Tags
as it turns out Barefoot Running Blogging Clogger Clojure Computer Science ebook Enlive Framework Functional Programming Gajure Genetic Algorithms Git hacking Haskell Heroku Incanter initiative iSight Lazy Evaluation Lisp Macports Math Open source OSX Package Management paul graham police Rails Rails3 Ruby running Security Shell Scripting sinatra Snow Leopard Spam startups success Syntax Thunks twitter web application Wordy WritingFollow Me?
I am @Unignorant
On Github
-
Recent Posts
- Rails3 Take Two: A small budgeting app
- CS Grad School Applications
- Listeria: Testing Rails 3 and Heroku
- On Standing Desks
- Nytimes Oracle (a Markov text generator)
- Security: Simultaneously Weak and Amusing
- High On Lisp
- Thunks and Haskell
- The Allure of the Asymmetrical
- Analyzing Word Frequencies with Clojure, Enlive and Incanter
- As it turns out is quite innocuous
- Gajure Now on Clojars
- Police Pursue and Capture a Barefoot Runner
- On Initiative
- How I develop on OSX
Archives
- August 2010 (4)
- May 2010 (1)
- April 2010 (3)
- March 2010 (3)
- February 2010 (4)
- December 2009 (4)
- October 2009 (4)
- September 2009 (1)
Recent Tweets
- A few resources for CS Grad School applications http://blog.ethanjfast.com/2010/08/cs-grad-school-applications/
- Blog about Listeria (rails3 + Heroku FTW) http://blog.ethanjfast.com/2010/08/listeria-testing-rails-3-and-heroku/
- RVM makes managing ruby gems so much easier http://rvm.beginrescueend.com/
- Just put together a good standing desk http://blog.ethanjfast.com/2010/08/on-standing-desks/
- Thanks to the creator of Vyquon for a lisp dialect which can actually be parsed through CIL http://github.com/overzeroe/Vyquon
Nytimes Oracle (a Markov text generator)
From this comment on HN, I came across a well-written article on generating sentences using a very simple Markov algorithm. The basic idea is that a body of text is divided into word-pairs, where the frequency of a given pair determines the likelihood of the generator producing the second word in that pair when presented with the first (a sentence ends when the generator encounters a period). It’s easy to construct sentences by randomly selecting a starting word from the input text, and feeding it to the generator.
Naturally, the kind of sentences that are output very much depends on the content of the input. For this reason, I have mined the ineffable wisdom that constitutes today’s Nytimes oped page, and fed it to my generator (its code is available on github). Much of the output is nonsense, but it occasionally produces some clever phrases. Here is a sample: