Iterate, iterate, iterate, and at the same time: sculpt, sculpt, sculpt! Using a Generative Grammar But taking the text at face value can introduce many problems into your project, and I’ll demonstrate why it’s important to craft an editorial voice and vision for your generative text projects, without leaning entirely on the generator. The text generated by a computer is often a great canvas, an excellent block of marble, and a good starting ground for any written content. In the title of this post, when I say “sculpting generative text,” I truly mean that. However, the same texts are often just as quickly criticized for being too unemotional, not taking all facts into account, missing the point, being awkwardly phrased, or all of the above. More recently applied in journalism, generative text is sometimes seen as a way to quickly produce editorial content without the bias of a human. When people talk about generative text, they often put machine-generated and human-generated passages side by side and challenge the reader to find the difference.
Sometimes, a subtle application of generative methods allows for unique experiences to surface in otherwise boilerplate content.
RANDOM SIGIL GENERATOR GENERATOR
While both are excellent examples of PGC, the generator doesn’t have to dominate the game or be the core of the engine in order to be effective.
When people talk about procedurally generated content, they often think of sweeping, monolithic titles like Spore or Dwarf Fortress. Generative text regularly receives attention in scholarly articles and real-world applications, but in the greater category of procedurally generated content, it sometimes gets lost in the mix. My experiments have yielded some ideas and practices that I’d like to share as I discuss techniques for sculpting generative text, specifically with Kate Compton’s Tracery. As a linguist, I’ve studied generative text for its abilities to trace new connections between things, people, places, and feelings in a way uncharted by humans. Even if we sometimes perceive emotion in a machine’s words (thanks to the ELIZA effect), a computer will never learn language the same way a human does, so can they deliver narratives with the same authenticity, accuracy, and appeal as a person?Īs a programmer, I’ve dabbled in generative text before as a method for creating novel interactive content for users, typically for games. We see words as a way for a human to express their thoughts and feelings, and while all words have concrete meanings, there are connotations and sentiments associated with certain words and phrases that we think a machine could never truly master as an unemotional third party.
As humans, we take our ownership of words very seriously.
0 kommentar(er)
