On 4/21/05, Jan T. Kim <[EMAIL PROTECTED]> wrote: > On Thu, Apr 21, 2005 at 01:01:54PM -0400, Roger D. Peng wrote: > > One important thing to remember, which I think some more > > experienced programmers may forget, is that R is two things---a > > programming language and an *interactive* system for statistics > > and graphics. Maintaining the "interactive-ableness" of R may > > have imposed certain design choices. I personally think the > > current S4 system of generics/methods is quite suitable for both > > the "programming" and "interactive" sides of R. > > That's certainly a valid point. A more "standard" kind of > object orientation does not necessarily impair interactive > use, however. Python is no less usable interactively than R, > for example.
Again, it depends on what you are doing. As much as I like Python (and I really do), interactive construction of (moderately complex) classes is a PITA (compared to class structures in S4, or prototypes with XLispStat which can be done on a single line instead of many). Now, interactive Java via Omegahat or Bean-shell is worse (more annoying) for class construction, for certain definitions of worse. Which brings up the problem with imprecision, that all examples are equally "usable", for various definitions of usability and contexts that programmers would work in. best, -tony "Commit early,commit often, and commit in a repository from which we can easily roll-back your mistakes" (AJR, 4Jan05). A.J. Rossini [EMAIL PROTECTED] ______________________________________________ R-devel@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-devel