[plt-scheme] Scheme and R

From: Neil Toronto (ntoronto at cs.byu.edu)
Date: Thu Mar 26 14:05:31 EDT 2009

Prabhakar Ragde wrote:
> Noel Welsh wrote:
>> On Thu, Mar 26, 2009 at 5:24 PM, Neil Toronto <ntoronto at cs.byu.edu> 
>> wrote:
>>> As a language it's rather weak and inconsistent.
>> I agree with this, and the rest of Neil's analysis.  It would be
>> simple to create a PLT Scheme language that extends the numeric
>> operators to vectors in addition to scalars. Perhaps one should also
>> replace the list operations with the vector equivalents. This could be
>> used to introduce R-esque programming.
> Thanks, Noel and Neil. I haven't seen anything to contradict these 
> points, or what Sungwoo Park says. I have zero influence on the choice 
> of the stats instructors to use R, and they seem to be doing little or 
> no language teaching, relying on Web materials and simple examples in 
> the text of assignment questions to guide students in their solutions. 
> The quality of the tutorials and books I have gathered so far made me 
> wonder if there was a way to take advantage of students' prior exposure 
> to Scheme to help them in their use of R. --PR

I would guess that the most difficult thing for the students to get used 
to is working with library functions that vectorize everything - that 
operate on lists or vectors of samples rather than single values - and 
then writing their own code the same way. I think what Noel is 
suggesting is to show them this in Scheme first as preparation for 
programming in R.


Posted on the users mailing list.