Tim O’Reily just compiled some ideas on “Huamn vs. Machine: The Great Challenge Of Our Time“. It actually suggested that an emergent inflexion of web 2.0 to be re-invented to a higher stage. The next stage will be the well designed mashup of human computing and machine computing which I used “P2R Computing” to described it before. I think it’s better to expand the idea to reflect Tim’s one.
Google was being very successful in the last past 8 years. They found a basic fact of Internet content democracy and formed the super search algorithm “Page Rank” to enable thousands of machines to work in a parallel way to support large scale indexing and searching request. The creative design of Adsense model helps Google collect money from large scale eye balls along with the booming of search behaviors. Undoubtedly, Google can earn more from the “Big Number” effect and for more years.
However, what I’m sure is that Google merely itself can’t bear is the global change of content structure since the new paradigm that Web 2.0 leads, whatever their investing in more machines, adjusting the Page Rank algorithm, or more activities to be more 2.0 alike. The problem to Big G is the internet content granularity has been dramatically downsized by both people and new kinds of technologies(like RSS and Atom). The connectivity between information will be not only described by physical links between web pages, but also a new layer of social filtering yet to be articulated by more new innovative social applications(like those RSS readaers, twiter.com, soup.io, etc.). Some recent found fake sites can present itself as a real content destination from Google’s search results, however, they are actually generated by machine programs to make money for their authors. The downsizing of virtual hosting and domain costs makes such things easily be mass produced. So actually, Google’s machine algorithm is now facing their distributed siblings from other hosts everywhere. Of course, people will be unsatisfied more and more.
Social Search, a new but long discussed concept is the right time to emerge in the coming two years(Am I optimistic?) Just a quick simple example first: When I search Google about “Human vs. Machine“, I can find Tim’s article, but obviously it’s not at a relevant position in my mind. And I can’t find any referential information to this search result. I switched to another mashup tool with the same query string and found more interesting results. I’m sure some of the links were really what I want and most importantly my own mind told me some are really relevant with some of the names I know and trust. Recently I compared many such just-in-case searches for in real context, for either blog composing or reference search, the “nearly social” search wins more than Google. Google may argue that skillful users can also realize social search based on their algorithm, but obviously Google didn’t index well in some cases I met (though I found Google added one missing index days later). The fundamental problem is Google still slowly adopt social layer in their base service. The missing link, Social Ranking(SR) , should be be introduced into future search system.
The Social Rank(SR) concept, which can be a perception inherited from Page Rank, yet to be developed deliberately based on one’s Social Portfolio(I dislike the term “Social Graph”, either from semantic or public acceptance) and calculated in a large enough scale to ensure its accuracy. Your Social Portfolios on each social application sites(Flickr, Youtube, Facebook, Twitter, Slideshare, etc.) will be crawled frequently to compile your SR. The more content you generated on those social sites, the higher your SR could be(but not linearly correlated). The trust relationships to connect those “Social Portfolio” will be another key factor to be compiled into your public SR. The people with best Sharism virtue will be recognized in this way. Like Tim, Joi Ito, Stehphen Downes, Dave Winner, Rebecca Mackinnon, etc.
In my humble vision, people will get more information in a streaming way from their daily social context instead of today’s individual keywords-driven machine searches(In some countries, the portal way still dominate, though). Their activities will be logged and fed to their trust circle. Your Social Portfolio with its inherent sensing capability will help you aggregate “right” information to avoid of a problem of “don’t know what you don’t know”. The Micro-pipeline is in shaping and more applications will emerge to support such “information streaming” time by time. Like Soup.io’s experiment.
The basic unit of the fluid in the micro-pipeline will be micro enough to support almost everyone’s participation and easy aggregation and remixture. Both Atom and Micro-format are in this category. But it’s questionable to apply larger granularity work as base content unit, like wiki. So Wikia’s effort could be on a too complex direction at least from the requirement of social searach on mass participation perspective.
Eventually, Social Search will be an entropy game, Google won the first round because they found the algorithm to map the less complicated web 1.0. And who will win next round to resemble the web 2.0 variety?