Filtering the Real-Time Web
The massive flow of information delivered by the emerging real-time web has some important inherent value, but building added value through intelligent extraction of information from that flow is the next step that online services are beginning to take. Social interaction design specialist Andrian Chan explored what this might look like in an excellent article today titled Activity Streams: Content and Flow that we'd like to highlight as our Real-Time Web Article of the Day.
"I think there are two distinct trends at work here," Chan writes. "One, the popularity and adoption of the stream as a form of social conversation. And the other, the conversion of realtime information into value that can be consumed outside the stream. Or to put it another way, the value of being in the flow, and of watching it from the river's edge."
In preperation for our forthcoming ReadWrite Real-Time Web Summit on October 15th and the publication of a major research report on the topic, we're highlighting one great article on the topic each day and briefing key companies working in this space. Chan's articulation of the things made possible by systems that filter the flow closely resembles what we're hearing from others we're talking to - filtering to add value is going to be key.
We're big believers in the value of the full, noisy flow as well - but the value latent in filtering is just as important.
"We can easily imagine a wide range of activities that are currently page or site-based being handled instead by the stream," Chan wrote today.
"Invitations, meeting requests, buying and selling, questions/answers -- these and much more could be transacted by means of messages 'off the page' and extracted or sorted out of streams by smart clients or aggregators. Analytics companies will have a gold mine of relationship data to scrape and visualize, for example, for use by those who want to see how influencers reach their audiences, around what topics, how quickly, with what redistribution, and so on.""The primary goals of interaction models used around the flow involve separating content from the conversational stream, extracting meta data where possible, assigning categories and embedding within content structures and navigational systems. Then the social challenge becomes making it accessible (search, browse, and categorization) and making it socially interesting (lists, rankings, votes, etc)."
Chan's full post goes into greater detail about both the value of the whole stream and what filtering could look like.
Chris Messina told us something similar in an interview today. "The river of news model doesn't have the handles regular people can grasp," he said. "It needs to be delivered in a way that respects your time, enhances the content or makes it easier to consume the content. My mom would like iPhone push notifications of pictures of me, for example, so she could comment in reply immediately. I hope this information can be made more actionable."
How do you picture services applying or enabling filtering to add value to the flow of the real-time web? Maybe you're a service provider working on this and maybe you're a consumer/participant excited about the possibilities. Let us know in comments, and if you can we'd love for you to join us for face-to-face conversations like this at the ReadWrite Real-Time Web Summit.
Share
Facebook
Comments
Subscribe to comments for this post OR Subscribe to comments for all ReadWriteWeb posts
This is how I picture:
http://nick.iss.im/2009/06/16/sharing-information-services-and-interactions-in-the-next-decade/
Its *context* that's going to be the challenge ( with filters are one of the tools that will provide context ).
Seeing a picture of my Auntie in real time, via my Flickr RSS, has little meaning beyond the minor "awwww" factor...
...but the pic's EXIF data, the geo RSS, its comments, the number of favorites ( favorited by whom? ), etc. arriving in real time, relentlessly, multiplied by one's entire social graph THAT will be the polar bear that wants to wrestle.
Definitely an interesting challenge, and one I don't think the majority of real time web users care enough about at the moment. But in time they will, as their real time streams move from being purely a conversation session as you describe to being a collection of meaningful information that they wish to interpret in differing contexts.
For me, a tool like Gist.com helps with the process (you guys have blogged about it a couple of times, and I did the same) but it isn't there yet.
There are certainly going to be interesting times ahead in assisting people in making sense of all of the "information" that we are collectively creating.
I know realtime is the new, new thing, especially in the Valley, but it has a huge limitation. There is no time available to develop meta data that separates the wheat from the chaff. Yes, it was realtime, but there was no decent filter. So we can use reputation right? What does that actually mean? Please don’t say number of Twitter followers. I get followed by “people” on Twitter with tens of thousands of followers, and 0 updates. Even if you came up with an algorithm that dealt with that type of problem, it can’t be based on reputation for the realtime subject, because reputation can’t develop in realtime. It has to have a past to give people’s reactions time to develop. But yes, right now if you say realtime, you certainly can get funded. In the Valley, at least.
@freisprecheinrichtung bluetooth
I know of one fairly large group of people that have come to expect ( demand? ) real time data - ebay bidders
If there is a group ( ten of millions of them ) who would appreciate new types of real time data, it would be them.
The firehose has its uses, especially in terms of trend detection / aggregation.
For 99% of folks, however, a filtered view isn't only preferred, it's required. Most companies simply don't have the resources, time, or desire to keep up with the massive torrents of information from Twitter/Soc-Networks/etc.
There's huge potential for solutions that help manage this information flood.
Rather than focus on user-facing apps, at Orchestr8 we've been building *capabilities*: semantic content mining & NLP web crawling technology exposed to our customers via a web API.
AlchemyAPI is being used by a companies across multiple verticals: social media monitoring firms, news aggregators / hyper-local news portals, trading firms, etc.
We're seeing folks do all sorts of interesting filtering operations against the real-time web: Topic-specific "filtered" Twitter streams, blog monitoring, analysis & routing of SEC company releases, etc.
Filtering is great and has been used in Knowledge Management for years. The problem lies with the taxonomy and getting everyone to agree that an apple is an edible apple or a computer.
It just goes downhill from there on a large scale.
Creating a filter is identical to posing a question. I can only do that from known facts and todays thoughts, so in one way or another filters will only "give me more of what I already know".
When someone comes up with a theory that the world really is round, an maybe older than a few thousand years, I might miss that exciting new thought.
So we lock us in and depend on making small windows of serendipity windows in the knowledge wall.
It's like a prison, it is :-)
Posted by: https://services.mozilla.com/openid/jchris
|
September 19, 2009 3:25 AM
Filtering Will Be Key in the Real-Time Web http://bit.ly/2aN2JH (RTW Article of the Day) [from http://twitter.com/marshallk/statuses/4041389930]
Posted by: Marshall Kirkpatrick
|
September 24, 2009 10:15 AM