jesg / dither Goto Github PK

see jesg/dither-java#2

simplify maintence with one core alogrithm for ipog written in c++11. ffi bindings for both the mri and jruby runtimes.

example:

$ rake compile
mkdir -p tmp/x86_64-linux/dither/2.1.0
cd tmp/x86_64-linux/dither/2.1.0
/home/jason/.rbenv/versions/2.1.0/bin/ruby -I. ../../../../ext/dither/extconf.rb
creating Makefile
cd -
cd tmp/x86_64-linux/dither/2.1.0
make
compiling ../../../../ext/dither/dither.cc
cc1plus: error: unrecognized command line option ‘-std=c++11’
Makefile:204: recipe for target 'dither.o' failed
make: *** [dither.o] Error 1
rake aborted!
Command failed with status (2): [make...]

Tasks: TOP => compile => compile:x86_64-linux => compile:dither:x86_64-linux => copy:dither:x86_64-linux:2.1.0 => tmp/x86_64-linux/dither/2.1.0/dither.so
(See full trace by running task with --trace)

need a generation strategy that can unify random walks with pairwise testing. a popular technique is to generate a bunch of potential solutions and take the best fit at each step.

• generator
• filter (constraints)
• best fit

this could be applied to:

• interactions between edges in a graph
• measuring coverage in a graph
• non-deterministic pairwise strategy

furthermore the alogrithm could be concurrent on the jvm.

generate random walks from a markov chain. this could be applied to usability testing and reliability. it would also allow the tester to include more complex user interaction like cycles.