This a brief account of what was discussed at the Cloud 1 meeting. I have also put the slides on this page rather than in the scriptorium which I will keep for 'big doc' rather than thought starters.
We met at the PMA Media Training Centre in Camden. Riva Elliott group member was our host. Present were Riva, myself, Chris Arning and John Kearon. Apologies from Tom Ewing and Mario Menti who had intended to be there.
I showed some slides about the scope of the project you will find these slides in the scriptorium on the webjam. Here are some of the issues that emerged:
What is a text? There was a query about whether the remit of the project is text as in words found on blogs and websites or a text as understood by a cultural analyst – any artefact which can be read and analysed – this would include just about everything on the web. We concluded that it meant the second however there are practical difficulties about linking and tagging to some of the content but in principle all web content is included.
Validation. Another massive topic but the question came what constitutes validation? The reason for bringing research thinking to bear on web content is partly because researchers are not permitted to announce findings which they cannot back up with some form of validation. A classic example is rentaquote where one respondent is taken out of context (and sometimes by the client too!). The researcher will need to show that what that person said represented the perceptions of other people as well. Agreeing what type of validation we were looking for would be difficult but we agreed that we need to find it. I made the distinction between the huntergatherer going to get fresh meat and the scavenger who brings in something dead or near dead they happened to find. Too much internet content held up as a kind of evidence was closer to roadkill whose origins and context were not well understood. So part of the remit of the project is to find ways to validate the content that we use for research.
The 4 components of the Cloud of Knowing project were usefully summarised by John Kearon – sourcing, hunting, grading and profiling.
Following on from the project slides. I then gave a brief description of the Purefold project and how that had seemed a useful approach for adapting as a research methodology. I was able to show how the RSS feeds from the original script ideas had been placed in a friendfeed and friendfeed visitors invited to evaluate and link to what had been sourced.
We also went onto discuss Demographic Replicators – another concept produced by David Bausola and Tom Himpe of Ag8. Which channels twitter feeds and blogs into a kind of amalagamated persona who represents that particular demographic, whether a 23yearold or someone living in Camden. We agreed that this kind of radical thinking fell within the Cloud of Knowing project because it was taking webcontent and turning into something which researchers could use. But was an unrecognisable advance on offline research. More like a car and less like the horseless carriages of the first generation of internet research.
Chris Arning gave a presentation about the semantic web - and the multiple levels which applying machine intelligence may give to analysing web content making revelatory connections which humans hadn’t seen for themselves but also generating new ones.
Sampling We closed with some discussion about sampling – using machine tools to guesstimate the identity of the authors of the content we were using and also the identity of research participants. Scoring models have become accepted practice in targeting for direct marketing. Using similar techniques it ought to be possible to use the content that we are looking at to establish the identities of those who created it. The gender guesser I posted on the blog last week is a simple example.
I have also since the session cobbled together a graphic about the Purefold adapt to the research session which is part jet engine and Bessemer blast furnace ! Its probably gobbledegook to you but it may make sense to those who were there.
Next steps - John was intent on exploring how to use Webjams! Riva is interested in how we might use the training she runs for journalists to harnes their talent and their ability to generate some publicity! I am intent on turning some of these ideas into research 'products' of which I reckon the jet engine blast furnace is an example. I would also like to do some more work on the Demographic replicators. Perhaps this should be a topic for the next meeting.