So I installed this software at work and have used it a few times on some fairly complex documents, as well as on some more straightforward ones. It’s o.k. as far as these things go – I guess I was hoping for a little more vulcan mind-meld type functionality; the reality is that even with the summaries it produces you still have to work through these, and they can be lengthy, and I question whether I don’t naturally do a better job myself in skimming through documents (e.g. it seems mostly to look at work frequency and placement, and summarize based on initial sentence in paragraph, though it is doing some analysis of structure that enables it to grab phrases that are buried more deeply.)
Still, an interesting experiment; when I get some time I’m going to dig further into the field of latent semantic analysis (partly based on Bruce’s recommendation that I do so) to see what developments there have been of recent days in this field. This as part of my belief that the solution that will emerge to find order amongst the chaos of the net will be a combination of automated techniques and methods that analyse actual behaviour of users and help aggregate their decisions and choices. – SWL