[R-SIG-Mac] Feedback to Apple?
Joerg van den Hoff
j.van_den_hoff at fzd.de
Wed Mar 21 18:14:05 CET 2007
as I said, the current state of X11 seems OK to me. it's in the bits and pieces
were some things could need a caring hand and generally, my mail was more a
pledge to keep X11 support in focus in the future. specifically, I mentioned
the 'sync' between utility/command and manpage and partly even missing manpage
(no example at hand, I only notice them by and then when really searching for
more generally, I'm also concerned with command line support in such a way
that HFS+ specifics are handled gracefully in all cases.
another example is behaviour of `.Xauthority': completely unreproducible
but with finite probability (say every two weeks) my G5 comes up from sleep
and the X11 server refuses access so I can't open a new xterm, say. only copying
over the old state of .Xauthority from the evening before (which I have
started to copy via a script prior to sleep) reestablishes the access.
I'm not sure, if this is a specific mac problem, but it looks that way.
On Wed, Mar 21, 2007 at 12:21:54PM -0400, Simon Urbanek wrote:
> I'm a bit puzzled about the X11-related statements here. Both
> mentioned "better X11 support" - what is really missing in Apple's
> X11? So far I have found Apple's X11 very enjoyable for all
> scientific purposes and much easier to work with that other 3rd-party
> X11 implementations.
> As for installing X11, yes, I would welcome if it was in included the
> standard install mode, but it is in no way hidden: all Mac OS X
> installation DVDs have it and you only have to start the Mac OS X
> installer - one double-click, that's all. (Besides, rumor has it that
> this may change for Leopard anyway.) For a class, why not just supply
> the package?
> Concerning R itself, my only worry (and a big headache) is the lack
> of Fortran support in Xcode. We have to install parallel compilers
> and system libraries which is a maintenance nightmare...
> On Mar 21, 2007, at 10:55 AM, Dan Putler wrote:
> >Let me second Joerg van den Hoff's call for better X11 support. I'm
> >using R via a modified version of the R Commander GUI (that uses the
> >tcltk package, which requires X11 on the OS X platform) in a couple
> >of different classes. One thing that I found frustrating is that
> >Apple seems to be going to greater lengths overtime to hide the
> >X11user.pkg. This makes installing R on OS X seem really complicated
> >to students. The installation of the course software is comparatively
> >much more straightforward on the Windows side, leaving the
> >unfortunate impression that Windows is more suited to statistical
> >On 21-Mar-07, at 4:49 AM, Joerg van den Hoff wrote:
> >>On Tue, Mar 20, 2007 at 07:37:14PM -0400, Geoffrey Hutchison wrote:
> >>>Hello fellow R users!
> >>>I'm currently serving as a committee member at MacResearch.org, a
> >>>website devoted to science, math, and statistics on Macs. In about
> >>>two weeks, we'll be visiting Apple to talk about issues, needs,
> >>>desires of science users and developers:
> >>>So I was writing to ask if you have particular questions or concerns
> >>>which I can proxy? I think this is particularly important, since
> >>>of the other MacResearch representatives are statistics folks.
> >>>(I'm a
> >>>chemist -- I'm a happy user of R for data exploration and analysis.)
> >>>For example, are all the R.app compiler problems solved? Do we need
> >>>graphing/plotting libraries for visualization from Apple? What
> >>>feedback can I pass along?
> >>>Many thanks,
> >>I don't have a R- specific issue. what I _can_ say is, that for
> >>Macs to be
> >>accepted in scientific environments it's mandatory, that the bare
> >>bones UNIX
> >>side of OSX, notably full screen X11 is working competitivly to,
> >>say, LINUX or
> >>FreeBSD machines. for scientific purposes I find the combination of
> >>some Aqua
> >>apps and many X11 apps and UNIX utilities especially helpful and
> >>superior to
> >>using anything else.
> >>right now X11 is OK (but not optimal), but I don't know, how the
> >>future support
> >>for that area will look like.
> >>so if you could transport the message, that many many tools require
> >>X11 and
> >>many scientists are UNIX 'addicts', this might help (although they
> >>sshould know).
> >>what I have noted is, that e.g. manpage support is no good: it
> >>seems that
> >>quite a few of them are out of sync with the actual state of affairs.
> >>one specific wish would be a 'canonical' rsync supporting HFS+
> >>stuff, such
> >>as aliases, too. (they did it with `cp', for instance). same for
> >>other UNIX
> >>utilities which copy data (dd, tar, ...).
> >>R-SIG-Mac mailing list
> >>R-SIG-Mac at stat.math.ethz.ch
> >R-SIG-Mac mailing list
> >R-SIG-Mac at stat.math.ethz.ch
More information about the R-SIG-Mac