>>>>> Abs Spurdle
>>>>> on Thu, 28 Mar 2019 12:26:36 +1300 writes:
>> the prison made by ancient design choices
> That prison of ancient design choices isn't so bad.
> I have no further comments on object oriented semantics.
> However, I'm planning to follow the following design
> pattern.
> If I set the class of an object, I will append the new
> class to the existing class.
> #good class (object) = c ("something", class (object) )
#even better ;-)
class(object) <- c("something", class(object))
> #bad class (object) = "something"
> I encourage others to do the same.
Indeed.
BUT also tell the thousands of people who do it -- including
somewhat famous R package authors --
*NOT* to use things such as
if(class(x) == "Date")
or
switch(class(x),
"Date" = ...... ,
"POSIXct" = ...... ,
....
...
stop("invalid class: ", class(x)))
BUT to always use
inherits(x, "....")
There may be rare exceptions where using class(x)[1] is
good, but I have seen many cases where class(x)[1] was used
and the R programmers found it smart they knew that
class(x) can be of length more than one, but really their code
would fail *exactly* because good R programmers do *prepend*
their S3 class extension/specialization to the already existing
class.
---
... and then, I do agree with Gabe that (in some cases), using
formal (aka "S4") classes is really what one should do in order
to get a clean interface.
Martin Maechler
ETH Zurich and R Core Team
______________________________________________
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
