[Rd] Another wishlist for R

[Patrick Burns]

Prof Brian Ripley wrote:

>On Mon, 26 Jan 2004, Duncan Murdoch wrote:
>
>  
>
>>On 26 Jan 2004 15:17:01 +0100, Peter Dalgaard
>><p.dalgaard at biostat.ku.dk> wrote :
>>
>>    
>>
>>>Either setup so that cat() will be used to print it (add class + print
>>>method) or return noquote(....) The latter will give this effect:
>>>
>>>      
>>>
>>>>noquote("function (x, y = NULL, type = \"p\", xlim = NULL, ylim = NULL, ")
>>>>        
>>>>
>>>[1] function (x, y = NULL, type = "p", xlim = NULL, ylim = NULL,
>>>
>>>i.e. include line numbers.
>>>      
>>>
>>I think this produces a nice display:
>>
>>    
>>
>>>tail.function
>>>      
>>>
>>function (x, n = 6) {
>>    lines <- matrix(deparse(x),ncol=1)
>>    
>>
>
>I would just use as.matrix(deparse(x)).
>
>  
>
>>    rownames(lines) <- 1:nrow(lines)
>>    colnames(lines) <- ''
>>    noquote(tail(lines,n=n))
>>}
>>
>>    
>>
>>>tail(plot.default)
>>>      
>>>
>>                                                                   
>>36         box(...)                                                
>>37     if (ann)                                                    
>>38         title(main = main, sub = sub, xlab = xlab, ylab = ylab, 
>>39             ...)                                                
>>40     invisible()                                                 
>>41 }                        
>>
>>Unfortunately, I doubt if people would really want the result to be a
>>matrix, so maybe a new class is what is needed.
>>    
>>
>
>I knew this happened, but I did not know what people wanted.  I can only 
>see this being used interactively (it at all: I would use page), so that 
>seems as good as any.
>
>  
>
I like Duncan's idea.  I doubt I have ever used head on a function
for real, but tail of a function can be handy for seeing the structure
of the return value.  I agree that this is only likely to be used 
interactively.

Pat