Here’s something thats been exercising me of late. We’re starting to work with a group of linguists who are keen to explore the use of a more flexible programming tool than the one they currently use (R) to analyse some large textual corpora.
As they are already aware that the tool they currently use has some restrictions, they’ve been self motivated so far and have been very open to the opportunities that Python and the NLTK (Natural Language Toolkit) can offer them. What is slowing them down is the realisation that it will need them to develop some new skills and for busy researchers (and many with no programming experience) this is proving to be a stressful time.
What we’ve done so far is to work with them to work through a few standard use-cases and provide some simple code templates so that they can learn some coding skills from scratch and then apply them straight away with concrete examples. By learning to program both ‘top-down’ and ‘bottom-up’ it has helped to reduce the steepness of the learning curve. It seems to have been a relatively successful approach so far and it is interesting to have a linguist’s take on the nature of Python.