ts and defaults

[Paul Gilbert]

>ts() in R is incompatible
>with ts() in S, and the name has already been taken for the constructor
>of the R ts class. (In case you still do not get it, there is already
>an R ts class.)

I have a very large set of test and ts() in R is suprisingly compatible with S,
even though it is implemented in a very different way. A lot of S time series
code runs directly in R with no changes. This does not affect my own code very
much, since I define my own classes too. In fact, most everyone who works much
with time series probably defines their own classes. You are not the first to
notice how bad the defaults are. But it does affect a lot of code which users
have, and for which they might want to use R, or even contribute.

>Do you want us to change ts in R not to be the constructor
>of the ts class, and thereby `break a lot of user code' in R?

No, I want you to change the name of your new improved class of time series and
leave "ts" alone or make it even more compatible with S.

Paul

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._