I've posted a [Scraperwiki scrape of the BBC In Our Time archive programme pages][1] that includes a column containing suggested reading items from the most recent programmes in an unstructured way; the single books column may contain multiple book references, in a variety of unstructured, though occasionally conventional, forms.
The scripts Orangeaurochs posted contains detectors and parsers of some of these unstructured formats, but I haven't integrated them in the scraperwiki - instead, the scraperwiki database may be regarded as a resource that needs parsing...
It is possible to run the In Our Time programme descriptions through Reuters' Open Calais system to identify the names and affiliations of guests included in the programme with a reasonable degree of reliability. (I wonder, should I try to parse out a column containing the paragraph that references discussants?) An example of what we can do with the participant data is described in the post 95[Visualising [Visualising OU Academic Participation with the BBC’s “In Our Time”][2]95 Time”][2]
[1]: http://scraperwiki.com/scrapers/in_our_time_books/
[2]: http://blog.ouseful.info/2011/01/21/visualising-ou-academic-participation-with-the-bbcs-in-our-time/