Challenge - Cohesive blocking
to Doug 14:32, 3 July 2008 (PDT)
Hi Doug, I haven't had a chance to look too deeply into this, but I did try to run the example and it was not a fast result. The code seems to be running as it should, and when I kill it a few minutes in it's just at a point deep in its recursion... I'll give it some more thought later, but for now I would guess that it could be a simple efficiency issue of R vs SAS. I've thought about reimplementing it in R without using recursion (not R's strong point), but that's not something I could take on in the immediate future. But again, I'm not sure where the holdup is with this example. I'll try to give it a deeper look over the next couple weeks as my summer calms down. Thanks for finding this, Peter
Here is the http://intersci.ss.uci.edu/wiki/pub/kconfilesNEW.zip (right click and save) full set of programs that James Moody is release into the public domain -- his SAS program for cohesive blocking Doug 14:36, 3 July 2008 (PDT)
Do we want to outsource (I have funds for this) to a place like Data art in Russia @ http://www.dataart.com/ ?
Scott D. White's advice: Knowing now that Peter has done an R version, it is much much easier to go from R to Matlab than from SAS to Matlab: there must be 100,000 programmers who know R and Matlab.
Natasa Kejzar knows both.