Frankensearch (Intelligent Media Search), a real time search collage powered by social media status
For regular search, finely sifting and weighting pages based on authority and rank is fine. Even if I only see the top 5-10 pages, most likely they'll contain the information or links to the information I'm searching for.
But real time search is something different. We're interested in what's happening, and we want to find out about it, as it occurs. What real time search should provide is a way to view everything that's happening in relation to a specific topic, business or person. We still need to filter out spam in real time, so there needs to be some authority or connection to real, and genuine status messages or links. But what do we use as a measure of trust and value? It's too soon for backlinks or any later structures to have emerged. For real time search to be effective we must trust the source by reviewing their track record. If we allow for some crowd like review or other filtering, real time tends to become near real time and loses it's freshness.
An example of real time search that is well received, and incredibly useful, is conference tagging. People at the conference, or who wish to engage with those at the conference, simply use the specified hash tag to join into the super sized chat room. Many conversations surrounding that hashtag have value, but there's a challenge in separating the signal from the noise. A simple start for social media and blog rating:
- historical measure of meta data or keywords derived from status updates associated with an business account or individual
- some type of influence measure (followers * reaction effect) from social crowd sourced sites
- outside sources of information that are trusted authorities for other non-real time data.
- weighted memory, system momentum