Social networks are the crest of the internet wave at the moment. They're even being credited with revolutions - literal ones like Tunisia, not just metaphorical. Personally, I've never really felt the attraction, but the phenomenon is fascinating, and I am now intrigued by examining how the structure of the process affects the discourse involved. To be more specific, consider the following categories -

'Linear' - the most usual, where all posts are simply arranged according to time of posting   [sounds 'real' and spontaneous ... but hard to track down specific comments, or extract specific themes & arguments]

'Staggered Linear' -  as used in IB online training - comments are entered as direct responses to a specific previous comment, creating a pattern of ever smaller branches, twigs and twiglets  [good for immediate relevance & responses to particular points ... but can't cope with complex argument: if your comment also involves other comments, it is hard to follow the cross-references]

'Topic Linear' - as for linear, but arranged in 'threads', i.e. specified subject areas  [easier to understand relevance ... but perhaps rather compartmentalised, and can't cope with interesting digressions or complex relations between topics]

I'm sure there are more species, but I would like to think up a structure which was more flexible, and made multiple linkages and patterns more accessible. Which brings me to the title ...

Tag Cloud Chat  - basic comments are entered as Linear, and may be stored as such  ... BUT no comment can be accepted without, say, a minimum of two tags entered. This generates a tag cloud, and when you click on one of the tag-words, you get all comments relating to that word / concept, arranged by time of posting.  This means that (a) people who enter comments can emphasise mutiple relevance and cross-reference, and (b) people who wish to follow the argument can select, as they wish, which aspect of the general discussion they want to follow in detail - and then change to look at another aspect.

How easy is this technically? In my ignorance, I can't see any real problem, using existing software ... but Jordi, our brilliant InThinking Webmaster, will surely be able to give us the answer. The point would be to achieve a balance between the flexibility of human thinking and the precision and speed of computer thinking.