Imagine a file system that supports greater agility in capturing meta-data about files, folders, and collections of related documents. Further consider that the file system is no longer bound to the LAN, but truly embraces Internet protocols and open standards such as Web services and XML. It would be possible to create and manage observations and annotations [e.g., blog posts] about file system objects just as you might with a file-system-enabled blog platform.

I'm not sure which approach is better—blogs that are more file-system aware, or file systems that are more meta-data aware. Whatever the case (and both ideas are probably good) I believe this is precisely the direction that Longhorn is heading.

Longhorn will include a new file system known as Windows Future Storage (WinFS) that not only will revamp the way data is stored on physical media, but also will make the process of finding and visualizing data and applications faster and easier. It won't surprise me to find that the XML interface to WinFS will provide Web services that make it simple to wrap blogging tools and similar personal publishing ideas around file system objects.

"In many work environments, often a needed document sits on someone's computer."—J Baumgart

This is a problem we tackled with a MyST application prototype called Sphere. In that example application, we created a model for blogging about documents—an approach that embraced the idea that enterprise information about documents was a useful step in chipping away at the goal of unlocking hidden knowledge resources that every enterprise possesses, but cannot see.

Because MyST supports private and public channels, Sphere can leverage this security and permissions model to provide the ability for some documents to be visible to enterprise employees whereas other documents are private. Another aspect of Sphere is the ability to simply make the existence of document (e.g., it's meta-data) visible to the company. Direct arbitration with the author then ensues before a document is shared. The collaborative nature of knowledge-sharing is as important as the knowledge itself; this idea embraces the common belief that knowledge workers are protective of what they know.

Another approach we've taken is the use of Microsoft Word as an authoring environment for MyST Weblogs called Office2Channel. This tool makes it possible to create documents in Word and simply publish them as blog posts. This opens the possibility for off-line composition and editing of blog posts while mirroring your local documents as Weblog items. In future updates of Office2Channel we will make it possible for selected parts of a document to post to the Weblog while the meat of the document remains localized.

Andre is right. To create a truly effective social network, you must consider what people do, not just who they know. You must extrapolate the social map based on social transactions—the activity surrounding personal interactions. This data must also be captured in candid ways as part of the natural process of working.

"To be a truly powerful technology, as well as both more relevant to the larger population as well as inherently more accurate, social networking systems should focus on an automated means of data-collection (the data itself has no ulterior motive or ego). One way to do this might be to insert (with the users permission), agents or listeners 'in-stream' to our existing communications systems (such as Spoke) and then focus the attention on the algorithms which accurately reconstruct a social map of our connections as well as the relative strength & context of the connection." -- Andre Durand

Two people that know each other, but rarely interact in meaningful ways, does not represent useful information. You must consider the social transactions as the most prominent aspect of a social network. And you must do it in the usual and customary course of the work process. When I invite you to look at my whitepaper, that's a social transaction. When you comment about my whitepaper, that act is additional social data. When Bob declines to look at something, that's also valuable social network data. Tracking events (or meta-data) relevant to social transitions is where social networking must go, and that's what the MyST platform does.

I've been working with Bud Gibson, a University of Michigan Business School professor, who is putting together what he is calling the BIT320 Distributed Learning Blogosphere as part of a course (BIT320) that he is teaching this Fall. Its an interesting idea and I'm eager to watch the course blogosphere evolve as the term unfolds.  (Bud is considering making the space public; I'll post a link if that turns out to be the case.)

Along the way, Bud, I, and others have had various discussions in an attempt to develop a clear mission statement, not only for the course itself, but for how the course relates to the larger world of knowledge-related issues.  This really goes to the heart of what Bill French and I have been working on for the past four years—initially while running the Elmer project for Starbase Corporation (which as since been acquired by Borland), and more recently, as MyST Technology Partners.

Here are the basic ideas in a nutshell (okay, several nutshells):

• Knowledge is a key asset, if not the key asset, of most organizations.
• In the KM world, a Holy Grail is effective capture, enhancement, persistence, and transfer knowledge. (The goal of the BIT320 course is more modest, but it is a step in this direction—specifically, to develop an appreciation of why this grail is, in fact, holy.)
• Successful solutions will be those that emerge organically through small, incremental behavior changes of knowledge workers rather than asking them to use some new "KM System"—history has shown the latter approach fails far more often than not.
• Emergent systems must provide low-friction scenarios for knowledge to traverse "the last mile" in both directions, i.e., knowledge capture and knowledge dissemination
• Capture must occur anywhere (and eventually, everywhere) a knowledge worker is already working—in Word, in Excel, in e-mail, in web browsers, on PDAs, on cell phones, etc.
• Dissemination mechanisms must similarly deliver the right information into the right context at the right time. This is essential for extracting knowledge (i.e., capacity to act wisely) from the information tsunami (i.e., enormous and ever-growing volumes of data) that is already overwhelming knowledge workers. This implies there must be many dissemination mechanisms—RSS, smart tags, search, research services, e-mail notifications, pager messages, cell phone notifications, IM, and so on, as well as new, "smart" ways to help workers find stuff and—perhaps more importantly—stuff find workers. There is no one-size-fits-all solution.

What will it take to get there?  That is, more or less, the question Bill and I have been focusing on for years.  Here are a few essential ingredients—there are, no doubt, many more, but I'm convinced these are all critical.

• We need unifying knowledge meta spaces (it is impractical to imagine that existing information systems will someday all integrate with each other).  Further, it must be possible to automatically merge (or virtually merge) multiple meta spaces to create larger spaces.
• We need a formalized concept of information object identity.
• We need a pervasive security and permissions model.  This model must be enforced for all information access.  (Consider, what should someone find when looking for information about "layoff planning"?  Clearly, the answer depends on who is asking the question.)
• We need plug-in architectures to define new business logic for not-yet-understood (or even invented) requirements.  New business logic should automatically inherit core capabilities such as object identity, security, etc.
• We need plug-in architectures for new (not-yet-invented) capture and dissemination technologies.  Again, these should inherit all core capabilities.
• We need complete separation of (knowledge) content from (presentation) form.
• We need international standards (like XSL) to describe transformation of information from one form to another.
• We need international standards (like topic maps) for the representation and interchange of meaning.
• We need international standards (like XML, SOAP, HTTP, etc.) for platform-neutral machine-to-machine communication.

The exciting part of all of this is that we, as an industry, are getting tantalizingly close to having all the building blocks.  At MyST, we've already done a number of interesting "KM" applications based on the MyST Web Services Platform and the MySmartChannels Weblog Application Server, one of which is Bud's course blogosphere.  We're always on the lookout for new, interesting knowledge-related applications to attack.  Feel free to contact us with ideas.

I don't profess to be a search engine expert, however, I make these observations in an attempt to better understand how people actually find information. Some SEO "experts" seem to think that there's little value in getting high rankings on phrases that are not popular. I think that's flawed thinking because popular terms are likely to be ambiguous term that almost always lead to more specific searches to achieve a "discovery transaction"; i.e., the point at which a user actually finds something of value.

And it's this basic understanding of how people find stuff that has caused me to regard traditional SEO as a fundamentally disruptable model. In fact, SEO is being disrupted now. It doesn’t take a rocket scientist to see the problem that Weblogs have created for SEO techniques; good or bad, it's disruptive.

So why are SEO techniques flawed (or disruptable)? Simple

The argument that low popularity terms and phrases are not competitive, is also flawed. There are many instances where two searches per month can yield millions of dollars in revenue. Dominating a "competitive" term is relative only to the person who shares that domain of interest. If you happen to sell secure RSS tools, it’s extremely relevant to the very few people that need secure RSS tools.

There are certainly businesses that want to dominate highly competitive terms like "plays" -

ESPN and Broadway are two that come to mind. However, highly competitive terms are [generally] ambiguous and in most cases encourage, and eventually demand users to enter search terms that are less ambiguous. Measuring the number of searches in a month for "plays" is irrelevant if that’s not the phrase that was used to find what they truly wanted. More importantly – you must measure the final key-phrase used that led to the discovery of actionable information and a commercial transaction. Furthermore, a good SEO strategy accounts for this behavior by recognizing that [depending on whose data you use] the average key-phrase is 3.6 words, not one or two. There's also the issue of the keyword ontology in your content.

Our business model is not about key-phrase domination. Rather, it’s about providing a method (and technology) to dominate a market segment with discoverable information (for enterprises as well as small Web site owners). A good example is Steve Fendrich (Pioneer Drama).

"We're closing in on 200,000 hits on the website. Should happen in a day or two. We have been getting a large number of new customers who tell us they are happy they found us. Thanks for all you two have done." -- Steve Fendrich, CEO, Pioneer Drama

Pioneer Drama is a drama content publishing and information provider. The owner (and customers) really appreciate the value of their publisher's diary and Weblog (built in MyST) which currently delivers high ranking for about 500 key phrases relevant to that business doman. One of the phrases is "Children’s Christmas Plays". Very few people search for this phrase, but when they do, his company is visible. In fact, Steve dominates the top ten (#1, #2, #7, and #8) Google results for this phrase. More importantly, he just writes about his company and products, and this happens automatically.

IMHO, the SEO industry does a great job of force-feeding results. However, they typically optimize for terms that are most popular, not terms and key phrases that are most popular and that also lead to a transaction or actionable event. As people search for content they begin the hunt with very abstract one, or two-word queries.  They continue to play with the phrases until they reach a point where they stop searching. By design, the search process causes the user to depart from phrases that are further away from "popular" terms, and close to "productive" terms. This is a key element in understanding how people find information and how to develop your overall SEO strategy. Popularity is good, but it's not the most important measuring stick.

Businesses typically attempt to dominate a term or phrase not on merit, but through brute force. This is an arms race mentality; one that is difficult to sustain but may produce good results over a short period of time; necessitating the re-hashing of the pages from time-to-time by an SEO expert. The alternative approach is that you dominate a set of search results based on knowledge (e.g., a cloud of content that represents a superior collection of information objects that are assembled in a way that search engines [and customers] appreciate). This is what MyST has [unintentionally] achieved.

Bud Gibson has a vision that (if successful) will connect a collection of blogs in a tighter and more meaningful way that tiptoes into a community space. It's called the The BIT320 Distributed Learning Blogosphere, and MyST Technology Partners (Andy specifically) is providing technical guidance to help Bud create a production quality blog system designed to support classroom and small team learning.

As we're all aware, relationships between blog posts and blogs can be established though a number of mechanisms such as track-backs, but to create a tighter coupling; one that creates value by leveraging the sum-of-the-parts, requires an authoritative server to handle things like unified search of the blogs. There are many other benefits that accrue to the idea of a SmartSpace that aggregates (or places an umbrella over) blogs about a common theme. Search is a big one, but other areas to consider include:

• Imagine a Microsoft Office Research Service that allows Office users to perform a look-up into this tight group of blogs from any Office document?
• Imagine Office users subscribing to a smart tag recognizer that automatically links all the blog post titles from any Word document back to the blog posts.
• Imagine users creating their own [custom] RSS feeds that slice across the entire collective of blogs.
• Imagine a recommendation engine that based on search queries can also tell you who is likely to know the most about a particular subject. We call this three-dimensional social networking because it embodies people that create information for people that know people that cinsume information. 
• Imagine a social network map based not just on who knows who, but who knows who that reads blog posts about 'x'.

Although Bud is focusing on small team learning environments, this model applies to knowledge management and many other business-related information and collaboration objectives.