[Rd] S3 vs S4 for a simple package

[Oleg Sklyar]

In EBImage I have a very similar situation (well, with more methods). It 
would be impossible to use S3 in my case as my data structures are 
images and I need at least two dimensions (but in fact use 3), thus 2 
variables to dispatch on in [, which are defined for multiple 
configurations like integer,missing; integer,integer or logical,missing. 
To make things faster in S4, I however did not put data into a separate 
slot, rather derived the class with contains="array", so effectively 
giving the class all the functionality of array, and only changing the 
default behaviour for a couple of methods. Obviously one can access the 
data low level through the @.Data slot, both in C and R. This was a real 
performance boost compared to putting data into a separate slot, 
otherwise even working with many images and large sets I have not really 
noticed any performance issues that would be connected with the speed of 
method dispatch (although I would not argue that there are none).

Regards,
Oleg

Dr Oleg Sklyar | EBI-EMBL, Cambridge CB10 1SD, UK | +44-1223-494466


John Chambers wrote:
> Prof Brian Ripley wrote:
>> On Mon, 7 Jan 2008, Robin Hankin wrote:
>>
>>   
>>> I am writing a package and need to decide whether to use S3 or S4.
>>>
>>> I have a single class, "multipol"; this needs methods for "[" and "[<-"
>>> and I also need a print (or show) method and methods for arithmetic +-
>>> */^.
>>>
>>> In S4, an object of class "multipol" has one slot that holds an array.
>>>
>>> Objects of class "multipol" require specific arithmetic operations;
>>> a,b being
>>> multipols means that a+b and a*b are defined in peculiar ways
>>> that make sense in the context of the package. I can also add and
>>> multiply
>>> by scalars (vectors of length one).
>>>     
> One thing you cannot do in S3 is to have methods that depend on anything 
> but the first argument.  Do you want something sensible for 1 + a when  
> a is a "multipol"? The default call to the primitive version may or may 
> not give you what you want.
>>> My impression is that S3 is perfectly adequate for this task, although
>>> I've not yet finalized the coding.
>>>
>>> S4 seems to be "overkill" for such a simple system.
>>>
>>> Can anyone give me some motivation for persisting with S4?
>>>
>>> Or indeed reassure me that S3 is a good design decision?
>>>     
>> Does performance matter?: S4 dispatch is many times slower than S3 
>> dispatch for such functions. (It is several times slower in general, but 
>> the difference is particularly marked for primitives.)
>>   
> Well, the question is whether performance of _method dispatch_ matters, 
> which it tends not to in many cases. And it would be good to have some 
> data to clarify "many times slower".  Generally, looking up inherited 
> methods is likely to take a while, but only the first time does R code 
> need to work out the inheritance. On repeated calls with the same 
> signature, dispatch should be basically a lookup in an environment.
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel