Dirk Eddelbuettel wrote:
On 31 January 2006 at 21:52, Greg Kochanski wrote:
| The thing I enjoy most about interacting with R developers
| is their grasp of the obvious.
|
| As I said, R incorrectly handled a user error.
|
|
| I get the feeling (correct me if I'm wrong) that users of R
| are assumed to be perfect?
|
Greg, this is Unix we're talking about. It has a certain set of assumptions,
one of which is that it is your (or your sys admins's) mistake if you try to
write somewhere where you have no permissions.
Never mind. I was trying to get beyond the obvious, and raise
the question of whether the ownership was set in the best possible
way, or if (perhaps) it might be better if some group other than
root owned the directory.
But, never mind. I've been working on lecture notes, planning
to explain R to a bunch of students. I've just come to the
conclusion that I don't want to do that after all. In my opinion
there are too many traps in R that will catch a novice,
and I can't bring myself to recommend R to them.
Part of the reason for that decision is that I've been
completely unable to elicit any sympathy from R developers
to problems that a novice user is likely to have.
That makes me suspect that there are far more hidden
traps than those that I have found, because it
seems likely that there has been no effort to remove them.
Indeed, I used S back in the 1990s when I was at Bell Labs,
and I find the current state of R eerily reminiscent of the
state of S then. It was simultaneously brilliant and
infinitely annoying, with elegant statistics coupled
with a complete lack of concern for the user. At the time,
it was basically research code, built by researchers for their
own use, so that was only to be expected.
R, I expected would be different. Most open source
projects are driven by bug reports, and consequently
a lot of the rough edges get removed. I expected something
like that to happen to R. I was mistaken.
Anyhow, I don't expect to trouble you further.
Return to your dreamy slumber where real users
don't make errors, and if they do, they deserve whatever
they get.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
