On Friday, April 12, 2013 6:42:15 PM UTC-7, Mike Anderson wrote:
My suggestion would be to think hard about how to break down the features into the simplest possible operations. The "Clojure way" is often about composing simple functions in powerful ways. Obligatory viewing: http://www.infoq.com/presentations/Simple-Made-Easy
Thanks, That's very helpful information. I can see how some of the existing bioconductor packages already have interfaces which work with this paradigm and some do not.
I am really excited to write data analysis libraries entirely in the same language we do the analysis itself in. I really love R, but I think having to write many of the packages themselves in C for sufficient performance makes it take much longer to develop new R packages.
I can see that writing functions to handle large sequencing data efficiently will be one of the most challenging issues, but I'm more in need of functions which can handle small molecules, and biochemical pathway data.
On Friday, April 12, 2013 7:31:24 PM UTC-7, Ben Mabey wrote:
In the early days of clojure there
was an Anathem meme going around and a number of libraries were
named after things and characters in the book.
Well I think I've found the right place then, this is one of my favorite books! I've heard that Neal Stephenson is himself a LISP programmer.