%in kempton park+277-882-255-28 abortion pills for sale in kempton park
R scala 17_05_2014
1. R / Scala
Ruslan Shevchenko!
ruslan@Shevchenko.Kiev.UA!
!
@rssh1 !
https://github.com/rssh
2. R/Scala interpolation: Why ?
❖ Packages !
❖ for near any thing in statistics exists
appropriative R package. !
❖ Culture!
❖ math-centric community!
❖ quick evaluation. !
❖ more research than development !
❖ Fun!
❖ cross two nontrivial (and very funny)
languages
3. R/Scala: context
ScalaNLP:!
Breeze: + something in REPL
scala-centric point of view
• build environment!
• enhanched REPL + graphics!
• wait until math people switch to one ;)))!
• better port only things which will live in
process mode.
4. R/Scala :context
R is beautiful More like human language than computer!
• 40 years of history!
• parts in C and FORTRAN!
• functional, immutability, lazy evaluation!
• antic implementation, crazy conventions !
• Object systems:!
• S3, S4, R5, R-proto!
All live together ;)
R is ugly!
• Slow!
• programming ‘in large’ in hard.
7. Simple, just add jar
unmanagedJars in Compile += {
import scala.sys.process._
val jarPath = List(“R","--slave","-e",
"""library("jvmr"); cat(.jvmr.jar)”""
).!!.trim
file(jarPath)
}
!
8. JVMR
Slow, run scala compiler in runtime!
if we want to call scala from R!
!
Integration via ‘string’ representations!
Very simple, in many cases it’s enough!
!
‘native’ R
9. RENJIN
❖ R interpeter in JVM!
❖ compiler toolchain for C and FORTRAN parts!
❖ gcj - based.!
❖ work only on linux.!
❖ R package with C parts => jar