Plenty of reading and researching morning, this time looking more closely at existing models of computer-automated fiction-editing software packages and services. I started by reading up on story-generation research projects – the highlight of which were a couple of papers by Neil McIntyre and Mirella Lapata (both based here in the School of Informatics): ‘Learning to Tell Tales: A Data-driven Approach to Story Generation ‘ and ‘Plot Induction and Evolutionary Search for Story Generation‘. Then I moved on to googling my way through the wide variety of ‘writer’s tools’ on offer via the internet as free downloads, one-off software fees, or automated editing services. A couple did come up which seem to be well received across the writing industry, though I have no personal experience of them: Scrivener (for putting a first draft of a full-length MSS together) and Final Draft (for screenwriters wanting to format their work to industry specifics). What I was really looking for, however, were suites of software for text analysis- plenty out there for non-fiction, few for fiction. Then I found FictionFixer: according to the sample evaluations, they’re directly in the area I’ve been looking into, but it was also interesting to see that the emotive-analysis we’ve been shuffling round the table is not a route they currently offer. Neither are the kinds of interfaces I’ve been thinking about: as a tool for writers, it seemed to be a very data-heavy output.
Some text-analysis demo highlights to share: a ‘guess the writer’s gender’ website, and an ‘assess your readability’ website. And a few lists which will point you to some of the other tools available for paying audiences (some offer free demos/trials).
The rest of the morning was spent refining my thoughts about potential relationships between the traditional 12 step hero quest narrative/journey and research project grant proposals. After lunch I went to visit Word Power books (conveniently located 2 mins walk from the Forum) to get a couple of postcards, and ended up accidentally buying The Suicicde Shop by Jean Teulé, a dystopian-looking novel I managed not to buy a month ago, but which I’ve been wondering about ever since. The turning point came today when I opened it on a random page and found a reference to Alan Turing (father of computer science/AI), which I decided was enough of a sign to justify the purchase.
Never ask someone to do something you haven’t done yourself. As an exercise in thinking through the sequences of the Hero’s Quest, I mapped out the different ideas & sketches I’ve had about Ada and Gil onto the 12-step pattern. It was quite fun, and turned what has been taking shape as a series of fragmentary conversations into the framework for a (heart-breaking) mystery story.
Back downstairs for the latter part of the day, to meet with Jon Oberlander and Jacques Fleuriot to talk about the hero quest/ research masterclasses. One interesting pathway which came out of that was the applicability of the model to a wider range of research activities – conferences, personal development – and also the importance of Step 3 in the quest when the hero refuses: time management! The Hero narrative is designed for existing stories and proposals are hypothetical projections, so there’s some work to be done in re-wording and identifying appropriate connections there.
Jon and I then swapped Jacques for Clare Llewellyn – a new PhD student in the School – to talk through ideas around using automated information retrieval to assess academic paper titles, with a view to maximising paper-acceptance and paper-citation. There’s existing work in punctuation which Clare has sent me some links to, so more on that next week. In the meantime I’ll leave you with my favourite new phrase of the day: ‘triangulation within positive wordspace’.