On Sat, 8 Nov 2008, Mark Kimpel wrote: > I found the article the "Y of R" in the latest R news to be very > interesting. It is certainly challenging me to learn more about how R works > "under the hood" as the author states. What is less clear to me is whether > this approach is primarily for teaching purposes or has a real world > application. What is meant by "fragility of reliance on the function > name defined as a global variable" as a downside to the classical recursive > formulation of function "s"? How can that impact the average R programmer? > > Beyond that, empiricist that I am, I decided to put the examples to the > test. My source code and output is below, but the bottom line consists of 2 > observations: > > - The Y function approach using csum is consistently slower on my machine > that the s function approach > - The Y function using csum gives recursive error with high input values > just like the s function does > - The Y function in fact reaches the limit of recursion BEFORE the s > function does > > Given that it is slower, is more cumbersome to write, and has a lower > nesting limit than the classical approach, I wonder about its utility for > the average programmer (or somewhat below average programmer like me). >
Thanks for your comments and to Gabor for some clarification. Your empirical study adds to our knowledge of the situation. I considered the implementation of Y in R to be of conceptual interest only, and I probably should have said that. Even the conceptual considerations may admit of improvement, as there are use-mention distinctions that are murky in various points in the text. But I will not be able to revisit this, apart from dealing with major misconceptions if such exist, in the foreseable future. ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
