[R-SIG-Mac] R and Mac distributions [Was: Problems loading Rcmdr on a Mac 10.7.2 and a Mac 10.6.8]

[Simon Urbanek]

On Sep 20, 2013, at 11:29 AM, peter dalgaard <pdalgd at gmail.com> wrote:

> 
> On Sep 20, 2013, at 08:49 , Prof Brian Ripley wrote:
>>> 
>> 
>> I am not so sure about that.  It will do no harm running CRAN binary R.  But it does mean that there are multiple versions of X11 things about and that will cause problems for other things (including possibly installing packages from source).
> 
> 
> Or, put differently, we could be shipping libraries that are incompatible with the system libraries (or 3rd party libraries for that matter) if those are not kept up to date. 
> 
> Of course, keeping versions compatible is exactly what package managers are for, on systems that have them. The way things work on commercial platforms tend to lead each applicationq into rolling everything you need into the application itself, which is - in the bigger perspective - wasteful and somewhat risky (as defective components only get replaced upon upgrade of the application). 
> 
> I know Simon distrusts them, but sometimes I do feel that we should work more closely with the people doing managed distributions for OSX, i.e. MacPorts, Homebrew, Fink. (Not sure how they'd cope with missing OS updates, though.)
> 

I don't have the time or resources to support more than one distribution. Given that the native port is what covers the vast majority of users, I feel that I'm spending my limited time well. However, I would certainly encourage people that want to create other R distributions to collaborate with us and I'd be very happy to share my experience and discuss issues with them. Historically, distributions like MacPorts or Fink had very poor quality of ports (to a point where R was not really working), mainly because the authors had apparently no idea about either R or OS X. I was never able to get any response to issues raised, so I gave up. The bright exception is Homebrew - it has much better design (it doesn't try to setup a parallel universe that breaks everything) and I was working with Max on technical issues in the early days, so the my experience was very positive. I didn't check it out recently (again, lack of spare time), but the design was very promising. 

Cheers,
Simon


> -- 
> Peter Dalgaard, Professor
> Center for Statistics, Copenhagen Business School
> Solbjerg Plads 3, 2000 Frederiksberg, Denmark
> Phone: (+45)38153501
> Email: pd.mes at cbs.dk  Priv: PDalgd at gmail.com
> 
> _______________________________________________
> R-SIG-Mac mailing list
> R-SIG-Mac at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
> 
>