[...] written previously about how social search won’t replace traditional search, how social relevancy rank may be used to deliver good results, and why the concept of social [...]

[...] I discovered when working on our Social Search paper. It has been argued before (one small example here) that Google’s PageRank algorithm is inherently social, as it aggregates information provided [...]

[...] written previously about how social search won’t replace traditional search, how social relevancy rank may be used to deliver good results, and why the concept of social [...]

[...] Personalmente,opino que la búsqueda social no sustituirá a la búsqueda tradicional, que un alto índice de relevancia social puede usarse para obtener buenos resultados de búsqueda, y que el concepto de búsqueda social supone una vuelta a un entorno familiar, y no hay nada que temer al respecto. Hoy hablaremos más específicamente sobre los tras “sabores” de búsqueda social que mejorarán la experiencia de búsqueda de los usuarios. [...]

[...] written previously about how social search won’t replace traditional search, how social relevancy rank can be used to deliver good results, and why the concept of social [...]

You rightly point out that “discussions” are a key component of social search — it’s not just about getting a piece of information, it’s about having an information exchange with a real person, that’s what makes it social. Sometimes you just want to have a conversation with someone who knows what they’re talking about, who shares your general taste… the best Social Search solutions will make that effortless and instant.

And to correct the record about Aardvark a bit …
- Aardvark is real-time (works via IM chat)
- Aardvark connects you to people with experience relevant to your question *in your network* (ie. friends and friends-of-friends) — so “experts” is a bit misleading.

Looking forward to seeing where your research goes!

I think you’re beating down the right track. What are the particular failure modes of current internet search, and which of these can be most successfully approached via social mechanisms? I agree with your point that early-phase information seeking tasks might be a great entrance point, since often this is where I don’t even know what my query should be yet. Often I find that asking a friend what I’m looking for gives me the terms to get started.

I see two issues still. One is that user needs are still not being met by existing solutions (however “social” or “unsocial” the algorithms may be). The other is that most of the social solutions that users do have access to today are in the style of collective harvesters. You’re right that Google’s search results are social in the sense that they are filtered by webmasters, but is that meaningful to you? I think it would be more meaningful is information were filtered through *your* social graph.

This issue is related to how important trust will be in social search. One paper (albeit in the organizational learning literature) found that trust was important for open-ended searches maybe more so than for well-defined problems. If this is true on the web, then the massively pooled social data that are Google’s current results, hence lacking an element of trust, may not be useful for the majority of searches that are exploratory in nature. (Although I think having personalized social recommendations could help with well-defined search queries too.)

Hi Brynn,

I’ve always thought of Google’s PageRank as the ultimate social algorithm. (A page is ranked higher when more people vote for it by placing a link to it.) Of course, the “people” here are not web users, but webmasters, but that only narrows down the set of people who you’re getting your search recommendations from; not the fundamental nature of the ranking algorithm.

How does this viewpoint supplement/coexist with your thoughts about social search and a pure non-social search engine?