[rkward-devel] a "misfeature"

[Deepayan Sarkar]

On 4/7/07, Thomas Friedrichsmeier
<thomas.friedrichsmeier at ruhr-uni-bochum.de> wrote:
> On Thursday 05 April 2007 22:03, Deepayan Sarkar wrote:
> > Ah, so lazy loading is bad enough and rk.get.structure is even worse :-).
>
> Exactly. Though actually I think lazy loading is not that bad at all. Given
> that GO is one heck of a large package, 44 seconds for reading the data may
> not be nice, but I wouldn't call it terrible, either.
>
> > Sure. It would be interesting to see how the Mac GUI behaves (since it
> > also shows the argument list in functions). I'll check when I get a
> > chance.
>
> I just had a quick look (still not too much time, at the moment), and it looks
> like it always fetches that information from R, instead of generating a

Yes, the evaluation seems to be done when the "(" key is pressed.

> persistent representation, like we do in RKWard. In fact, this approach has
> some obvious advantages, and may be preferrable in the long run - if only R
> would allow concurrent threaded evaluations (which still seems far away
> despite at least some efforts, esp. by Luke Tierney and Duncan Temple Lang,
> as far as I can see).

Well, coding practices vary of course, but I don't think it's very
common to either write code or go into an object browser like
interface while R is busy (except maybe the latter when debugging).
It's certainly not worth it to me if it comes at the cost of huge
package loading times. But then, no GUI can be everything to everyone,
and you'll have to decide what your priorities are.

-Deepayan