What the evangelists of the Semantic Web say is true: the bigger the web gets, the harder it is to find what we want to know. To that end, the perennial concern for providing "answers" has again popped up-and this time indicators point toward a substantive change in the tone of the web, along the lines of the paradigm shift from directories to search engines. Two recent developments typify this change-Facebook Questions and Ask.com.
Facebook Questions is a new feature that allows anyone on Facebook to ask questions of the entire set of 500 million Facebook users-and to contribute answers to the questions of others. This is a major new feature, and a major way to leverage Facebook's loyal users. Incidentally, they're not loyal to Facebook, they're loyal to the friends they have connected to with Facebook and to the sense of connection itself. It's those friends who'll be asking and answering for each other-and that's why this scheme is likely to work.
This has all happened before, sort of.
In 2008, Search Wikia debuted as a social-network powered search engine. It was meant to be something of a hybrid between a traditional automated web crawler and an aggregator of live, real-time, human responses to search queries. The project, noble in aspiration, promptly failed. It was ambitious, but ultimately lacking in a solid, thickly-stitched social network. Jimmy Wales and friends seemed to think the Wiki design itself would be enough to foster the growth of dense community for the purposes of Search Wikia. It wasn't.
So 2010 may be a better year for Q&A. Facebook's 500 million worldwide users is a nation-sized base of contributors. And, in the case of Facebook, each user has a stake in the process. If you answer well, you're rewarded with improved social standing amongst the circle you care about most: again, it's your "friends" you're working for. This stands Facebook far better than Search Wikia, though to be fair Facebook has no apparent aspirations for becoming a search engine itself. It knows how to stay in the box.
When the service is rolled out to the general user (right now it's in extremely limited beta), every user will see a Questions option on their Wall controls. Blake Ross (director of product management at Facebook) blogged about the upcoming feature on July 28. From his post: "To help us show your question to the most relevant people and ensure the best answers, you can tag it with a specific topic. For instance, if you have a question about what type of camera you should buy, you could tag it with ‘Photography.' If you want to find the best bike routes in the area, you might tag it with ‘Cycling.' The questions you ask will be shown to people who have expressed interest in the particular topics you tag, as well as to your friends and friends of friends."
Tagging does it, then. When I search Facebook for "Frederick Barthelme" or "Frank Black" or "Rowing," the result returned is either a profile (for Barthelme) or a fan or "interest" page. When Questions kicks in, one of the categories will be Questions and Answers. Screenshots of this feature are available in Ross' blog post.
I'm not bowled over by Facebook's internal or site-wide search feature now, so I can only hope search gets more robust in time for Questions.
The New Ask.com
The other big example of the shift in the tone of the web is Ask.com's new features, available for beta testing by invitation. It has rolled out a new paradigm for old Jeeves & Co. just in time to compete with Facebook Questions.
Ask.com has now re-imagined itself as something of a "decision engine" (ring a bell?). But as it happens, Ask.com does a damn good job of answering questions. It pulls in blogs, syndicated news sources, and even has ported the entirety of Wikipedia into its own domain (see http://www.ask.com/wiki/Traditional_Chinese_medicine for a sample). "Where's the best pub in Fort Worth?" brings back useful info from Ask. Rife with "Featured Sponsors" (clearly labeled), Ask still manages to present top-rated results with contact info for each business, and locations on a map. But the ratings? These are generated by third-party site users from Citysearch. The maps? Microsoft-generated. Ask.com doesn't have Facebook's millions of dedicated netizens-it has to ask for help from the open web...and from other registered Ask users.
Your Ask.com profile keeps track of what questions you've asked (basically, your search history), what questions you've answered, and lets you track popular questions from other user profiles. This reminds one of Search Wikia, but (again) Ask.com lacks a built-in network of "friends" and has no apparent way to ready-port relationships from other networks.
All this makes Ask.com something of an answer aggregator, rather than an answer generator. It pulls in the sources of information that most likely match the terms in your query, and that looks good-it looks like an intelligent response from some kind of systematic ontology. But it's not.
Others are in on this, too. Bing tried some marketing along the lines of answering personal and daily-life type questions ("bing and decide"). Hunch has played its hand (as you may have read here), and its efforts have even lately been covered by WIRED (see August's piece: http://www.wired.com/magazine/2010/07/ff_caterina_fake/). Hunch ups the ante on search "decisions" by openly, earnestly, and rigorously creating a profile of each user who asks Hunch a question. Hunch doesn't just play at answering questions. Hunch has built an architecture that allows it to answer questions to/for each user profile with exactitude and authority. So far, no other service can claim to do this.
This seems to be a peak season for interest in accessing the web on a human-language level, and for expecting intelligent, specific, and original output for natural language queries. This precurses the intentional and orderly schema of the Semantic Web to come.
Ask.com, Facebook, and even Microsoft for that matter, seem to have decided that when folks search the web, they search in their own folksy voices-or that they should and that their services should respond in kind. And it's clear that Facebook and Ask.com will compete to position themselves as a first choice for asking for answers on the open web.