Re:Machine Learning for Hackers Table of Contents
by programnature on Thursday February 09, 2012 @02:38AM
(#3570184)
The problem with domain-specific systems like R is that, as software systems, you can't "hack" them far outside of their contraints.
Integration into other systems (like the web, or robust big-data pipelines) can't be taken for granted. Doing "standard" stuff becomes nonstandard, thus hacking can hit a wall.
Re:Machine Learning for Hackers Table of Contents
by simonw on Thursday February 09, 2012 @04:11AM
(#3570359)
R is open source, and there are plenty of "bridges" to talk to it from other languages (e.g. http://rpy.sourceforge.net/ )
Re:Machine Learning for Hackers Table of Contents
by plessthanpt05 on Wednesday February 08, 2012 @08:23PM
(#3569391)
I use R almost daily, and sure, for tasks involving math (& statistical analysis in particular) R is fantastic; however, to say "for hackers" (at least in my mind) would involve a less domain specific language -- like a scripting language (python or some such similar language) that has statistical analysis capabilities. Perhaps a matter of semantics, but in a similar vein, I imagine a book of the same title that used Octave would not be considered "for hackers" by most due to the specific/constrained functionality.
Just seems like "hacker" is such a buzz word now that it gets tacked onto a lot of stuff just to generate interest.
If anything, it seems like a language more suited for hackers than Python for tasks involving math--it looks like it has a more integrated environment for fast iteration , particularly for generating nice plots of the data you're working with. It also seems like a language designed and optimized exactly for this sort of activity.