first floor. For the hospital architect, all of these components are directly related to radiology and require explicit tradeoffs in the overall layout. But for the mother to- be, it only makes sense that the ultrasound machine would be on the ground floor, just inside the entrance. She's six months pregnant, for crying out loud. Carrying this constraint of fixedness into the online world creates unnecessary confusion. At a 401K site, keeping all the account management tools under the "Account Management" tab makes eminent sense to the designer looking at the information architecture of the whole site. But to the user, there's no reason that a link to the "Modify monthly deposit" tool can't appear in the "Account Management" area, and next to the account balance, and next to the report on monthly deposits.  

The Birth Of Context Architecture

The digital world delivers what the physical world can't—the unpredictably perfect path for each user. Unlike in the physical hospital, online, the expectant mother could make a beeline to her ultrasound results, organized alongside complementary nutrition advice, while the insurance call center agent could find those same results filed neatly by claim status. But how to summon such precise organization from a chaotic world where IT can't possibly afford such hand-built customization?

The answer is "context architecture," a remaking of the three core ideas of physical architecture for the digital world based on the interplay of three kinds of context.

The first is the context of the user, which can be expressed through explicit actions and profile information. We might know some behaviors characteristic of expectant mothers in general, and even about the past actions of specific patients. That's all helpful information when calculating what content and functionality to serve up.

The second idea is the context of the data. With every step, our user must always know all the possible (but only the valid) next steps given the specific data set. Think of this as navigation menus that constantly update to tell the user all of the choices available to her. And as common sense would dictate, none of the choices lead to a dead end—obvious in the physical world, but all-too-common online.When our heroine has gone up to the second floor, the lobby shop should no longer be an option, and when the secretary opens up the appointment book, an entirely new set of choices becomes available that weren't shown in the parking lot.

And the third idea is the context of the business. These are restrictions and recommendations based on policies and business rules. Expert knowledge, access permissions, and rules of thumb can be generalized and invoked in context. The security guard, the information desk, and the counselor in the hospital can all have their say at the appropriate moment online.

Taking advantage of these three kinds of context yields user-centric apps made of:

DYNAMIC PLACES

We begin with waypoint and destination pages, which give us a sense for where we are, where we've been, and where we can go next. In the digital world, these choices can be different for each user, as informed by their profiles and histories. So think of place now as the sum of our breadcrumb trail, our menus, and their options. Moreover, place can be augmented with dynamic recommendations—either different destinations or recommended next steps. For example, the mother who found her ultrasound results next to a pamphlet on prenatal nutrition discovers valuable material she didn't know to ask for. An information architect didn't have to place these items together manually because we could use the context of the user (this is a pregnant woman), the data (ultrasounds and prenatal nutrition are both related to pregnant women), and the business (doctors are advised to push nutritional advice to pregnant women) to make this recommendation in context.

SPONTANEOUS PATHS

Paths in the digital world can provide many equivalent ways to reach the same end, each customized to the individual user. A mother might find her ultrasound online by entering her social security number and then browsing under a menu subcategory that offers recent test results. An insurance agent might reach the same file by looking at his newest cases, refined by orders marked as unpaid, and finally narrowed by ones who've recently contacted the call center. The path is now determined by the user, as she chains together individual steps dynamically assembled based on what she has done so far, the possibilities in the data, and any business rules. Instead of the designer saying, "This is the one path," the user says, "I'll do this next," and makes a path—an unpredictably perfect path—in the process.

ADAPTIVE PLANS

The overall plan continues to relate waypoints, destinations, and paths, but given the far greater number of possible relationships, it takes on new challenges and opportunities.The essential change here is a move from categories to characteristics that allow category assembly on-the-fly. So now instead of guessing that a patient might expect to find her ultrasound results filed under the category "active cases" rather than "patient histories," the ultrasound is instead reachable by any of its many characteristics. Think of the path our mother-to-be or the insurance agent—each traveled as two of many custom categories, each composed of the characteristics particular to that one file.That introduces the possibility of an endless number of on-the-fly categories; so how do we make that into an orderly experience and bring it into reality?

Building with Context Architecture

Fortunately, this mass customization of information can actually be easier to model than a static, one-size-fits-all approach. The key is to take a faceted approach to the data model. Over the past few years, faceted classification has emerged as a best practice, revived from the theories of Ranganathan, the father of modern library science. What is faceted classification? Every instance of an entity— a document, a book, a consultant, etc.—simply possesses the set of characteristics that describes it uniquely. The characteristics are grouped by facets, which are orthogonal lenses with which to view the world. For example, a directory of consultants could be classified by several different facets like job skills, preferred working locations, performance rankings, availability dates, and so on. In fact, a lot of enterprise content is already classified this way, particularly if it's a document with fields and metadata, or records from a relational database.

Next, a faceted classification system needs a way to query it, making faceted navigation and search possible. The facets of the entities play three roles in this exchange. First, they are objects of a search, providing factual context for query matches. For example, a search for "Java" to find J2EE-certified consultants will match "java developer" in the "Job Role" field and "five years' Java experience" in the résumé. But the former is considered more authoritative than the latter. Second, using the relationships among the facets, the engine must also find all the other characteristics of the matching consultants, such as location, availability, and hourly rate. Third, the entire list of matching consultants is summarized according to each one of the characteristics possessed by the consultants in the list. Not only does the full list of matches come back, but also dozens of refinement menus, driven off the locations, availability windows, and hourly rates of just those consultants. This dynamic summarization happens with each step the user takes, constantly showing all the possible (but only the valid) next steps at every step.

Finally, the architect must have control over the display. The display must enforce restrictions and make recommendations. On the restriction side, iron-clad security must handle the unpredictable intersections of user context and data context. For example, if consultants in the directory happen to include specialists with a government security clearance, the user context and data context will reconcile a set of permissions to determine which consultants can be shown.The results list includes only those consultants, and just as importantly, the menu options include only the characteristics of those consultants, lest confidential information about the specialists' skills leak out.

The display should also make recommendations within context. This is accomplished through adaptive rules that let managers and site owners specify the kind of thing to do in a kind of situation. For example, if a staffing manager is browsing a list of available consultants, it's probably good to display recommendations for transportation options to get them to the job. If I'm looking at a list of offshore consultants, it probably makes sense to let a staffing manager know the contact information for our regional travel contractors, but if they're domestic, we might promote the latest HR policy changes on air travel. We let the system identify when to fire the rule, and let the rule determine how to properly execute given the context of the user and the data.When the offshore consultant and the regional contractor share the same country in the location facet, our rule fires, and our data model gracefully supplies all the context we need.

The Pruitt-Igoe housing project looked grand on paper because it seemed to accommodate everyone, but in reality, it didn't serve any particular individual well. The online world shatters the constraints of the physical world, delivering an unprecedented ability to design for the individual. Information architecture that works by analogy to the physical world is no longer enough.

Context architecture is the path forward, dropping each user in the unique place at the intersection of the person, the data, and the business.

Click Here for a Copy of the Complete Article in PDF Format

ABOUT THE AUTHOR

PAUL SONDEREGGER (psonderegger@endeca.com) is Endeca's chief strategist.

SearchInform Technologies, Inc. has announced a new direction for the company's software into the field of information security. SearchInform MailSniffer provides information security and can prevent the leakage of confidential information.

SearchInform MailSniffer RC intercepts all email traffic on the network protocol level, indexes the intercepted messages, and enables the user to conduct search through them. Regardless of the email client, MailSniffer has access to all sent and/or received messages on a given computer. SearchInform MailSniffer RC is designed to enable the user to conduct quick quality full text search with due consideration to stemming, thesaurus, and word location in a phrase. The search is conducted through all incoming and outgoing correspondence not only in the body of the letter, but also in its attributes and even in the contents of attached files. All intercepted information gets indexed and stored into the database, so that even if a message is deleted from the mail client, its contents will still be available for search. When viewing the history of correspondence between two people, MailSniffer displays it in a chronological order.

User access rights differentiation system implemented in SearchInform MailSniffer RC with its flexible settings is designed to allow for various access patterns. The administrator can assign the access rights in such a way that head of a certain department will be able to view the mailboxes of his subordinates, while they will only be able to access their own. SearchInform MailSniffer RC main features, include: previewing correspondence history with one recipient; full text search in messages and attached files with due consideration to stemming; a similar search feature; control over employees‚ correspondence; and user access rights differentiation.

(www.searchinform.com)

Information architecture, content management and taxonomy consultancy Earley and Associates has announced its acquisition of Wordmap Ltd., a vendor of enterprise taxonomy management software, for an undisclosed sum.

Earley and Associates' Principal Seth Earley, reports: "For some time we have seen Wordmap as a point of technical excellence in this field. Earley and Associates can strengthen that capability with the knowledge and experience we have of designing and compiling taxonomies. Making taxonomies operational in content management and search is a focus for our firm, and we look forward to working with existing Wordmap customers as well as other large enterprises looking for a holistic approach to taxonomy--a solution rather than just a product."

Bill Hutchison, Wordmap CEO, comments: "This is a very exciting opportunity for Wordmap to make the most of its innovative technologies. Our colleagues at Earley and Associates will allow us to bring together an outstanding software product with some of the best taxonomy consulting skills in the market, and these will complement each other and help our customers achieve a much better result, more quickly. Taxonomy is about much more than technology, and we think that we now have the mix of software and skills to make taxonomy work for the enterprise."

Earley adds, "An interesting aspect of the Wordmap suite that differentiates it from many products on the market is the integration with content tagging and search. After all is said and done, taxonomies are only useful if they are presented to the user in a meaningful way. Wordmap modules have the ability to do that without a lot of API level coding. The tagging module overcomes many limitations of content management tools in presenting and applying taxonomies for tagging. The navigation module is an easy way to add faceted search (also called guided navigation) without having to acquire additional faceted search tools. In these ways, Wordmap adds value to existing search and content management environments."

Earley & Associates will still focus on solving customer problems and not just selling them a tool, according to Earley. "If the tool is appropriate, then there is a terrific synergy," he says, "But no tool is appropriate for every circumstance, so we intend to remain a consulting firm and bring in Wordmap as appropriate." Wordmap will continue to develop and support the Wordmap range of products, which now includes:

Wordmap Designer--a system dedicated to the management of enterprise taxonomies.

Wordmap Content Tagger--a bridge between the taxonomy and the enterprise's content management systems; supports detailed tagging of documents by reference to the taxonomy as they are uploaded into third-party repositories.

Wordmap Intelligent Text Classifier--an advanced statistical text classifier based on Support Vector Machine algorithms.

Wordmap Navigator--a tool that takes published taxonomies from Wordmap Designer and converts them into HTML browsing interfaces.

Technorati has announced the launch of a new service called WTF (Where's the Fire), which encourages users to write explanations about why a given search term is hot--right now. Since users can also vote for the WTFs they think are most helpful, Technorati can highlight the best WTFs for the community.

(www.technorati.com)

Siderean Software, a provider of relational navigation solutions, has announced that the company is a participant in the Oracle Secure Search Initiative. The Secure Search Initiative is designed to allow software vendors like Siderean to partner with the company to develop a set of secure connectors that enable Oracle Secure Enterprise Search 10g users to access information from disparate enterprise data sources and applications. Oracle Secure Enterprise Search 10g users can leverage Siderean's flagship Seamark Navigator relational navigation solution to illuminate previously unseen relationships across all formats of digital information to produce a 360 degree view of available information.

With Seamark Navigator, Oracle Secure Enterprise Search 10g users have a view of the full range of assets that are available for decisions, tasks, and projects. Seamark for Oracle is a solution specifically for Oracle Secure Enterprise Search 10g. The solution's contextual content view shows the entire scope of available information and incorporates precision search to help users retrieve what they need. Seamark for Oracle Secure Enterprise Search 10g operates as a web service, allowing existing or new Oracle search deployments to be refined to give users an entirely new, insight-driven experience. Featuring a point-and-click administration interface, Seamark for Oracle Secure Enterprise Search 10g was designed to allow administration of the navigation experience from the Oracle interface. Additionally, the solution uses the Resource Description Framework (RDF) and Web Ontology Language (OWL) standards from the World-Wide Web Consortium. The Seamark relational navigation engine uses metadata facets to guide users to relevant content, generating views into the available assets based on a user's navigation. Item counts for each metadata facet, as well as the number and type of metadata facets displayed, are updated with each user click. Additional features for Seamark for Oracle Secure Enterprise Search 10g include: Relational Navigation; Navigation Administration; Controlled Vocabulary Administration; Platform Support; and Secure Infrastructure.

Clarabridge, a text-mining software company, has announced its participation in the Oracle Secure Search Initiative. The company's involvement in the initiative is designed to enable it to provide integration capability between Oracle Secure Enterprise Search 10g and major business intelligence (BI) tools. The enhanced Oracle-Clarabridge offering will consolidate knowledge stored in various BI applications across the enterprise and make it securely available to business users from an internet-like search box. The Clarabridge BI connector retrieves all reports from enterprise servers and pushes them to an indexing engine, which automatically performs updates as necessary. Results from one BI tool can be presented against another to ensure that the user finds the precise report desired, no matter what the underlying technology.

Inxight Software, Inc., a provider of enterprise software solutions for information discovery, has announced the general availability of Inxight SmartDiscovery Awareness Server for Oracle Secure Enterprise Search 10g. Inxight has also announced its participation in the Oracle Secure Search Initiative. The new Inxight SmartDiscovery Awareness Server for Oracle Secure Enterprise Search 10g, powered by Inxight's ThingFinder entity extraction, clusters search results on the fly, enabling users to filter their Oracle Secure Enterprise Search 10g result sets by the people, companies, places, concepts, and other information contained within them. It augments keyword in context summaries with the "Top Mentions" contained within the full text of the document. In addition, personalized page and term alerts automatically inform users of new information of interest in their Oracle Secure Enterprise Search 10g indexes, open web or other federated sources. Inxight SmartDiscovery Awareness Server for Oracle Secure Enterprise Search 10g allows users to simultaneously search not only their Oracle Database and applications, but any public web or Deep Web site, or other data sources using a single query. Awareness Server then allows users to see at a glance and filter search results sets by the relevant people, companies, products, and other information contained within them.

(www.siderean.com; www.clarabridge.com; www.inxight.com; www.oracle.com)

JumpTap, Inc., a provider of mobile search and advertising, has announced a new investment from WPP, a communications services group. This investment will provide clients of WPP companies, collectively buyers of media, with access to mobile search and advertising solutions. Through this partnership, both parties will also partner to develop new techniques to target and profile customers via the mobile phone. The partnership will also focus on developing case studies with strong brands who want to market to the mobile user. WPP joins JumpTap's existing investors, including General Catalyst, Redpoint Ventures, BCE Capital, and Valhalla Partners.

(www.jumptap.com)

eMag Solutions, an international electronic discovery company, has announced the availability of its new eMag PreVu software application. eMag PreVu, a raw data preview software application, provides a structured view into otherwise unstructured data, designed to allow users to better understand the scope of case information prior and subsequent to Meet & Confer sessions. This data is sortable and searchable by a variety of fields, including subject matter, keywords, content, context, custodian, metadata, and others.

Updates to the Federal Rules can shorten the time frame from the time a suit is filed to the Meet & Confer session where opposing sides sit down to discuss materials, processes, and schedules for discovery. The application allows users to view data structures and files in their raw native format prior to discovery processing. The PreVu user is also able to flag (include or exclude) at the tape, hard drive, directory, folder, and individual file level to expedite culling. PreVu also can allow users to more fully understand the data they are facing. The earlier attorneys get an overview of the subject matter, the more readily they can assess the merits and weaknesses of the case and prepare the strategy accordingly. This stretches beyond civil litigation to also include compliance and regulatory requests.

In addition eMag PreVu provides the ability to search and sort by server, custodian, directory, file type or date created/modified/accessed as well as perform full searches, including: subject matter, keywords, content, context, custodian, metadata, as well as fuzzy and wildcard searches. This application also offers idiom recognition and analysis of language structure down to word recognition at the root level. Users can save searches and their results performed using eMag PreVu and even create reports based on the search and sorting criteria they determine. The eMag PreVu is designed to allow companies and law firms to sort, search, and cull raw data in almost any manner.

(www.emaglink.com)

Kazeon has announced an alliance with Catalyst, a provider of litigation document review and extranet software, and CAAS, a legal and information risk management consulting company, to deliver an ediscovery and records management solution. This partnership leverages Kazeon's Information Server IS1200-ECS to decrease the time and cost of the ediscovery process. Catalyst's litigation support platform is designed to enable legal teams to organize and collaborate on search results generated by Kazeon while CAAS' discovery process management plays a role in guiding those organizations through the legal aspects of the ediscovery process for litigation and records retention.

(www.kazeon.com; www.catalystsecure.com; www.caasny.com)

Google Inc. has announced that the Google Mini offers search features for finding and sharing information within small businesses and departmental groups, including document and user-level security, as well as access to any business content through Google Onebox for Enterprise.

Starting at just $1995, the new Google Mini features: Secure Search, specially enhanced to support the information-sharing needs inside of small businesses and departmental workgroups, the Google Mini offers document and user-level security across all business content. Google's access control capabilities integrate with existing security systems; Google OneBox for Enterprise, the Google OneBox for Enterprise feature is designed to allow businesses to provide secure access to any information--such as contact and calendar info, HR benefits, sales leads, or purchase order status--through the Google search box; and other Site Search Improvements, designed so that site administrators can link the Google Mini search results page with Google Analytics to provide more detailed information about how people use search on their site in order to improve the overall site experience. The new Google Mini also automatically generates sitemaps, designed to allow webmasters to expose more public content for crawling and indexing by Google.com. The Google Mini is offered in versions that search from 50,000 up to 300,000 documents, includes a year of support, and is available for purchase online.

(http://mini.google.com)

Exalead, a provider of search software for business and the web, has announced the infusion of $15.6 million by Qualis SCA. Qualis, a long-time investor in Exalead, becomes its majority shareholder. The capital will be used to support international expansion and continued development of its new web search engine. In addition, it is designed to allow Exalead to open new subsidiaries in Europe.

Exalead has also announced the general availability of the newest version of its enterprise search software, exalead one:enterprise, designed to provide users with access to content and data--both structured and unstructured--regardless of format or location. exalead one:enterprise 4.5 offers a new user interface with greater search refinement options, improved performance for both 64-bit and 32-bit system environments, expanded language and file format support, as well as new management tools for administrators.

Features of this version include: the ability for customers to select from three user interfaces (UI) to meet the needs of employees: the UI available in exalead one:enterprise 4.0; the new, streamlined UI found on Exalead's web search engine for business-related searches inside the firewall; and, a white label version for organizations hoping to customize the look and feel from top to bottom; a more streamlined UI design than prior versions, but continues to provide Exalead's Search by Serendipity navigation technology and advanced features like phonetic, fuzzy matching, approximate search, word-stemming, thumbnail image previews, search term highlighting and spelling correction; compatibility for most major operating systems, including Windows, Linux, Solaris and UNIX, exalead one:enterprise 4.5 adds support for IBM AIX versions 5.2 and higher; expanded language support for Dutch; support for more than 320 file formats, including native support for Microsoft Office 2007; an updated connector for Microsoft Exchange; the availability of new CSS (Cascading Style Sheet) templates; and new exalead one:search APIs (Application Programming Interfaces) available so that administrators can add custom capabilities using the XSL (eXtensible Style Language) open standard.

(www.exalead.com)

Information Builders has unveiled WebFOCUS Magnify, a search and navigation tool that categorizes search results and supplements them with analysis and reporting capabilities. Magnify uses the metadata from Google or other search engines to index structured data records and provide access to all WebFOCUS capabilities through the search interface.

The company explains that Magnify captures data on a message bus. Using standard integration technology from iWay Software (an Information Builders company), Magnify enriches the messages, adds metatags and submits it to the search engine indexing mechanism, avoiding the need for crawling data stores, particularly database records, combining structured data in databases with unstructured search. WebFOCUS Magnify leverages the metatags and provides results in a navigation tree to guide users to the information they need.

Features of WebFOCUS Magnify include:

• dynamic categorization of search results,
• search-driven parameterized reports,
• dynamic directories, and
• search engine agnosticism.

SearchInform Technologies Inc. has introduced a new feature--SearchInform SDK, called to solve the problems of search for external developers. The API is integrated into custom applications and performs search through any data sources.

The new feature is designed to enable implementation of the SearchInform search module into any information infrastructure with further ability to conduct search through it or build a new information management system on the SearchInform SDK platform. SearchInform SDK features: COM technology designed to ensure integration; an indexing mechanism with support of a large number of file formats and databases; support of any non-standard data sources, proprietary data, and custom document systems; indexing speed from 15 to 30 Gb/hour; phrase search with due consideration to stemming and thesaurus; a Similar Search feature performing full-text search in the document contents; and technical support in the on-line mode.

(www.searchinform.com)

Fios Inc., a provider of electronic discovery and litigation readiness services, has announced new consulting services. Fios' new Review Management Services are designed to help clients assess and identify gaps in existing or proposed review strategies, identify opportunities for process improvement, and implement recommendations to improve attorney efficiency. Fios' Review Management Services are an expansion of the litigation readiness services the company has been offering since 2003.

Fios Inc. has also announced the official launch of its third sponsored information portal, TechnologyCounsel.org. This new web portal is dedicated to supporting the role of Technology Counsel and the legal practitioners who are bridging the gap between law and information technology. TechnologyCounsel.org offers news, trends, research, legal blogs, job boards, and articles focused on legal and technology issues related to litigation, governmental investigations, and compliance. In addition to news, articles, and industry research, TechnologyCounsel.org features a new industry Blog, entitled "Enterprise Matters," which provides observations on issues that technology counsel professionals can apply to their practices. The blog is supported by Mary Mack, Esq., and Daniel Pelc, Esq., Fios' Office of Technology Counsel, and electronic discovery experts.

(www.fiosinc.com; www.technologycounsel.org)