Semantic Web: The Missing Pieces

Kate Ray put together an excellent short film (15 min) on the semantic web. See “Web 3.0″ on her blog. The topics covered are spot on, and the people interviewed are thought leaders in field, including Clay Shirky, David Weinberger and Tim Berners-Lee.

The arguments closely mirror a recent presentation I gave at the IA Summit in Warsaw called “The Navigation Layer: Making Sense of it All.” It was comforting for me to hear others making these same arguments, for one. But there seemed to be a big piece of the puzzle missing in Kate’s Web 3.0 video, which I explain at the end of this post.

Below you’ll find a correlation of themes from the Web 3.0 video and my presentation at the Polish IA Summit:

WEB 3.0 VIDEO: “The core problem is that our ability to create information has far exceeded our ability to manage it.”

MY PRESENTATION: Yup – information overload is the primary challenge we face online. What’s more, to paraphase David Weinberger, the solution to too much information is to add more information; namely, metadata. I then identify three sources of explicit metadata: expert created (e.g., controlled vocabularly), algorithmically created (e.g., entity extraction), and user created (e.g, tagging)

But metadata alone isn’t helpful. It needs some kind of structure.

WEB 3.0 VIDEO: “What if I could start to put things together in all that flow of information? In order to do that you need some sort of structure.”

MY PRESENTATION: Just as there are three types of metadata, there are three types of structures: expert created (e.g., a taxonomy or thesaurus), algrithimically created (e.g., tag cloud), and user created (e.g., social groups on bookmarking sites).

Which types of metadata and structures are best? I strongly believe it’s not an either/or question.

WEB 3.0 VIDEO: Are ontologies needed for the semantic web? Many of the experts in the field believe so. Others dissent. David Weinberger: “We are always going to be filtering the filters that filter our filters that filter our filters.”

MY PRESENTATION: We will need ALL types of metadata and all types of structures to deal with information overload. Solutions may mix top-down ontologies with algorithmically created metadata that users can tag as they see fit. The filters needed can (and probably) will be a mix of types of metadata and types of structures. See my article in the ASIST Bulletin, “Navigating The Long Tail,” for more on this.

It’s all about relationships:

WEB 3.0 VIDEO: “With semantic web technologies, the links mean something.” “If I have enough of those relationships it builds context. And context is what it’s all about.”

MY PRESENTATION: Amen. We need to move beyond just findability. We need to help people make sense what they find. To do this, you need to expose relationships between metadata in the navigation layer. Conext is critical here.

WEB 3.0 VIDEO: Clay Shirky: “Content, as it turns out, is not king”…”How do you filter things to create more value than you currently get?”… “If I were going to start a news business tomorrow, I’d start a news business designed to produce not one new bit of news. But instead, to aggregate news for individuals in ways that matter to them.”
John Hebeler: Google doesn’t scale to meet the challenges of trillions and trillions of information objects on the web. Instead, “you have to do all of the integration in your head.”

MY PRESENTATION: I show a screen shot of the LexisNexis Analytics Dashboard, a producted I helped design. This solution provides a media analyst or a reputation manager, for instance, the big picture of what’s happening in the news without having to read one news article. It aggregates metadata and presents it in charts and graphs that show relationships and patterns. Of course, you can link through to the content if needed, but very often it’s the picture of aggregated metadata that’s most insightful. So the curse of information overload, or more specifically metadata overload, becomes a blessing if you can harness the patterns it shows in a way that’s ready for human consumption.

WEB 3.o VIDEO: Clay Shirky: “Does the world make sense, or do we make sense of the world?”

MY PRESENTATION: We make sense of the world. Hence, sense making as a field of inquiry is absolutely required.

The Missing Pieces

The video has an overall techno-philosophical tone to it. What’s missing from the core of the discussion are human factors, I believe. In a previous post, I point to a quote from Ora Lassila, semantic web guru. He says:

After 10+ years of work into various aspects of the Semantic Web and its constituent technologies, I am now fully convinced (read: no longer in denial) that most of the remaining challenges to realize the Semantic Web vision have nothing to do with the underlying technologies involving data, ontologies, reasoning, etc. Instead, it all comes down to user interfaces and usability. Somehow, I repeatedly run into a situation where some use of Semantic Web technologies that would make a nice end-user application is “blocked” by the fact that the user interface is the real challenge.

How true.

WEB 3.0 VIDEO: “There’s a massive amount of potential, but no real tools to harness it.”

MY PRESENTATION: Beyond the technical tools that will harvest data and metadata, we need interfaces into the data that a.) mortal human beings can easily use and b.) help people make sense of the data they are viewing.

In my presentation I then point to four key challenges the design of interfaces for semantic technologies face:

• Representation - How information is displayed affects how it’s found and understood.
• Interaction - The ability to interact and manipulate information is important in sense making, particularly in digital environments.
• Context - The topical and thematic structures applied to metadata define scope and the boundaries between domains and niche markets.
• Time - Showing trends over time can in many situations provide the most meaningful insight.

So, it seems to me we’ve bashed around the technical and philosophical aspects of the semantic web long enough. I say until we can figure out how the promise of the semantic web can help average people in the daily lives, we need to shift focus to the human factors involved.

This is where information architects can potentially help. Unfortunately, not only do I not see different groups reaching out to each other across the aisle, I don’t see IAs moving towards solutions for the semantic web. Instead, the field is drifting off in the direction of genric user experience design (which is also important, but perhaps too broad to directly address the challenges of the semantic web and sense making). Sense making is an area of interest that has been around for decades. It’s had recent growth in interessant with things like sense making workshops at the CHI conferences.

To reiterate a point previously made:

The success of next-generation information systems depends much more on human factors than on more sophisticated technologies.

But beyond the human factors, it’s also not clear what the business value of the promise of the semantic web is. So in addition I also see a need to shift focus to business aspects of semantic technologies on the wbe.

In my opinion, the missing pieces in the current discussions of the semantic web and semantic technologies are twofold: 1.) users and 2.) business value:

The general semantic web discussion needs a shift in focus to the design of user interfaces that help people make sense of massive amounts of aggregated metadata while at the same time bringing value to businesses.

This is why I find projects like Pivot so interesting, which I mentioned in a previous post. The creators talk about joy and information experience, rather than data and technology.

The Problem With The Semantic Web: Usability

Check out Duane Degler’s presentation User Interfaces for the Semantic Web. In skimming it, I came across this quote from semantic web guru Ora Lassila, which comes from his blog post Semantic Web Soul Searching:

After 10+ years of work into various aspects of the Semantic Web and its constituent technologies, I am now fully convinced (read: no longer in denial) that most of the remaining challenges to realize the Semantic Web vision have nothing to do with the underlying technologies involving data, ontologies, reasoning, etc. Instead, it all comes down to user interfaces and usability. Somehow, I repeatedly run into a situation where some use of Semantic Web technologies that would make a nice end-user application is “blocked” by the fact that the user interface is the real challenge.

…

For a long time (longer than I have worked on the Semantic Web) I have wanted to build systems that work on users’ behalf. Semantic Web is one of the enabling technologies, a means to an end, and not the end itself. Every time I look critically at the current use of (information) technology, I cannot help but wonder how it is possible to actually get away with the approach taken today (where substantial burden is placed on the users).

This is a really powerful quote for anyone involved in UI design with semantic technologies. But for those of us from the user-centered design camp, it’s not really surprising or new. Having someone like Ora Lassila say it is.

In two related presentations I gave in the past couple of years, I make similar claims. In Navigating The Long Tail, I point out that the cost of having more and more metadata is noise. And that’s fundamentally a UI problem. In The Navigation Layer: Making Sense Of It All, I suggest four main challenges of UI design in hyper-information-rich contexts to assist in sense making:

• Representation - How information is displayed affects how it’s found and understood.
• Interaction - The ability to interact and manipulate information is important in sense making, particularly in digital environments.
• Context - The topical and thematic structures applied to metadata define scope and the boundaries between domains and niche markets.
• Time - Showing trends over time can in many situations provide the most meaningful insight.

Yet, most of the attention in discussions around semantic capabilities goes to the technological aspects. Sure, there are many UI projects–many of them academic–but these are few in comparison.

What’s more, beyond UI and usability concerns, I believe the semantic web and semantic technologies in general face another non-technical challenge: business models. We’ve not really figured out what the value is to customers and how to make money–at least not on a broad scale.

My general recommendation for any business looking to leverage semantic technologies is to spend as much time (if not more) figuring out a.) UI issues and how you’re going to bring value to customer and b.) the business model and how you’re going to make money.

The success of next-generation information systems depends much more on human factors than on more sophisticated technologies.

Entity Cube

I just came across Entity Cube from Microsoft Asia. The researchers mention right up front that accuracy of the entity extraction and categorization still needs work, so I won’t even comment on that. (Listing “Al Roker” in the category “Academic” seemed odd, to say the least).

From the site:

EntityCube is a research prototype for exploring object-level search technologies, which automatically summarizes the Web for entities (such as people, locations and organizations) with a modest web presence…EntityCube generates summaries of Web entities from billions of public Web pages that contain information about people, locations, and organizations, and allows for exploration of their relationships. For example, users can use EntityCube to find an automatically generated biography page and social-network graph for a person, and use it to discover a relationship path between two people.

So it’s really metadata (the summaries) about metadata (the entities)–a coffee table but about coffee tables, so to speak. Looks like we have so much metadata from the web, that we need help understanding it before we can use it. And it’s all done automatically–by machine algorithms. No need for IAs or the like, right?

Well, not exactly.

Turns out there is a ton of good ol’ hand-created IA on Entity Cube, primarily in the form of faceted navigation, both on the home page as starting points into an entity as well as in results lists to group results. But the site doesn’t really do a job with the information interaction design. They could have a much better experience with the information, in general. After all, it’s not just about getting the metadata (through entity extraction) but also how it’s presented and used.

It’s interesting, though, that browsing seems to be in the foreground of the product. Right on the home page, below the search input field, there are lots of links to scan and browse. Just another trend I’ve noticed recently–browse it “in.” Is browse the new search?

There’s also a visualization tool that attempts to present a wealth of metadata in a visual display. It’s pretty good, but I suspect it would have limited use and appeal in a broader context. That is, I don’t see the average web use “getting” the visualization quickly.