TTT a DATA/POEMS/WORDS The output of the robots can be treated in a simplistic way, like a data set. For example, it took six days of exhibition to get a partial sentence of coherent English: the phrase 'or deem to ask.' One of the things the project sets out to do, is to ask whether language might emerge from a combination of software and numbers. This first partial sentence might be grounds for cautiously thinking yes. Recently the phrase 'dad my map is pup' resulted, which clearly has one form of English sentence structure, if not the appropriate words. How long before we get structure and content together - that is the question. 'Nick Sow is so oh' might be a phrase said about someone, but we are after something a bit more complex. The Information Officers at Puke Ariki museum, all keep an eye for long words. If you look at your phone when typing a text and using predictive mode, you will see that the more times you need to hits the keys for letters, the fewer options there are for words. So getting a long word means that the robots have to move to precisely the correct order of cylinders, to get the word. Given their path is upset or randomised by bumping into eachother, this means that randomness must interact with the programmed actions of forward, backwards, turn left, turn right. It has amazed me that such a simple set of commands would generate words let along string a number together to make a strange kind of poetry. The orginal intention was to incorporate a fractal or Julia set into the movement commands, but given text and sentences were generated with such simple instructions, things were left as they were. If we do incorporate fractal or Julia set trajectories, it will be for another iteration of the project. While six or seven letter words are relatively uncommon, truly oputstanding have been two eight letter words output in the first two weeks of exhibition. The words are 'glimpsed' and 'infidels'. 1. 2. 3. our 4. 5. 6. 7. sos 8. 9. 10. pi 11. 12. worm 13. 14. 15.
|