subject.cz/josef

Quite recently, some of the world’s best bloggers (Note: ‘best’ in the ‘number of readers’ metrics) were involved in quite an emotional discussion about improving the twitter search by adding some notion of author’s equity (TechCrunch, Scoble, Scoble, TechCrunch and TechCrunch again).

Two equity metrics were discussed so far:

1. Number of followers the author has.
   
2. Number of retweets for an individual tweet.
   

As Scoble rightly pointed out, deriving the importance of tweets by number of followers is tricky, mostly because it doesn’t express author’s competency in a subject domain the concrete tweet is from/about. He is giving an example of himself tweeting about ’supply chain management’, which is certainly not the reason why most people follow him on twitter. He also rightly suggests to use the metadata (like retweets, favorites, links, clicks, reshares on the friendfeed,… etc.). However, all these solutions are just very basic and incomplete, as the problem is more general. It is IMO generally about measuring the quality of contributions to the community.

This is what the Community Equity (CEQ) is all about. Yes, we have it implemented in our internal SunSpace project, which is a content-centric wiki system with social networking features, but it could be very easilly applied to the asynchronous communication use case of twitter, friendfeed, blogs and possibly of the whole global social network as such.
It is however important to realize, that twitter is not a standalone service - it plays quite an important part in the ecosystem of myriads of various online services, where users create, share and discuss various types of content. Thus, measuring the quality of twitterers and their tweets could be done by the following simple semantic mashup of:

1. Author’s contributions to various content services, e.g. blog posts, videos on youtube, images on flickr, slides on slideshare, …, and yes, tweets on twitter. The more services are used, the better (=more accurate) the results.
   
2. Metadata related to these contributions, like ratings, comments, downloads, reuses, …, retweets, numbers of followers, and last but not least tags. Some of the metadata are more or less specific to the individual services - for example ratings are quite commonly used across different services, while retweets are specific to twitter-like ones only, etc.
   
3. Tags define topics and thus subject domains the contributions are from/about/related to.
   
4. CEQ service implementation(s), computing equity values for individual contributions, contributors and consequently tags (thus answering the question What is author’s expertize/competency in a given domain/topic?).
   CEQ implementation may be generally different for different (types of) services. In SunSpace, we currently have it for the document-centric wiki use case, where things like ratings, downloads, views and reuses are used as inputs for the computation. This relates to the metadata specific to different services mentioned above.
   This btw. means, that Scoble and others are actually discussing the meaning of metadata (like no. followers, retweets, etc.) in the context of the twitter service. Its obvious that there are not enough types metadata in twitter - for example no ratings and tags - and that’s why it needs to be mashed up with more services…
5. Search using CEQ values for filtering/ordering of search results.

We have already been working on making this part of the next SunSpace release rolling out sometime early this year. To have this fully working on the public web, full-featured semantic web with shared concept models and open social networks would be needed. And that won’t happen soon.
But it makes sense to start with small now - the discussion has already started, which is good. It would also be nice to allow users to tag and rate tweets. Forwarding (=reuse) of tweets may be another feature worth considering.
Btw. - did you know that CEQ will be open sourced and public soon?

This entry was posted on Friday, January 2nd, 2009 at 1:07 am and is filed under Uncategorized. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.